blender/source/gameengine/VideoTexture/ImageBase.cpp

704 lines
16 KiB
C++
Raw Normal View History

/*
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
-----------------------------------------------------------------------------
This source file is part of VideoTexture library
Copyright (c) 2007 The Zdeno Ash Miklas
This program is free software; you can redistribute it and/or modify it under
the terms of the GNU Lesser General Public License as published by the Free Software
Foundation; either version 2 of the License, or (at your option) any later
version.
This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License along with
this program; if not, write to the Free Software Foundation, Inc., 59 Temple
Place - Suite 330, Boston, MA 02111-1307, USA, or go to
http://www.gnu.org/copyleft/lesser.txt.
-----------------------------------------------------------------------------
*/
/** \file gameengine/VideoTexture/ImageBase.cpp
* \ingroup bgevideotex
*/
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
#include "ImageBase.h"
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
extern "C" {
#include "bgl.h"
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
}
#include "GL/glew.h"
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
#include <vector>
#include <string.h>
#include <PyObjectPlus.h>
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
#include <structmember.h>
#include "FilterBase.h"
#include "Exception.h"
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
#if (defined(WIN32) || defined(WIN64)) && !defined(FREE_WINDOWS)
#define strcasecmp _stricmp
#endif
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// ImageBase class implementation
// constructor
ImageBase::ImageBase (bool staticSrc) : m_image(NULL), m_imgSize(0),
m_avail(false), m_scale(false), m_scaleChange(false), m_flip(false),
m_staticSources(staticSrc), m_pyfilter(NULL)
{
m_size[0] = m_size[1] = 0;
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
m_exports = 0;
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
}
// destructor
ImageBase::~ImageBase (void)
{
// release image
if (m_image)
delete [] m_image;
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
}
// release python objects
bool ImageBase::release (void)
{
// iterate sources
for (ImageSourceList::iterator it = m_sources.begin(); it != m_sources.end(); ++it)
{
// release source object
delete *it;
*it = NULL;
}
// release filter object
Py_XDECREF(m_pyfilter);
m_pyfilter = NULL;
return true;
}
// get image
unsigned int * ImageBase::getImage (unsigned int texId, double ts)
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
{
// if image is not available
if (!m_avail)
{
// if there are any sources
if (!m_sources.empty())
{
// get images from sources
for (ImageSourceList::iterator it = m_sources.begin(); it != m_sources.end(); ++it)
// get source image
(*it)->getImage(ts);
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// init image
init(m_sources[0]->getSize()[0], m_sources[0]->getSize()[1]);
}
// calculate new image
calcImage(texId, ts);
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
}
// if image is available, return it, otherwise NULL
return m_avail ? m_image : NULL;
}
// refresh image source
void ImageBase::refresh (void)
{
// invalidate this image
m_avail = false;
// refresh all sources
for (ImageSourceList::iterator it = m_sources.begin(); it != m_sources.end(); ++it)
(*it)->refresh();
}
// get source object
PyImage * ImageBase::getSource (const char * id)
{
// find source
ImageSourceList::iterator src = findSource(id);
// return it, if found
return src != m_sources.end() ? (*src)->getSource() : NULL;
}
// set source object
bool ImageBase::setSource (const char * id, PyImage * source)
{
// find source
ImageSourceList::iterator src = findSource(id);
// check source loop
if (source != NULL && source->m_image->loopDetect(this))
return false;
// if found, set new object
if (src != m_sources.end())
// if new object is not empty or sources are static
if (source != NULL || m_staticSources)
// replace previous source
(*src)->setSource(source);
// otherwise delete source
else
m_sources.erase(src);
// if source is not found and adding is allowed
else
if (!m_staticSources)
{
// create new source
ImageSource * newSrc = newSource(id);
newSrc->setSource(source);
// if source was created, add it to source list
if (newSrc != NULL) m_sources.push_back(newSrc);
}
// otherwise source wasn't set
else
return false;
// source was set
return true;
}
// set pixel filter
void ImageBase::setFilter (PyFilter * filt)
{
// reference new filter
if (filt != NULL) Py_INCREF(filt);
// release previous filter
Py_XDECREF(m_pyfilter);
// set new filter
m_pyfilter = filt;
}
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
ExceptionID ImageHasExports;
ExceptionID InvalidColorChannel;
ExpDesc ImageHasExportsDesc (ImageHasExports, "Image has exported buffers, cannot resize");
ExpDesc InvalidColorChannelDesc (InvalidColorChannel, "Invalid or too many color channels specified. At most 4 values within R, G, B, A, 0, 1");
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// initialize image data
void ImageBase::init (short width, short height)
{
// if image has to be scaled
if (m_scale)
{
// recalc sizes of image
width = calcSize(width);
height = calcSize(height);
}
// if sizes differ
if (width != m_size[0] || height != m_size[1])
{
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
if (m_exports > 0)
THRWEXCP(ImageHasExports,S_OK);
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// new buffer size
unsigned int newSize = width * height;
// if new buffer is larger than previous
if (newSize > m_imgSize)
{
// set new buffer size
m_imgSize = newSize;
// release previous and create new buffer
if (m_image)
delete [] m_image;
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
m_image = new unsigned int[m_imgSize];
}
// new image size
m_size[0] = width;
m_size[1] = height;
// scale was processed
m_scaleChange = false;
}
}
// find source
ImageSourceList::iterator ImageBase::findSource (const char * id)
{
// iterate sources
ImageSourceList::iterator it;
for (it = m_sources.begin(); it != m_sources.end(); ++it)
// if id matches, return iterator
if ((*it)->is(id)) return it;
// source not found
return it;
}
// check sources sizes
bool ImageBase::checkSourceSizes (void)
{
// reference size
short * refSize = NULL;
// iterate sources
for (ImageSourceList::iterator it = m_sources.begin(); it != m_sources.end(); ++it)
{
// get size of current source
short * curSize = (*it)->getSize();
// if size is available and is not empty
if (curSize[0] != 0 && curSize[1] != 0) {
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// if reference size is not set
if (refSize == NULL) {
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// set current size as reference
refSize = curSize;
// otherwise check with current size
} else if (curSize[0] != refSize[0] || curSize[1] != refSize[1]) {
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// if they don't match, report it
return false;
}
}
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
}
// all sizes match
return true;
}
// compute nearest power of 2 value
short ImageBase::calcSize (short size)
{
// while there is more than 1 bit in size value
while ((size & (size - 1)) != 0)
// clear last bit
size = size & (size - 1);
// return result
return size;
}
// perform loop detection
bool ImageBase::loopDetect (ImageBase * img)
{
// if this object is the same as parameter, loop is detected
if (this == img) return true;
// check all sources
for (ImageSourceList::iterator it = m_sources.begin(); it != m_sources.end(); ++it)
// if source detected loop, return this result
if ((*it)->getSource() != NULL && (*it)->getSource()->m_image->loopDetect(img))
return true;
// no loop detected
return false;
}
// ImageSource class implementation
// constructor
ImageSource::ImageSource (const char * id) : m_source(NULL), m_image(NULL)
{
// copy id
int idx;
for (idx = 0; id[idx] != '\0' && idx < SourceIdSize - 1; ++idx)
m_id[idx] = id[idx];
m_id[idx] = '\0';
}
// destructor
ImageSource::~ImageSource (void)
{
// release source
setSource(NULL);
}
// compare id
bool ImageSource::is (const char * id)
{
for (char * myId = m_id; *myId != '\0'; ++myId, ++id)
if (*myId != *id) return false;
return *id == '\0';
}
// set source object
void ImageSource::setSource (PyImage * source)
{
// reference new source
if (source != NULL) Py_INCREF(source);
// release previous source
Py_XDECREF(m_source);
// set new source
m_source = source;
}
// get image from source
unsigned int * ImageSource::getImage (double ts)
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
{
// if source is available
if (m_source != NULL)
// get image from source
m_image = m_source->m_image->getImage(0, ts);
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// otherwise reset buffer
else
m_image = NULL;
// return image
return m_image;
}
// refresh source
void ImageSource::refresh (void)
{
// if source is available, refresh it
if (m_source != NULL) m_source->m_image->refresh();
}
// list of image types
PyTypeList pyImageTypes;
// functions for python interface
// object allocation
PyObject * Image_allocNew (PyTypeObject * type, PyObject * args, PyObject * kwds)
{
// allocate object
PyImage * self = reinterpret_cast<PyImage*>(type->tp_alloc(type, 0));
// initialize object structure
self->m_image = NULL;
// return allocated object
return reinterpret_cast<PyObject*>(self);
}
// object deallocation
void Image_dealloc (PyImage * self)
{
// release object attributes
if (self->m_image != NULL)
{
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
if (self->m_image->m_exports > 0)
{
PyErr_SetString(PyExc_SystemError,
2011-09-03 02:15:49 +00:00
"deallocated Image object has exported buffers");
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
PyErr_Print();
}
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
// if release requires deleting of object, do it
if (self->m_image->release())
delete self->m_image;
self->m_image = NULL;
}
}
// get image data
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
PyObject * Image_getImage (PyImage * self, char * mode)
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
{
try
{
unsigned int * image = self->m_image->getImage();
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
if (image)
{
// build BGL buffer
int dimensions = self->m_image->getBuffSize();
Buffer * buffer;
if (mode == NULL || !strcasecmp(mode, "RGBA"))
{
buffer = BGL_MakeBuffer( GL_BYTE, 1, &dimensions, image);
}
else
{
int i, c, ncolor, pixels;
int offset[4];
unsigned char *s, *d;
// scan the mode to get the channels requested, no more than 4
for (i=ncolor=0; mode[i] != 0 && ncolor < 4; i++)
{
switch (toupper(mode[i]))
{
case 'R':
offset[ncolor++] = 0;
break;
case 'G':
offset[ncolor++] = 1;
break;
case 'B':
offset[ncolor++] = 2;
break;
case 'A':
offset[ncolor++] = 3;
break;
case '0':
offset[ncolor++] = -1;
break;
case '1':
offset[ncolor++] = -2;
break;
// if you add more color code, change the switch further down
default:
THRWEXCP(InvalidColorChannel,S_OK);
}
}
if (mode[i] != 0) {
THRWEXCP(InvalidColorChannel,S_OK);
}
// first get the number of pixels
pixels = dimensions / 4;
// multiple by the number of channels, each is one byte
dimensions = pixels * ncolor;
// get an empty buffer
buffer = BGL_MakeBuffer( GL_BYTE, 1, &dimensions, NULL);
// and fill it
for (i=0, d=(unsigned char*)buffer->buf.asbyte, s=(unsigned char*)image;
i<pixels;
++i, d+=ncolor, s+=4)
{
for (c=0; c<ncolor; c++)
{
switch (offset[c])
{
case 0: d[c] = s[0]; break;
case 1: d[c] = s[1]; break;
case 2: d[c] = s[2]; break;
case 3: d[c] = s[3]; break;
case -1: d[c] = 0; break;
case -2: d[c] = 0xFF; break;
}
}
}
}
return (PyObject*)buffer;
}
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
}
catch (Exception & exp)
{
exp.report();
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
return NULL;
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
}
Py_RETURN_NONE;
}
// get image size
PyObject * Image_getSize (PyImage * self, void * closure)
{
return Py_BuildValue("(hh)", self->m_image->getSize()[0],
self->m_image->getSize()[1]);
}
// refresh image
PyObject * Image_refresh (PyImage * self)
{
self->m_image->refresh();
Py_RETURN_NONE;
}
// get scale
PyObject * Image_getScale (PyImage * self, void * closure)
{
if (self->m_image != NULL && self->m_image->getScale()) Py_RETURN_TRUE;
else Py_RETURN_FALSE;
}
// set scale
int Image_setScale (PyImage * self, PyObject * value, void * closure)
{
// check parameter, report failure
if (value == NULL || !PyBool_Check(value))
{
PyErr_SetString(PyExc_TypeError, "The value must be a bool");
return -1;
}
// set scale
if (self->m_image != NULL) self->m_image->setScale(value == Py_True);
// success
return 0;
}
// get flip
PyObject * Image_getFlip (PyImage * self, void * closure)
{
if (self->m_image != NULL && self->m_image->getFlip()) Py_RETURN_TRUE;
else Py_RETURN_FALSE;
}
// set flip
int Image_setFlip (PyImage * self, PyObject * value, void * closure)
{
// check parameter, report failure
if (value == NULL || !PyBool_Check(value))
{
PyErr_SetString(PyExc_TypeError, "The value must be a bool");
return -1;
}
// set scale
if (self->m_image != NULL) self->m_image->setFlip(value == Py_True);
// success
return 0;
}
// get filter source object
PyObject * Image_getSource (PyImage * self, PyObject * args)
{
// get arguments
char * id;
if (!PyArg_ParseTuple(args, "s:getSource", &id))
return NULL;
if (self->m_image != NULL)
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
{
// get source object
PyObject * src = reinterpret_cast<PyObject*>(self->m_image->getSource(id));
// if source is available
if (src != NULL)
{
// return source
Py_INCREF(src);
return src;
}
}
// source was not found
Py_RETURN_NONE;
}
// set filter source object
PyObject * Image_setSource (PyImage * self, PyObject * args)
{
// get arguments
char * id;
PyObject * obj;
if (!PyArg_ParseTuple(args, "sO:setSource", &id, &obj))
return NULL;
if (self->m_image != NULL)
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
{
// check type of object
if (pyImageTypes.in(Py_TYPE(obj)))
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
{
// convert to image struct
PyImage * img = reinterpret_cast<PyImage*>(obj);
// set source
if (!self->m_image->setSource(id, img))
{
// if not set, retport error
PyErr_SetString(PyExc_RuntimeError, "Invalid source or id");
return NULL;
}
}
// else report error
else
{
PyErr_SetString(PyExc_RuntimeError, "Invalid type of object");
return NULL;
}
}
// return none
Py_RETURN_NONE;
}
// get pixel filter object
PyObject * Image_getFilter (PyImage * self, void * closure)
{
// if image object is available
if (self->m_image != NULL)
{
// pixel filter object
PyObject * filt = reinterpret_cast<PyObject*>(self->m_image->getFilter());
// if filter is present
if (filt != NULL)
{
// return it
Py_INCREF(filt);
return filt;
}
}
// otherwise return none
Py_RETURN_NONE;
}
// set pixel filter object
int Image_setFilter (PyImage * self, PyObject * value, void * closure)
{
// if image object is available
if (self->m_image != NULL)
{
// check new value
if (value == NULL || !pyFilterTypes.in(Py_TYPE(value)))
VideoTexture module. The only compilation system that works for sure is the MSVC project files. I've tried my best to update the other compilation system but I count on the community to check and fix them. This is Zdeno Miklas video texture plugin ported to trunk. The original plugin API is maintained (can be found here http://home.scarlet.be/~tsi46445/blender/blendVideoTex.html) EXCEPT for the following: The module name is changed to VideoTexture (instead of blendVideoTex). A new (and only) video source is now available: VideoFFmpeg() You must pass 1 to 4 arguments when you create it (you can use named arguments): VideoFFmpeg(file) : play a video file VideoFFmpeg(file, capture, rate, width, height) : start a live video capture file: In the first form, file is a video file name, relative to startup directory. It can also be a URL, FFmpeg will happily stream a video from a network source. In the second form, file is empty or is a hint for the format of the video capture. In Windows, file is ignored and should be empty or not specified. In Linux, ffmpeg supports two types of device: VideoForLinux and DV1394. The user specifies the type of device with the file parameter: [<device_type>][:<standard>] <device_type> : 'v4l' for VideoForLinux, 'dv1394' for DV1394; default to 'v4l' <standard> : 'pal', 'secam' or 'ntsc', default to 'ntsc' The driver name is constructed automatically from the device types: v4l : /dev/video<capture> dv1394: /dev/dv1394/<capture> If you have different driver name, you can specify the driver name explicitely instead of device type. Examples of valid file parameter: /dev/v4l/video0:pal /dev/ieee1394/1:ntsc dv1394:ntsc v4l:pal :secam capture: Defines the index number of the capture source, starting from 0. The first capture device is always 0. The VideoTexutre modules knows that you want to start a live video capture when you set this parameter to a number >= 0. Setting this parameter < 0 indicates a video file playback. Default value is -1. rate: the capture frame rate, by default 25 frames/sec width: height: Width and height of the video capture in pixel, default value 0. In Windows you must specify these values and they must fit with the capture device capability. For example, if you have a webcam that can capture at 160x120, 320x240 or 640x480, you must specify one of these couple of values or the opening of the video source will fail. In Linux, default values are provided by the VideoForLinux driver if you don't specify width and height. Simple example ************** 1. Texture definition script: import VideoTexture contr = GameLogic.getCurrentController() obj = contr.getOwner() if not hasattr(GameLogic, 'video'): matID = VideoTexture.materialID(obj, 'MAVideoMat') GameLogic.video = VideoTexture.Texture(obj, matID) GameLogic.vidSrc = VideoTexture.VideoFFmpeg('trailer_400p.ogg') # Streaming is also possible: #GameLogic.vidSrc = VideoTexture.VideoFFmpeg('http://10.32.1.10/trailer_400p.ogg') GameLogic.vidSrc.repeat = -1 # If the video dimensions are not a power of 2, scaling must be done before # sending the texture to the GPU. This is done by default with gluScaleImage() # but you can also use a faster, but less precise, scaling by setting scale # to True. Best approach is to convert the video offline and set the dimensions right. GameLogic.vidSrc.scale = True # FFmpeg always delivers the video image upside down, so flipping is enabled automatically #GameLogic.vidSrc.flip = True if contr.getSensors()[0].isPositive(): GameLogic.video.source = GameLogic.vidSrc GameLogic.vidSrc.play() 2. Texture refresh script: obj = GameLogic.getCurrentController().getOwner() if hasattr(GameLogic, 'video') != 0: GameLogic.video.refresh(True) You can download this demo here: http://home.scarlet.be/~tsi46445/blender/VideoTextureDemo.blend http://home.scarlet.be/~tsi46445/blender/trailer_400p.ogg
2008-10-31 22:35:52 +00:00
{
// report value error
PyErr_SetString(PyExc_TypeError, "Invalid type of value");
return -1;
}
// set new value
self->m_image->setFilter(reinterpret_cast<PyFilter*>(value));
}
// return success
return 0;
}
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
PyObject * Image_valid(PyImage * self, void * closure)
{
if (self->m_image->isImageAvailable())
{
Py_RETURN_TRUE;
}
else
{
Py_RETURN_FALSE;
}
}
int Image_getbuffer(PyImage *self, Py_buffer *view, int flags)
{
unsigned int * image;
2011-09-01 02:12:53 +00:00
int ret;
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
try
{
// can throw in case of resize
image = self->m_image->getImage();
}
catch (Exception & exp)
{
// cannot return -1, this creates a crash in Python, for now we will just return an empty buffer
2010-06-06 00:35:48 +00:00
exp.report();
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
//return -1;
goto error;
}
if (!image)
{
// same remark, see above
//PyErr_SetString(PyExc_BufferError, "Image buffer is not available");
//return -1;
goto error;
}
2011-09-01 02:12:53 +00:00
if (view == NULL)
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
{
2011-09-01 02:12:53 +00:00
self->m_image->m_exports++;
return 0;
}
ret = PyBuffer_FillInfo(view, (PyObject*)self, image, self->m_image->getBuffSize(), 0, flags);
if (ret >= 0)
self->m_image->m_exports++;
return ret;
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
error:
// Return a empty buffer to avoid a crash in Python 3.1
// The bug is fixed in Python SVN 77916, as soon as the python revision used by Blender is
// updated, you can simply return -1 and set the error
static char* buf = (char *)"";
2011-09-01 02:12:53 +00:00
ret = PyBuffer_FillInfo(view, (PyObject*)self, buf, 0, 0, flags);
if (ret >= 0)
self->m_image->m_exports++;
return ret;
VideoTexture: improvements to image data access API. - Use BGL buffer instead of string for image data. - Add buffer interface to image source. - Allow customization of pixel format. - Add valid property to check if the image data is available. The image property of all Image source objects will now return a BGL 'buffer' object. Previously it was returning a string, which was not working at all with Python 3.1. The BGL buffer type allows sequence access to bytes and is directly usable in BGL OpenGL wrapper functions. The buffer is formated as a 1 dimensional array of bytes with 4 bytes per pixel in RGBA order. BGL buffers will also be accepted in the ImageBuff load() and plot() functions. It is possible to customize the pixel format by using the VideoTexture.imageToArray(image, mode) function: the first argument is a Image source object, the second optional argument is a format string using the R, G, B, A, 0 and 1 characters. For example "BGR" means that each pixel will be 3 bytes, corresponding to the Blue, Green and Red channel in that order. Use 0 for a fixed hex 00 value, 1 for hex FF. The default mode is "RGBA". All Image source objects now support the buffer interface which allows to create memoryview objects for direct access to the image internal buffer without memory copy. The buffer format is one dimensional array of bytes with 4 bytes per pixel in RGBA order. The buffer is writable, which allows custom modifications of the image data. v = memoryview(source) A bug in the Python 3.1 buffer API will cause a crash if the memoryview object cannot be created. Therefore, you must always check first that an image data is available before creating a memoryview object. Use the new valid attribute for that: if source.valid: v = memoryview(source) ... Note: the BGL buffer object itself does not yet support the buffer interface. Note: the valid attribute makes sense only if you use image source in conjunction with texture object like this: # refresh texture but keep image data in memory texture.refresh(False) if texture.source.valid: v = memoryview(texture.source) # process image ... # invalidate image for next texture refresh texture.source.refresh() Limitation: While memoryview objects exist, the image cannot be resized. Resizing occurs with ImageViewport objects when the viewport size is changed or with ImageFFmpeg when a new image is reloaded for example. Any attempt to resize will cause a runtime error. Delete the memoryview objects is you want to resize an image source object.
2010-02-21 22:20:00 +00:00
}
void Image_releaseBuffer(PyImage *self, Py_buffer *buffer)
{
self->m_image->m_exports--;
}
PyBufferProcs imageBufferProcs =
{
(getbufferproc)Image_getbuffer,
(releasebufferproc)Image_releaseBuffer
};