- Support for OS specific/native audio devices/backends has been added, that is PulseAudio (Linux), WASAPI (Windows) and CoreAudio (MacOS).
- New sound effects have been added, namely Modulator and Equalizer. Thanks to Marcos Perez for contributing the Equalizer.
- File stream info: if an audio file contains multiple streams you can choose which one to process instead of taking the first one (this feature is only supported by ffmpeg, not libsndfile).
- API Change: double instead of float for time values for more precise timing control.
- There have been lots of bugfixes, which are basically the majority of all changes.
- And some other minor improvements were implemented as well.
Detailed list of changes:
d4042d9 Port changes in Blender to upstream.
b60fb45 Equalizer
ab04e84 Fixes
8f0c305 Fix build error with MSVC 17.4+ ported from Blender.
ce44342 Minor documentation update.
cdcb3f4 Migrate from distuils to setuptools for python module.
21eccef Fix FindFFTW to find the float version fftw3f.
ab15e2f Bugfix: API change in new ffmpeg version.
a097be8 Clang format file added, valid from now on.
2fc9fb7 Porting bugfix from Blender upstream.
bb655b7 Bugfix: wrong sample size computation for PulseAudio.
a150495 Bugfix: Buffer did not support buffers > 2 GB.
034645c Update for ffmpeg 5.
932739c Bugfix: WASAPI hangs.
4fcd47c WASAPI: fix bug when switching the default device while there is no playback.
a16fbd2 Python API: fix to get convolution in the python API.
27ac5c1 WASAPI: always switch to default audio device.
1b03e6c Bugfix: catch exception if file cannot be read.
- FFMPEG file writing crashed, also corrected pts for encoding
- silenced Doxygen warnings about undefined defines
- C++ API:
- ResampleReader uses specs instead of sample rate
- Bindings API:
- writing sounds to files
- reading sound data, specs and length
- resampling sounds
- CMake/Building:
- first steps towards building for Mac
- windows builds copy dlls automatically
- Python module:
- using distutils instead of setuptools
- added numpy as dependency
- Documentation:
- added windows building and plugin documentation
- disabled html timestamps in doxygen
- updated sphinx template
- build binding documentation without installing the python module
Detailed list of changes:
326a300 Documentation: windows, dll copying now done automatically.
54cac4f Windows: install dlls.
65c2d78 Bindings: Sound length and specs properties.
c38da70 Bindings API: adding resampling.
374822f Documentation: Added windows and plugin documentation.
a9dc5b9 Python module: add numpy as dependency.
c933a02 C API: implement new API based on the python API.
ac54c52 Python API: silence numpy warnings.
c9491bb Python API: checking for a positive sample rate.
4eb1fa8 Python API: reorder functions.
ec7c00b Sphinx update and fixes.
e16d979 FFMPEG: correct pts during encoding.
7ab3935 Documentation: git path fix.
28d77bb Python: use distutils directly instead of setuptools.
1f43284 Silence doxygen warning about undefined defines.
0d52458 CMake: improvements and fixes for building on Mac.
37daedf FFMPEG: bugfixes for file writing.
780ca2a ResampleReader API change
4d9863d Python API: Optimization for cached sounds' data access.
ea04fee Python API: read sound data and create sound buffers as well as getting the specs of a sound.
335b293 Python sound writing API.
36a7252 Pkgconfig: use cmake configured library directory.
5503908 Doxygen: disable html timestamps.
Initial Release of Audaspace 1.0
================================
Audaspace has been the internal audio library of blender since blender 2.5. It is now released as a standalone library to be used in other projects as well.