Speex: Difference between revisions
Jump to navigation
Jump to search
(outdated code in OPUS TOOLS) |
(sorry for editing an "inactive historical page" ... but SpeexDSP is undocumented) |
||
| Line 12: | Line 12: | ||
== Usefulness, Speex DSP == | == Usefulness, Speex DSP == | ||
In 2009 the source code of Speex was split into "Speex codec" and "Speex DSP". While the codec is deprecated, Speex DSP is still useful and still under development, it contains among others resampling code (written originally 2007 for the Speex codec) used by several other audio projects, including OPUS TOOLS (actually OPUS TOOLS 0.1.9 from 2014 contains an outdated version of the resampling code from 2009 with minimal modifications). | In 2009 the source code of Speex was split into "Speex codec" and "Speex DSP". While the codec is deprecated now (has been since 2013), Speex DSP is still useful and still under development, it contains among others resampling code (written originally 2007 for the Speex codec) used by several other audio projects, including OPUS TOOLS (actually OPUS TOOLS 0.1.9 from 2014 contains an outdated version of the resampling code from 2009 with minimal modifications). | ||
== Tasks == | == Tasks == | ||
Revision as of 01:44, 25 July 2016
The Speex codec is deprecated! Xiph recommends you use the superior Opus codec instead.
Website
The Speex homepage has all the project info.
There is also a Speex FAQ.
Hardware
See Speex hardware for a partial list of supported hardware
Usefulness, Speex DSP
In 2009 the source code of Speex was split into "Speex codec" and "Speex DSP". While the codec is deprecated now (has been since 2013), Speex DSP is still useful and still under development, it contains among others resampling code (written originally 2007 for the Speex codec) used by several other audio projects, including OPUS TOOLS (actually OPUS TOOLS 0.1.9 from 2014 contains an outdated version of the resampling code from 2009 with minimal modifications).
Tasks
These are some improvements that could be made to Speex.
Let us know if you'd like to work on one of them.
- Speech/signal processing (DSP design)
- Improve noise suppression (get rid of musical noise) and residual echo suppression
- Improve packet-loss concealment (PLC)
- Re-write the built-in voice activity detector (VAD)
- Improve the 2.15 kbps vocoder mode (there are even 4 unused bits left to use)
- Algorithmic optimizations (see if some searches can be simplified/approximated)
- Complete fixed-point (DSP development)
- Wideband
- VBR
- Rest of the narrowband modes
- Preprocessor (noise suppression, AGC)
- Jitter buffer
- Arch-specific optimization
- More...
- Tune (playing with parameters)
- Noise weighting filter
- Perceptual enhancement
- Features (plain C programming)
- Implement maximum VBR bit-rate
- Implement peeling (write functions to strip some of the bits)
- Peel high-band (wideband -> narrowband)
- Transform 24.6 kbps mode to 15 kbps mode
- Documentation
- Use questions from the mailing list to create a better FAQ on this wiki
- Update the Speex manual based on recent papers
- Improve libspeex documentation
- Write good example code
- Split off the documentation of SpeexDSP ... the latest release "speexdsp-1.2rc3.tar.gz" from 2015 still contains "The Speex Codec Manual Version 1.2 Beta 3" from 2007 ... it documents the codec part, not the DSP part, further [[1]] contains some dead links, and it does not prominently document that only the codec is deprecated, but the DSP part is not