Difference between revisions of "OpusTodo"

From XiphWiki
Jump to: navigation, search
(Other: link to mozilla logo design bug)
(31 intermediate revisions by 6 users not shown)
Line 1: Line 1:
== Code ==
+
== 1.1-beta ==
  
=== For IETF draft ===
+
* Tune mode switching decisions
* Multi-channel signalling (done, needs more testing)
+
** Use SILK up to higher rates on voice
* Make opus-compare fail for single seriously trashed frames?
+
** Adapt stereo SILK/CELT threshold based on stereo width
 +
* Tune hybrid rate allocation
 +
* Figure out how to use speech/music detection optimally
 +
** find optimal switching time (low energy/tonality)
 +
* Improve variable frame size
 +
* Tune transient detector?
 +
* Use ALLOC in tonality analysis
 +
* LOTS of testing
  
== SILK issues ==
+
== Lower priority ==
* <s>CLANG ARITHMETIC UNDEFINED at <silk/decode_core.c, (108:40)> : Op: *, Reason : Signed Multiplication Overflow, BINARY OPERATION: left (int32): 916258817 right (int32): -3 </s>
 
* <s>CLANG ARITHMETIC UNDEFINED at <silk/decode_core.c, (108:40)> : Op: +, Reason : Signed Addition Overflow, BINARY OPERATION: left (int32): -418535217 right (int32): -1832517634</s>
 
* Two PLC IOC
 
* LBRR IOC
 
* CLANG ARITHMETIC UNDEFINED at <silk/CNG.c, (154:48)> : Op: +, Reason : Signed Addition Overflow, BINARY OPERATION: left (int32): -7107 right (int32): -2147478272
 
* CLANG ARITHMETIC UNDEFINED at <silk/VAD.c, (258:44)> : Op: +, Reason : Signed Addition Overflow, BINARY OPERATION: left (int32): -2138692500 right (int32): -16675982
 
  
 
+
* Handle packets with PLC frames followed by FEC
=== Later ===
+
* Better handling for the case where FEC has a different bandwidth than the current mode
* Usable command-line tools (<s>opus is a major regression from libcelt right now</s>)
+
* PLC transitions on unprotected SILK-SILK bandwidth changes?
* Smart automatic mode decision
 
* psymodel based VBR
 
* Remove copy in inverse MDCT
 
* Save some float<->int conversions
 
  
 
== Spec ==
 
== Spec ==
 +
* Ogg mapping. See [[http://tools.ietf.org/html/draft-ietf-codec-oggopus IETF draft]]
 +
* Matroska mapping. See: [[MatroskaOpus]]
 +
* RTP payload format See [[http://tools.ietf.org/html/draft-spittka-payload-rtp-opus IETF draft]]
  
* Finish codec draft
+
== Website ==
* Ogg mapping (including multi-channel). See: [[OggOpus]]
+
* De-uglify webpage - some suggestions: write about codecs obsoleted by OPUS (Speex, CELT, Vorbis(?), and the prop. ones), write about implementations (is there only one so far?), comparison table (Opus, Vorbis, Speex, ..., MP5) of features (channels, freq, bits per sample, license, language (C89), integer impl. (Vorbis decoder only, Opus YES, ...), future use in video files (Theora? Dirac? WebM? other future codecs...), audio files for storage (like Vorbis, no raw Opus defined, only inside OGG), ...  
* RTP payload format
+
* Promotional material (some nice free or Public domain sounds in Opus format)
  
 
== Other ==
 
== Other ==
  
* Logo See: [https://bugzilla.mozilla.org/show_bug.cgi?id=689261 Mozilla bug 689261] for some discussion
 
* Test vectors
 
* Listening tests
 
* Documentation (at a minimum every exported symbol should have complete and accurate documentation)
 
* Add content to opus-codec.org
 
** The above documentation
 
** Presentations
 
** Examples and test results  (hyperlink to Monty's demo, gmaxwell's HA results page, etc)
 
 
* Oggz-validate (should also validate opus toc)
 
* Oggz-validate (should also validate opus toc)
  
 
== Opus-tools ==
 
== Opus-tools ==
* Build infrastructure (e.g. autotools)
 
 
* A simple real time streaming example tool
 
* A simple real time streaming example tool
* Multichannel support
+
* Replaygain (half done— needs a gain tool)
* Replaygain
+
 
* Testing (incl. jenkins automation)
+
== Experiments ==
 +
 
 +
* Test exp_analysis and void_my_warranty.patch
  
== Third party software ==
+
 
* Support in ekiga
+
== Future work ==
* Support in mumble
+
* Smart automatic mode decision
* Support in asterisk
+
* psymodel based VBR
* Support in icecast
+
* Remove copy in inverse MDCT
* Support in firefox (rtcweb and in ogg)
+
* Save some float<->int conversions
* Support in VLC
+
* Improvements to LP mode CBR (greg has some code)
* Support in ogg123
 
* Support in ffmpeg
 
* Support in rockbox
 
* Support in foobar2000
 
* Support in gstreamer
 
* Support in mplayer
 
* Support in xmms
 
* Support in oggdsf
 
* Support in xiphqt
 

Revision as of 09:58, 12 March 2013

1.1-beta

  • Tune mode switching decisions
    • Use SILK up to higher rates on voice
    • Adapt stereo SILK/CELT threshold based on stereo width
  • Tune hybrid rate allocation
  • Figure out how to use speech/music detection optimally
    • find optimal switching time (low energy/tonality)
  • Improve variable frame size
  • Tune transient detector?
  • Use ALLOC in tonality analysis
  • LOTS of testing

Lower priority

  • Handle packets with PLC frames followed by FEC
  • Better handling for the case where FEC has a different bandwidth than the current mode
  • PLC transitions on unprotected SILK-SILK bandwidth changes?

Spec

Website

  • De-uglify webpage - some suggestions: write about codecs obsoleted by OPUS (Speex, CELT, Vorbis(?), and the prop. ones), write about implementations (is there only one so far?), comparison table (Opus, Vorbis, Speex, ..., MP5) of features (channels, freq, bits per sample, license, language (C89), integer impl. (Vorbis decoder only, Opus YES, ...), future use in video files (Theora? Dirac? WebM? other future codecs...), audio files for storage (like Vorbis, no raw Opus defined, only inside OGG), ...
  • Promotional material (some nice free or Public domain sounds in Opus format)

Other

  • Oggz-validate (should also validate opus toc)

Opus-tools

  • A simple real time streaming example tool
  • Replaygain (half done— needs a gain tool)

Experiments

  • Test exp_analysis and void_my_warranty.patch


Future work

  • Smart automatic mode decision
  • psymodel based VBR
  • Remove copy in inverse MDCT
  • Save some float<->int conversions
  • Improvements to LP mode CBR (greg has some code)