[FFmpeg-trac] #8081(avcodec:new): support of g729b codec

FFmpeg trac at avcodec.org
Sun Aug 18 16:59:29 EEST 2019

#8081: support of g729b codec
             Reporter:  AlexanderU   |                    Owner:
                 Type:  enhancement  |                   Status:  new
             Priority:  wish         |                Component:  avcodec
              Version:  git-master   |               Resolution:
             Keywords:  g729         |               Blocked By:
             Blocking:               |  Reproduced by developer:  0
Analyzed by developer:  0            |

Comment (by AlexanderU):

 Replying to [comment:6 cehoyos]:
 > Replying to [comment:4 AlexanderU]:
 > > yes. Are there any specific way to confirm that?
 > My question was if you used the file that you attached here to create
 the output file (that you also attached). Afaik, the difference between
 G.729 and G.729B is that G.729B is able to send (via network) the
 information that the sender is silent which is a shorter package than an
 actual audio package, this cannot be done in a wav file, see ticket #7604.
 If your file would contain such ''2-byte Silence Insertion Descriptor''
 packets, the output file would have to be longer (but it is too short even
 if the input does not contain any Silence Insertion Descriptor packets).
 > My suspicion is that ACELP.KELVIN uses other coefficients than G.729 or
 has other small differences.

 I agree with you. ACELP.KELVIN looks like modification of g729. I don't
 know how deep modification is. Sample file contains stream with 11 bytes
 per packet. g729 has 10 bytes per packet and g729b has 12 bytes per packet
 (2 silence insertion descriptor + 10 bytes of data)

Ticket URL: <https://trac.ffmpeg.org/ticket/8081#comment:8>
FFmpeg <https://ffmpeg.org>
FFmpeg issue tracker

More information about the FFmpeg-trac mailing list