which is desined to handle both AMR(NB) and G711 codecs. I am using SDL to provide us with access to the audio hardware(have also got microphone input working using a patch I recvd from this news group). The VoIP app currrenly receives the AMR or G711 RTP frames (via RTP/UDP), each rtp payload contains 6 AMR encoded speech frames at AMR encode mode 7 i.e each speech frame is 31 bytes. Each speech frame represents 20ms of speech. When decoded to PCM is results in 160 bytes per frame. I now need to present this PCM data to the audio output callback but am wondering how to do this with regards to the frequency that it is expecting data? I have currently opened the audio device with a 30250 frequency as this seems to work best with the microphone but my AMR was recorded at 8000, or should I just feed the audio callback with as much data as it asks for when is ask for it?
Currently I am just dumping the 6 speech frames (160*6=960 bytes) into the audio callback buffer as they arrive but the playback sounds far too quick but I can just make out the actual speech so I know the decode is working.
Sorry if my q’s seem obvious but I have only really used SDL before to playback sound data from supported formats like .wav
Any help is appreciated and many thanks for a great product.
Regards,
-Chris Dobbs~
Sorry but please ignore my last message as our useless mail router seems
to have buggered it up and chopped half of it out!
All,
I am currently working on a project to design a cross-platform (linux
and WIN32) VoIP API for both audio and video. I am planning to handle
both AMR(NB) and G711 audio codecs and am using SDL to provide us with
access to the audio hardware(have also got microphone input working
using a patch I recvd from this news group). The VoIP app currrenly
receives the AMR or G711 RTP frames (via RTP/UDP), each rtp payload
contains 6 AMR encoded speech frames at AMR encode mode 7 i.e each
speech frame is 31 bytes. Each speech frame represents 20ms of speech.
When decoded to PCM is results in 320 bytes per frame (160*2). I now
need to present this PCM data to the audio output callback but am
wondering how to do this with regards to the frequency that it is
expecting data? I have currently opened the audio device with a 30250
frequency as this seems to work best with the microphone but my AMR was
recorded at 8000, or should I just feed the audio callback with as much
data as it asks for when is ask for it?
Currently I am just dumping the 6 speech frames ((160*6)*2= 1920 bytes)
into the audio callback buffer as they arrive but the playback sounds
far too quick but I can just make out the actual speech so I know the
decode is working.
Sorry if my q’s seem obvious but I have only really used SDL before to
playback sound data from supported formats like .wav Any help is
appreciated and many thanks for a great product.
Regards,
-Chris Dobbs