Hi,
I am using the iLBC speech codec on Android. I have not changed the original
code I obtained from the RFC. I have just written JNI bridges to access the C
code. There are no exceptions during encoding and decoding. I am using 8KHz,
16bit, Mono. I am using the 30ms encoding option. Whenever I playback decoded
audio, it plays faster than the original rate it was recorded at. I have checked
the playback options of the system. It is the same as the settings I used while
recording. I have performed a test also. I recorded original PCM to a file. Then
while playing back, I recorded the decoded PCM. There is a small difference of
400 bytes in file sizes. I analyzed the waves in SoundForge, by visual
inspection, both are ALMOST similar (few differences here and there).
I am new to speech coding, someone, please help me with this problem.
Also, I thought If i resample decoded data, the rate maybe adjusted. Is this a
viable solution? If so, please send me some C code (for ARM processor) or Java
code or a link to the same.
Thanks,
Earlence
iLBC on Android
Started by ●August 14, 2009
Reply by ●August 14, 20092009-08-14
Earlence-
> I am using the iLBC speech codec on Android. I have not changed
> the original code I obtained from the RFC. I have just
> written JNI bridges to access the C code. There are no exceptions
> during encoding and decoding. I am using 8KHz,
> 16bit, Mono. I am using the 30ms encoding option. Whenever I
> playback decoded audio, it plays faster than the original
> rate it was recorded at.
This sounds like the easier problem to debug. In what file format are you storing decoded audio? Suggest to store it
to .wav format, and then try playback on a WinXP machine. Anything even slightly wrong with your basic parameters
(Fs, num channels, codec type, etc) will jump out and can be debugged easily.
-Jeff
> I have checked the playback options of the system. It is the same as the settings I used
> while recording. I have performed a test also. I recorded original PCM to a file. Then while playing back, I recorded
> the decoded PCM. There is a small difference of 400 bytes in file sizes. I analyzed the waves in SoundForge, by visual
> inspection, both are ALMOST similar (few differences here and there).
>
> I am new to speech coding, someone, please help me with this problem.
>
> Also, I thought If i resample decoded data, the rate maybe adjusted. Is this a viable solution? If so, please send me
> some C code (for ARM processor) or Java code or a link to the same.
>
> Thanks,
> Earlence
> I am using the iLBC speech codec on Android. I have not changed
> the original code I obtained from the RFC. I have just
> written JNI bridges to access the C code. There are no exceptions
> during encoding and decoding. I am using 8KHz,
> 16bit, Mono. I am using the 30ms encoding option. Whenever I
> playback decoded audio, it plays faster than the original
> rate it was recorded at.
This sounds like the easier problem to debug. In what file format are you storing decoded audio? Suggest to store it
to .wav format, and then try playback on a WinXP machine. Anything even slightly wrong with your basic parameters
(Fs, num channels, codec type, etc) will jump out and can be debugged easily.
-Jeff
> I have checked the playback options of the system. It is the same as the settings I used
> while recording. I have performed a test also. I recorded original PCM to a file. Then while playing back, I recorded
> the decoded PCM. There is a small difference of 400 bytes in file sizes. I analyzed the waves in SoundForge, by visual
> inspection, both are ALMOST similar (few differences here and there).
>
> I am new to speech coding, someone, please help me with this problem.
>
> Also, I thought If i resample decoded data, the rate maybe adjusted. Is this a viable solution? If so, please send me
> some C code (for ARM processor) or Java code or a link to the same.
>
> Thanks,
> Earlence