I'm playing with using amr-wb for podcast recordings on my android. I'm really size constrained, and it seems amr-wb does a better job than opus in the 12 kbps range. But the same as opus ~ 16 kbps. See:
http://research.nokia.com/files/public/%5B..._Opus_Codec.pdf (http://research.nokia.com/files/public/%5B16%5D_InterSpeech2011_Voice_Quality_Characterization_of_IETF_Opus_Codec.pdf)
I'll be using this over the car radio, so a relatively noisy environment.
The Library of Congress uses amr-wb for some of its sound recordings. In its description of the codec:
The available bitrates are: 6.60, 8.85, 12.65, 14.25, 15.85, 18.25, 19.85, 23.05 and 23.85 kbps. The lowest rate that provides excellent speech quality in a clean environtment is 12.65 kbps. Higher rates are useful in background noise conditions......
http://www.digitalpreservation.gov/formats...fdd000255.shtml (http://www.digitalpreservation.gov/formats/fdd/fdd000255.shtml)
Now if I need to go up to the 16 kbps range, I might as well use opus, which I use for all my music. But does it make sense that a higher bitrate makes any difference in noisy conditions ?? If I have excellent speech quality at a low noise level, don't I have excellent speech quality at a high noise level? I may need to increase the volume to compete with the noise, but the speech quality hasn't changed. And a higher bitrate doesn't increase the volume. Correct?
That refers to the noise of the recording, not the noise of the playback environment.
Ah. So if I'm transcoding, then 12.65 kbps will work? Great.
Thanks