3 ms·
I'm glad to see the SIG improving the baseline audio codec. SBC is/was a weird little codec. It's like a dumbed-down MPEG-1 Layer-1 audio, but with only 8 sub-
by prutschman 7y ago
I'm glad to see the SIG improving the baseline audio codec. SBC is/was a weird little codec.
It's like a dumbed-down MPEG-1 Layer-1 audio, but with only 8 sub-bands. Like layer-1, the filterbank isn't perfect reconstruction, so the encode/decode cycle causes aliasing noise even before any quantization takes place.
(At a previous job I had to make a proprietary extension that improved the stop-band attenuation of the prototype filter, otherwise the speaker companies didn't even want to talk to us, the THD+N looked so bad.)
The only psychoacoustic model is an optional static biasing of which sub-bands gets bits allocated.
On the other hand, you can encode/decode entirely in software on a 24 MHZ arm7tdmi. (arm7 is ARMv3, not ARMv7, confusingly).