Re:Announcing Bluetooth LE Audio
by 0-_-0 on Tuesday January 07, 2020 @01:46PM
(#21983159)
Opus is not low complexity
[ Parent
| on HN ]
Found this:
LC3 at 32 kbps (LC3 32) provides significantly better audio quality than Opus-CELT at 32 kbps and complexity level 0 (OPUS_v114_c0 and COPUS_v114_c0).
https://www.etsi.org/deliver/etsi_tr/103500_103599/103590/01...
Re:Announcing Bluetooth LE Audio
by throw0101a on Tuesday January 07, 2020 @03:34PM
(#21984485)
I'm of the opinion that Codec 2 [0][1] should pause on trying to perform better at lower and lower bit rates, and put some work into higher ones.
Getting down to 700 and even 450 b/s is neat and all, but there seems (to me) to be a gap between 4000 and 10000 b/s in the patent-free space.
opus_256k can decode in real time on ~40MHz arm9 from 1999, or 16MHz DSP core https://www.synopsys.com/dw/ipdir.php?ds=arc_audio_opus_code...
For comparison, Synopsys' LC3 codec implementation: https://www.synopsys.com/dw/ipdir.php?ds=arc_lc3_codec
It doesn't seem to be a particularly lightweight codec compared to Opus, or particularly performant.