[opus] Confusion on CELT or Silk

Amit Ashara ashara.amit at gmail.com
Thu Dec 31 18:44:31 PST 2015

Hello Joshua,

I understand. My main aim to first understand the processing overhead when
decisions have to be made for a fixed audio in signal, reduce it so that I
can then begin making an update to reduce the memory (on chip flash and
RAM) footprint.


On Thu, Dec 31, 2015 at 7:58 PM, Joshua Bowman <silverbacknet at gmail.com>

> Opus will always use whatever mode it feels best encodes the frame, though
> you can use opus_encoder_ctl(ctx,OPUS_SET_SIGNAL(signal)); where signal is
> either OPUS_SIGNAL_VOICE or OPUS_SIGNAL_MUSIC to influence the decision.
> SILK won't be used when picking incompatible frame sizes, but you're not
> really supposed to know or care what underlying algorithm makes the magic
> happen, since neither SILK not CELT stands alone in Opus. Simply feeding
> the correct audio type should be sufficient.
> Oh, I see your previous message that you're benchmarking. You can import
> opus_private.h and use opus_encoder_ctl(ctx,OPUS_SET_FORCE_MODE(mode));
> where mode is MODE_SILK_ONLY, MODE_CELT_ONLY, or MODE_HYBRID. Obviously,
> don't do this in a real application! The numbers will be slightly different
> due to bypassing the mode analyzer, so make sure to test the public API
> primarily.
> On 12/31/2015 5:17 PM, Amit Ashara wrote:
> Hello All,
> Based on reading the documentation and some online resources, it seems
> that there are 2 codecs CELT and SILK. However when calling opus_encode*
> API, it is not clear which of the two are used?
> Is there a specific mode to be selected to use one or the other?
> Regards
> Amit
> _______________________________________________
> opus mailing listopus at xiph.orghttp://lists.xiph.org/mailman/listinfo/opus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.xiph.org/pipermail/opus/attachments/20151231/2b95d52c/attachment.htm 

More information about the opus mailing list