opus - Oct 2015 - recommended opus bitrate / opusenc setting for general?

On Sat, 2015-10-24 at 22:16 -0700, Thomas Daede wrote:
> Everything above 96kbps on that table is speculative, as the highest
> multi-participant listening testing done was at 96kbps. Here's the
> results from that test, if you're curious:
> 
> http://listening-test.coresv.net/results.htm
> 
> As you can see, at that rate Opus ranged from slightly perceptible to
> imperceptible. Also importantly, note how few of the donors were able
> to
> give significant results on most tracks. This makes a 128kbps or
> higher
> test impractical as there would not be enough listeners capable of
> hearing a difference. The HydrogenAudio wiki should probably be
> updated
> accordingly.

So basically, with 100 kB/s or say 128 kB/s one should be more or less
safe and very close to transparency, right?!


Are there any bitrates recommended (e.g. for quality or performance
reasons)? E.g. that one should use 128 kB/s rather than 127 kB/s or is
this anyway moot, since with VBR it's just a target value?

When I increase the encoding complexity to 10, this won't have any
noticeable effect on the decoding complexity, right?

Last but not least, since my purpose would be listening to music, is it
best to set the framesize to its maximum (60ms)?
And will this have any bad side effects, like on gapless playback or
so?


Thanks,
Chris.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 5313 bytes
Desc: not available
Url :
http://lists.xiph.org/pipermail/opus/attachments/20151026/1e664233/attachment.bin

Christoph Anton Mitterer <calestyo at scientia.net>
wrote:
> So basically, with 100 kB/s or say 128 kB/s one should be more or less
> safe and very close to transparency, right?!
Yes, assuming you mean kb/s (kilobits, not kilobytes, per second), and
no more than 2 channels (stereo).
>
> Are there any bitrates recommended (e.g. for quality or performance
> reasons)? E.g. that one should use 128 kB/s rather than 127 kB/s or is
> this anyway moot, since with VBR it's just a target value?
For VBR it is just a target value and the actual value will vary.
There is no need to choose any specific value; the target does not
even have to be an integer number of kb/s.
>
> When I increase the encoding complexity to 10, this won't have any
> noticeable effect on the decoding complexity, right?
10 (the maximum) is the default encoding complexity in opusenc, and
assuming you are encoding on a laptop or desktop from the past decade
there is little reason to use another value.  The complexity setting
is intended to impact the complexity of encoding only, not the
complexity of decoding.
>
> Last but not least, since my purpose would be listening to music, is it
> best to set the framesize to its maximum (60ms)?
> And will this have any bad side effects, like on gapless playback or
> so?
Sizes larger than 20ms reduce encoder flexibility and are generally
beneficial only at low bit rates, or when there is a large per-packet
overhead (e.g. when transmitting via RTP).  For your use case 20ms
(the default) is the most efficient.  Gapless playback works with any
frame size.

 - Mark

On 26/10/15 05:21, Mark Harris wrote:
>> Last but not least, since my purpose would be listening to music, is it
>> best to set the framesize to its maximum (60ms)?
>> And will this have any bad side effects, like on gapless playback or
>> so?
> 
> Sizes larger than 20ms reduce encoder flexibility and are generally
> beneficial only at low bit rates, or when there is a large per-packet
> overhead (e.g. when transmitting via RTP).  For your use case 20ms
> (the default) is the most efficient.  Gapless playback works with any
> frame size.
My experience confirms this: for 96kbps, 60ms doesn't gain anything and
the resulting file is even slightly bigger. When encoding podcasts mono,
16kbps, 60ms have a bit of filesize advantage. I can't hear the quality
difference and the size advantage is so tiny (maybe 2-3%) that bothering
with framesize is actually pointless.

That said, I would consider that encoding 60ms frames could provide more
redundancy for the encoder for more compression, but that is not
currently the case. Too bad. Maybe in a future Opus release. This sounds
like a missed opportunity.

-- 
Jes?s Cea Avi?n                         _/_/      _/_/_/        _/_/_/
jcea at jcea.es - http://www.jcea.es/     _/_/    _/_/  _/_/    _/_/  _/_/
Twitter: @jcea                        _/_/    _/_/          _/_/_/_/_/
jabber / xmpp:jcea at jabber.org  _/_/  _/_/    _/_/          _/_/  _/_/
"Things are not so easy"      _/_/  _/_/    _/_/  _/_/    _/_/  _/_/
"My name is Dump, Core Dump"   _/_/_/        _/_/_/      _/_/  _/_/
"El amor es poner tu felicidad en la felicidad de otro" - Leibniz

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
Url :
http://lists.xiph.org/pipermail/opus/attachments/20151026/0a8e9354/attachment.pgp