[opus] Opus vs AAC (endurance test)

[Jean-Marc Valin]

On 11/16/2017 06:55 PM, Gregory Maxwell wrote:
> 
> Opus is specifically designed to survive tandeming but you need to keep
> the frames aligned and not mess with the gain, which your tools probably
> do not do.

Actually, what happens is that if you use 44.1 kHz, the resampling in
the decoder causes a fractional-sample delay in opusdec. There's a fix
in master, but I'm not sure it's been committed.

I actually ran the test a while ago and that delay makes a huge
difference in the cascading performance because of the alignment.

	Jean-Marc

> On Thu, Nov 16, 2017 at 10:58 PM, encrupted anonymous
> <sergeinakamoto at gmail.com <mailto:sergeinakamoto at gmail.com>> wrote:
> 
>     using iTunes i've noticed that AAC is
>     very good at re-encoding own lossy sound.
>     let's test Opus!
> 
>     neroaacenc.exe -q 0.75 -if 000.wav -of 001.m4a
>     neroaacdec.exe -if 001.m4a -of aac001.wav
>     wavdiff.exe 000.wav aac001.wav
>     Comparing 000.wav - aac001.wav...
>     Max diff: -17.3867dB
>     RMS diff: -33.0851dB
>     Mean diff: -32.4582dB
> 
>     opusenc.exe --bitrate 512 "000.wav" 001.opus
>     opusdec.exe 001.opus opus001.wav
>     wavdiff 000.wav opus001.wav
>     Comparing 000.wav - opus001.wav...
>     Max diff: -22.5646dB
>     RMS diff: -39.0425dB
>     Mean diff: -38.7372dB
> 
>     Opus @482kbps is much better than AAC @288kbps.
>     now let's see what happens after 10 passes
> 
>     Comparing 000.wav - aac010.wav...
>     Max diff: -16.1286dB
>     RMS diff: -32.3361dB
>     Mean diff: -31.715dB
> 
>     AAC stepped back just a little
> 
>     Comparing 000.wav - opus010.wav...
>     Max diff: -7.61385dB
>     RMS diff: -20.3666dB
>     Mean diff: -20.1286dB
> 
>     Opus made complete disaster, but HF looks
>     good (so it isn't an error in frames
>     synchronisation of wavdiff program).
> 
>     And no, i'm not going to do ABX test at high
>     bitrates because i'm not wavdiff.exe, and i'm
>     not going to do ABX tests at low bitrates
>     because i don not use them.
> 
>     FLAC is good, but MIDI is better. Lossy codecs
>     "reverse-engeneering" audio back to MIDI, that's
>     what they do :)
> 
> 
>     <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
>     	Virus-free. www.avast.com
>     <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
> 
> 
>     <#m_1631935980190637709_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> 
>     _______________________________________________
>     opus mailing list
>     opus at xiph.org <mailto:opus at xiph.org>
>     http://lists.xiph.org/mailman/listinfo/opus
>     <http://lists.xiph.org/mailman/listinfo/opus>
> 
> 
> 
> 
> _______________________________________________
> opus mailing list
> opus at xiph.org
> http://lists.xiph.org/mailman/listinfo/opus
>