[opus] Compexity vs Quality

[Olle Frimanson]

Hi,

We measure the absolute timing of encode/decode for a OPUS mode, this is then transferred to MIPS using the CPU frequency and MIPS/Hz which comes from corresponding ARM architecture.

We used default for all OPUS settings apart from complexity  settings.

These figures are for the fixed point version, we have tested floating point as well  which had slightly higher figures on same platform.

BR/Olle

Från: Hoene [mailto:hoene at uni-tuebingen.de]
Skickat: den 15 januari 2013 11:47
Till: Olle Frimanson; opus at xiph.org
Ämne: AW: [opus] Compexity vs Quality

We will make same soon (Feb).
I would be happy to understand how you did your perform measurement and which parameters you used.

Christian


--
Universität Tübingen, Sand 13, 72076 Tübingen, Germany
Tel +49 7071 2970532, Fax +49 7071 5220
http://kn.inf.uni-tuebingen.de/staff/hoene.html

Von: opus-bounces at xiph.org<mailto:opus-bounces at xiph.org> [mailto:opus-bounces at xiph.org] Im Auftrag von Olle Frimanson
Gesendet: Dienstag, 15. Januar 2013 08:15
An: opus at xiph.org<mailto:opus at xiph.org>
Betreff: [opus] Compexity vs Quality

Hi,

The complexity settings in Opus affects the CPU load a lot, below are some preliminary figures (MIPS) we got on Samsung Galaxy S2 1.2 (1.2 GHz ARM7)

8KHz codec encode/decode - OPUS/8000 C0        56.51
8KHz codec encode/decode - OPUS/8000 C5        106.06
8KHz codec encode/decode - OPUS/8000 C10       156.55
16KHz codec encode/decode - OPUS/16000 C0       99.19
16KHz codec encode/decode - OPUS/16000 C5       196.00
16KHz codec encode/decode - OPUS/16000 C10     293.12

My question is if anybody have "real" testing on how complexity setting impact speech quality?

BR/Olle
________________________________
No virus found in this message.
Checked by AVG - www.avg.com<http://www.avg.com>
Version: 2012.0.2221 / Virus Database: 2638/5533 - Release Date: 01/14/13
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.xiph.org/pipermail/opus/attachments/20130115/fd739f88/attachment-0001.htm