[vorbis] MP3/Vorbis comparision

[Giuliano Pochini]

On 11-Feb-2003 Frank Grotelueschen wrote:
> Giuliano Pochini wrote:
>>High frequency boost in mono files is quite low. Can you try the same
>>test with stereo samples ? I think the results will not be same as good.
> 
> I have repeated the test with stereo white noise. Both channels are
> totaly independent from each other. [...]
> The stereo ogg-files have nearly the same high frequencies boost as
> the mono ogg-files.
> The frequencies response off all mp3-files is total flat.

I had time to do a quick test too. I tried to encode mono and stereo
white noise at various quality level. I used audacity to do the
spectral analisys I don't know how much reliable it is. Stereo samples
are a bit different from mono ones. At -q3 I can see a noticeable steep
rise of about 2dB at ~6KHz. At -q5 it is centered over 10KHz and from
-q6 I can't see it anymore. That means lossy stereo coupling has
something to do with high frequency boost. I had no time to test all
quality levels.
There is another fact that can be quite annoiyng. My noise samples
were at half amplitude (-16K/+16K). After I encoded/decoded them,
the amplitude was a lot higher than that. I can see it in the background
window of your spectrograms too. It happens up to -q6. IMHO it's caused
by the way vorbis encodes (lossy) phase informations.

> Filesize is around 10 MB.
> Because i have only 10 MB space for storing, it was necessary to delete
> the mono-files before uploading the stereo-files, sorry.

Well, noise is noise... the only interesting thing are the graphs :)

<p>Bye.

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.