[vorbis] as if you haven't seen this yet

[Ross Levis]

On a similar note, I heard a good way of testing the transparency of a codec
is to encode a song, decode it to WAV, re-encode the decoded WAV etc about
10 times.  It highlights the imperfections very well.  I tried this some
time ago with LAME using CBR & VBR, & VBR was much worse.  Haven't tried it
with vorbis.

Ross.

Patrick McLean wrote:
> I'm not on the list, so cc any replies directly to my address please.
> 
> I saw this discussion reading the archives, and thought I 
> would mention an
> idea I saw in the discussion on slashdot. Someone suggested 
> that a good way
> to do a test of lossy formats would be to take about  3 or 4 
> clips of around
> a minute each, encode them to vorbis, mp3, mp3pro, wma and 
> whatever else you
> want to test, then decode them back to .wav files and post them for
> download, not giving any clue as to which format they were 
> originally in
> (like name them clip1.wav, clip2.wav) then have people 
> download and evaluate
> them (providing the original .wav for a reference point, of course)
> 
> Once a preset number of people have evaluated them, and sent 
> back their
> comments on each (mabe have a form that they can fill out) 
> then release teh
> results telling which file was in what format.
> 
> I think this would probably be the best way to objectively 
> test all the
> available lossey sound formats, and I think that testing with 
> as many people
> as possible is probably preferrable, and more accurate to 
> testing with just
> a few so-called experts.
> 
> I might do this sometime if I have some free time and I 
> figure out what
> tracks would be appropiate to use, they would probably have 
> to be either
> non-copyrighted or I would have to have to get permission from the
> copywright owner (unless 1 or 2 minute clips would be covered 
> under fair
> use, I don't know)
> 
> 
> --- >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.
> 

--- >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.