[vorbis] Interactive RC3 quality analysis graphs

Mercier, Dave dmercier at ea.com
Tue Jan 22 17:59:11 PST 2002



When I first tried RC3, I tried -q 0 on a somewhat mellow acoustic track
(encoded to 56 kbps), and was amazed - although you could tell it wasn't CD
quality there weren't any harshly annoying artifacts. I then tried a busy
hard rock track, and the results were drastically different (pretty
distorted). 

So it seems the results are highly dependant on the music you encoding at
quality level 0 (I'm guessing you encoded classical or something acoustic?).

I've encoded and listened to about 15 of my CD's at -q 3 now, and I'm very
impressed. Although I haven't A/B'd original vs -q 3 Vorbis, I haven't been
able to pick anything offensive out (like the crappy cymbals you can get
with lower bit rate MP3). Maybe it's just a matter of not being tuned to the
type of artifacts Vorbis produces (took me a while to realize how bad MP3
was initially), but hopefully not. The bit rate averages about 110 kbps at
-q 3. The only possible bad thing I've noticed is the high frequencies don't
seem to be very present in the scope, but my ears can't casually discern a
quality loss with an A/B.

Anyway, RC3 seems to be rock, I'm very impressed!

Thanks,
Dave.

-----Original Message-----
From: Kenneth Arnold [mailto:ken at arnoldnet.net]
Sent: Tuesday, January 22, 2002 5:42 PM
To: vorbis at xiph.org
Subject: Re: [vorbis] Interactive RC3 quality analysis graphs

<p>On Tue, Jan 22, 2002 at 08:19:09PM -0500, Monty wrote:
> People are seriously bothering to ABX/expect any kind of quality from
> -q 0?  The 'fuck it, there's no hope, just chop bits off' mode? Eek.

You mean the virtually transparent to the unskilled / casual listener
mode? I consider myself to have better than average ears, yet I really
have to listen carefully back and forth to hear the difference between
-q0 and original. -q0 Vorbis seems a little duller on the track I just
tried, but it's barely noticable, and for those who consider WMA at
48k "near-CD quality", I doubt most people would really care.

<p>--- >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.



More information about the Vorbis mailing list