Help - Search - Members - Calendar
Full Version: Lame 3.97, -q3 vs -q2
Hydrogenaudio Forums > Lossy Audio Compression > MP3 > MP3 - General
NeoRenegade
I understand that now -q3 is used rather than -q2 in the default settings (used to be the alt presets).

Is -q3 used because it has shown to be better quality, or is it just used because it's deemed "as good as" -q2 but faster?
Yaztromo
QUOTE(NeoRenegade @ Oct 23 2005, 10:41 PM)
I understand that now -q3 is used rather than -q2 in the default settings (used to be the alt presets).

Is -q3 used because it has shown to be better quality, or is it just used because it's deemed "as good as" -q2 but faster?
*



I encoded two files. One with "-V2 --vbr-new" and one with "-V2 --vbr-new -q2". Whilst LAME did report different Q values, they were bit for bit identical huh.gif

QUOTE
INFO (foo_bitcompare) : Comparing:
INFO (foo_bitcompare) : location: "file://C:\Home\Incoming\testnoqchange.mp3" (0)
INFO (foo_bitcompare) : location: "file://C:\Home\Incoming\testq2.mp3" (0)
INFO (foo_bitcompare) : No differences in decoded data found.
INFO (foo_bitcompare) : Finished successfully.
DreamTactix291
I believe --vbr-new only has 3 real -q values based on what Gabriel has said.

-q0 through -q4 are bit identical
-q5
-q6 through -q9 are bit identical

Now if I'm wrong I'm sure Gabriel or one of the other LAME devs will correct me, but I'm pretty sure I read what he posted once correctly.
rutra80
It was me [:
Anyway, NeoRenegade isn't explicitly asking about --vbr-new, and in old VBR mode there still are ten -q levels.
To answer his question, I'm not aware of any listening tests showing that lower -q values produce higher quality files (thogh of course it is designed to produce higher quality files), so I guess that -q3 being default is a choosen compromise between (unperceptably) higher quality & speed - LAME devs probably decided that -q value any lower than 3 would slow things down too much without giving noticably better quality.
DreamTactix291
Actually I believe current --vbr-old's -q3 is the same as the old -q2. I believe the numbers were bumped down when the experimental and super slow -q0 was implemented at some point after the 3.90 branch. Again I could be wrong but I'm sure I've read that here too.
NeoRenegade
Thanks for the good replies so far.

To elaborate on my original post, I'm messing around with CBR. I'm guessing this is the same as --vbr-old, with reference to -q values and their application?
DreamTactix291
AFAIK -q values are the same for --vbr-old as for CBR. --vbr-old reuses something from the CBR models that --vbr-new doesn't but I honestly can't remember what it was, something to do with the quantisation step I believe though I may be way off on this as I can't claim to understand the internals of the LAME source all that well.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2008 Invision Power Services, Inc.