I have ripped 20+ CDs with -V 0 --vbr-new (LAME 3.97) and I have only now realized, that the setting -q5, I was using some time in the past was still turned on in my settings. I have tried to ABX the mp3s I've got with these awkward settings, and I can't tell the difference, however I would like some opinions, how bad have I screwed my mp3s up? I don't like the idea of reripping... I know, if I can't ABX it, why worry. But I can't help a feeling, that something is sub-par now. I need some reassurance. What does -q5 at a high setting, such as -V 0 do anyway?
Heck, I tested -q9 with -V 0 --vbr-new now and I still can't ABX it...

As a sidenote... the music is of extreme complexity as I get bitrate of 240 kbps average with -V 2 --vbr-new instead of 190 kbps average (and I know VBR is unpredictable...).

Some technical explaination of what is happening with such idiotic combination of settings would be very much appreciated!

Most members here in the forum would either tell you: "read the wiki, the wiki is there for a purpose", "leave the -Vx presets alone" or "other-statements-that-would-make-you-feel-stupid."

IIRC, the default setting for the -V presets in LAME 3.97 is -q3. Turning to high quality mode is -q2. Although you can't ABX at -q9, if you still bit feel paranoid, you can turn it all the way up to -q0 (the highest -q setting). This might give you peace of mind, and possibly shave off bits compared to the default settings.

LAME 3.98 beta versions render the -q switch unoperable, however.

A q value of 4 and above has no effect with vbr-new, so changes within that range that won't give you any higher quality.

Using a q value of 5 apparently changes the noise shape amplification (you can see the difference in settings with the --verbose switch for LAME). I don't know in what degree that affects the quality, but if you haven't noticed it yet, you're probably fine. No need to re-rip IMO.

I think as long as the mp3s are un-ABXable from the soundtracks on the original CDs, you're good to go.

...or simply make you wait a whole lot longer.

I tested a while back some very old lame versions and -q2 was no great improvement over q5. At 128k q7 is clearly inferior to the default using lame 3.97 / 3.98..

I don't think you have much to worry about but I would be annoyed too. Which frontend gave you q5 ?

Thanks for all your comments! I know the main thing here is that the mp3 are not ABX-able compared to the original. I just needed some peace of mind. shadowking comforted me the most, saying that -q2 is not so much better than -q5 . I'll just leave those mp3s as they are then.

Frontend I used was EAC with lame.exe. And I had put the -q5 switch at the end last time I was testing something at 112 kbps and playing arround with the settings. Because one can't see the whole command line and my -q5 was at the end, I have just changed my otherwise -V 4 --vbr-new to -V 0 --vbr-new (because of the complexity of the files I have encoded, i.e. 240 kbps average at -V 2 --vbr-new) and completely forgot about -q5 at the end. Actually I forgot about it last time I was playing around, so you can imagine that if I didn't remember then, that now I have really completely forgotten about it, until I checked my mp3s with EncSpot yesterday...

Oh, and as a curiosity... those files, that come as 240 kbps average at -V 2 --vbr-new, have an average of "only" 270 kbps at -V 0 --vbr-new and curiously low 170 kbps average at -V 3 --vbr-new. Strange samples for lame, hehe.

Maybe I should start a new topic for my next question, but I have decided that I wont...
Is the folowing true for VBR also?

-q 0: use slowest & best possible version of all algorithms.

-q 2: recommended. Same as -h. -q 0 and -q 1 are slow and may not produce
significantly higher quality.

-q 5: default value. Good speed, reasonable quality

-q 7: same as -f. Very fast, ok quality. (psycho acoustics are
used for pre-echo & M/S, but no noise shaping is done.

-q 9: disables almost all algorithms including psy-model. poor quality.

Thank you for a nice discusion, that will hopefully evolve.

EDIT: Typos.

In 3.98 q7~9 are mapped to q6 for vbr.. A good thing.

V2~V0 are affected by sbf21 issue that is why the bitrate is bloated with hf heavy music.

wait for 3.98 final and re-rip

Thanks, but I'm already at piece with those mp3s. I would rerip if I would accidentally use -q9 with -V 0 --vbr-new, but with -q5 I'm OK now. I thought it was worse than it is.

I'm glad to hear that 3.98 ignores -q switch with -V. This means that after 3.98 this can never happen to me again.