Fraunhofer Q100 setting vs LAME V0, How does it compare? Options

[metal]

I currently have a subscription to eMusic, and out of curiosity using some open source mp3 utilities I wrote a short utility to dump the MP3 headers to the screen. What I found was interesting, and so I have come to (what appears to be) the best forum on the web for discussion of such matters.

First I was not happy to find that when music from eMusic is encoded with the Fraunhofer encoder, they are inserting into the id3v2 header a PRIV field. In this field is encoded everything they need to trace the download to me which includes date, time, session ID, and some other identifiers. I don't care so much unless my portable MP3 player or computer gets lost or stolen, and then if some kid uploads my music to some share site I would probably be hunted down like a dog... But thats not really why I'm making this post.

What I also found while scanning the id3v2 headers is that when eMusic encodes with Fraunhofer, they use the "VBR q100" settings. Are these settings any good? The only posts I could find about this Fraunhofer setting were these two, which state the q100 settings uses a 16 kHz cutoff. The audio graph looks not quite as good as LAME V0 graph.

These are the posts:
Maximizing VBR Quality, Squeezing out the last 2%
Question about VBR and --strictly-enforce-iso


Personally, I use LAME when I rip my CDs using the "-V0 --vbr-new -q0 --lowpass 19.7" settings. Any comments on those settings would be welcome too.

Thanks.

[greynol]

Unless you can ABX the differences what is done to the audio spectrum does not matter.

We listen with our ears, not our eyes.

[Kohlrabi]

Personally, I use LAME when I rip my CDs using the "-V0 --vbr-new -q0 --lowpass 19.7" settings. Any comments on those settings would be welcome too.

I'd avoid all the special flags and start from a much lower -V setting, like -V5. If the resulting encodes are not transparent to you, gradually decrease the -V setting. I think you might be surprised at the level that sounds transparent to you. As a safety margin you might then just choose -V one level better than the one that was transparent. It's of course "safe" (read: as good as LAME VBR encodes will get) to choose -V0 and not do any testing, which potentially takes up a lot more space than the files produced by the above mentioned method.

This post has been edited by Kohlrabi: Feb 28 2012, 21:31