Which LAME VBR settings do you use?, Poll regarding LAME VBR preferences Options

[quackalist]

Voted V0, 3.98, not that I imagine it betters V2 which I used to use other then rarely but storage isn't at a premium now so I allow a large 'comfort zone' just because.

Fact is, I have a lot of FLAC's on my HD too, and if I have any temporary storage issues I recode a lot to V0 while I ponder another HD

This post has been edited by quackalist: Nov 7 2008, 02:05

[Kitsuned]

My two cents buys a vote for:

Lame 3.98
-V3 setting (great for portables given gbs of space)
no floating points (never saw the need)

[ajfoucault]

3.98.2
V0
Real me do this!
JK~!

[antman]

Lame 3.98.2/V5.7

[somemightsay]

3.98.2

-V0

[xSerpentx]

V4
LAME 3.98

I'm actually re-ripping. I want to save space and I noticed from reading here so much that my current files (V0) is way overkill and I will mostly be listening to these V4 mp3's on my iPhone or iPod classic.

Before you guys say anything...I'm really not interested in lossless.

[Frankie]

Mostly V4 for use with my iAudio F2. Up to V0 for other purposes.
3.98.2
Floating point? Never used it. But still I think it's a good thing that we are given the choice to use it.....

[rick.hughes]

I play my FLAC archive on my home stereo, and only use mp3 on my portable. Since it only has 12GB and is primarily used in the car I use -V6 by default. If I ever notice any problem I will just encode that one song again rather than use more space than needed on all the others.

[trout]

-V 1.337

[skope]

Hey everyone! I stick to using 3.90.3modified --alt-preset standard. It's slow and probably outpeformed qualitywise, still it feels solid to me. It's like a tank.

EDIT: Oh, I forgot to mention that I play my mp3's in my iPod Classic 80GB with KOSS Porta Pro's. I encode mainly hip-hop music which usually gives me pretty good file sizes on the V2 (standard) setting. At home I play the CD's in my stereo.

This post has been edited by skope: May 9 2009, 22:22

[PaJaRo]

When I transcode from a lossy source (I know it is not a good practice, but I do it for portability to my DAP) I usually use -V 4.
When I transcode from a lossless source, again to listening the music in a DAP, I use -V 3.
I mostly use flac on my computer (since hard disk space is not a problem).

I'm using lame 3.98.2 I've never used floating point, but it could be a good idea.

This post has been edited by PaJaRo: May 10 2009, 19:15

[Pike84]

What is the practical significance of using floating point values in this case? I tried searching, but there seems to be very little information on the subject .

This post has been edited by Pike84: May 12 2009, 09:40

[halb27]

IMO the advantage of non-integer -V values is with listening tests. You can adjust quality setting to achieve an average bitrate which gives a fair comparison with the contenders.
So Lame 3.98 -V5.7 was used with last fall's 128 kbps mp3 test. I used -V1.5 (--lowpass 16.7) for a fair comparison with dBpoweramp's FhG encoder @ CBR 192 in a private listening test. (As I was content with the results I sticked with -V1.5 though difference in quality and bitrate towards -V2 or -V1 is expected to be negligible).

This post has been edited by halb27: May 12 2009, 11:30

[Pike84]

So, it's just about being able to use decimal numbers, in addition to integers? It's not like decimals couldn't be used with fixed-point values?

This post has been edited by Pike84: May 13 2009, 09:04

[halb27]

So, it's just about being able to use decimal numbers, in addition to integers? It's not like decimals couldn't be used with fixed-point values?

I don't know if I understand you correctly. If it's about the number of digits after the decimal point: it's possible to use more than 1 digit though I can't imagine anybody has the need for it.

[Pike84]

Well, I'm not sure if I understand this floating point stuff very well myself, and I guess computers using binaries has something to do with it...

Anyway, what I meant was that is it necessary to use floating point values to be able to use any non-integer value for the quality setting of LAME? I guess I'm being a bit lazy for not wanting to read a lot about the floating point implementations with computers and LAME here, but I was hoping that someone could dumb it down for me .

[halb27]

It's not about implementation, just about the -V value given on the command line.

Guess it wasn't a good idea refering to 'floating point' values at all. In our context of -V value 'floating point value' should read 'non-integer value'.

This post has been edited by halb27: May 13 2009, 15:35

[Pike84]

Oh, ok. I thought it had some deeper meaning behind it, but I guess not .

[ZinCh]

Sometimes I am using floating point to get target size.

[gasmann]

It looks like I'm stuck with LAME 3.98.2 -V 5 for my portable. Maybe I could go even lower (I've never tried) because -V 5 is generally transparent for me. I'm using MP3 only for my old portable having 256 MB. But it is of high quality (hell, it was expensive!) and I'm fine with 4 hours of music, that's enough for me (apart from that it's not very practical to have a lot of music on it because it has no folder navigation, you have to skip through all songs to find what you want).

This post has been edited by gasmann: May 18 2009, 19:39

[d_headshot]

What are floating point values?

[WonderSlug]

What are floating point values?

Numbers with decimal values


Integers --> 1, 2, 3, 128, 1353698 and so on


Floating Point --> 1.6, 12.1, 313.453, 895.3237952

This post has been edited by WonderSlug: May 20 2009, 04:42

[lock67ca]

Most of my collection is ripped with 3.97b at vbr new but I've been using 3.98 at -v2 since it was released. I may try using -v5, since I'd like to get more stuff crammed onto my portables with as little quality loss as possible.

[GeneV]

I went through a couple of test with a few pop/rock songs, comparing ABR 160 and 192 vs. VBR at similar bitrates.
I made ABX-Tests and the outcome is that ABR clearly sounds better.
It should be mentioned that I always use the -q 0 switch, which increases encoding time but seems to deliver significantly better quality.

I also compared different ABR bitrates. The optimum seems to be 192. I found that 224 or more do not produce audible improvements,
while 160 is only slightly less in qualitiy and 128 a tolerable value, if size matters at least as much as quality.

So my default encoding options are: --abr 192 -q 0

Peace
Gene

[halb27]

... I made ABX-Tests and the outcome is that ABR clearly sounds better. ...

Do you use 3.98?
I've been an ABR advocate for a long time before 3.98 came out (maybe I still am at very high bitrate ABR 250...300, but it's not very relevant to me any more).
With 3.98 however I can hardly imagine samples where ABR is clearly better than VBR in your bitrate range though ABR certainly is fine too.