IPB

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Closed TopicStart new topic
From FLAC to v2 VBR - 2 Issues, Basic encoding advice
IDefyAxioms
post Apr 21 2013, 11:25
Post #1





Group: Members
Posts: 2
Joined: 21-April 13
Member No.: 107771



Hey all,

I'm using Foobar2000 + LAME 3.99.2.5 to convert my FLAC files to mp3. Currently what I do is load the desired files into Foobar, then Convert -> 190kbps v2 VBR, and convert. My first question is: should I be doing anything else to get the most out of what I'm encoding? Are there any presets that would make the files "better" while keeping a relatively low file size?

As for my second issue: Is v2 the best for me? I haven't done any ABX listening tests, and though the wiki suggests v6-v4, I've gathered elsewhere (head.fi) that v2 might actually be the best of both worlds for quality/low file size. Should I be using something different?

Thanks!
Go to the top of the page
+Quote Post
bernhold
post Apr 21 2013, 11:58
Post #2





Group: Members
Posts: 41
Joined: 22-March 13
Member No.: 107330



Question 1:

Not really, that should be the best way. You could have a look at the quality setting of LAME (the -q switch), if I remember correctly, it defaults to -q 2 which is not the highest quality setting (the best is -q 0), so you could set it to higher quality and slower encoding. However, while this sounds promising, the highest settings have little to no effect on the perceived quality, at least that's what I read and observed myself. For general purpose encoding, to my knowledge, the -q switch is the only option to tweak quality (other than the presets and bitrate distrubution modes). Check out http://lame.cvs.sourceforge.net/viewvc/lame/lame/USAGE for the additional settings.

Here's an excerpt from the LAME manual:

QUOTE
=======================================================================
algorithm quality selection
=======================================================================
-q n

Bitrate is of course the main influence on quality. The higher the
bitrate, the higher the quality. But for a given bitrate,
we have a choice of algorithms to determine the best
scalefactors and huffman encoding (noise shaping).

-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.


Question 2:

-V2 should be perceptually transparent in most cases. The question is, what's your equipment? Are you encoding for a mobile music player? Then V2 is more than enough, even overkill I'd say. If you are listening at home with hi-fi equipment, V2 is still perceptually transparent for most people and most situations, but if you're paranoid, go for the best possible quality (CBR 320kbps), however that kind of defeats the purpose of compression because you don't get a lot of bang for your buck with that mode. On my laptop with $60 headphones, even -V5 is transparent to me. But again, it depends on your equipment, the type of music and your ears. I strongly recommend to do an ABX listening test (search for the foo_abx plugin).

This post has been edited by bernhold: Apr 21 2013, 12:13
Go to the top of the page
+Quote Post
o-l-a-v
post Apr 21 2013, 12:08
Post #3





Group: Members
Posts: 123
Joined: 18-July 12
From: Norway
Member No.: 101545



LAME is said to be transparent for most peaople at 192kbps
http://wiki.hydrogenaudio.org/index.php?title=Transparency

But only you can tell what sounds transparent to you. ABX is the best way to do that.

For encoding options regarding LAME, see here:
http://lame.cvs.sourceforge.net/viewvc/lam...ml/switchs.html

There are other formats that will preserve audio quality at lower bitrates than LAME/MP3. AAC for instance, is said to be transparent for most people at 150kbps.
QAAC is considered one of the best atm. You should download TAudioConverter and start testing for yourself. It's a program that makes it much easier to switch between settings compared to Foobar + it has all those codecs included: http://sourceforge.net/projects/taudioconverter/

There is also a new Lame alpha, which is said to fix some artifacts/bugs lame 3.99 (and earlier) has. Try it: LAME 3.100.alpha2 / LAME 3.100.alpha2 64bit
Go to the top of the page
+Quote Post
db1989
post Apr 21 2013, 12:25
Post #4





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



“Basic encoding advice” is not the province of MP3 Tech.

QUOTE
Are there any presets that would make the files "better" while keeping a relatively low file size?

As for my second issue: Is v2 the best for me? I haven't done any ABX listening tests, and though the wiki suggests v6-v4, I've gathered elsewhere (head.fi) that v2 might actually be the best of both worlds for quality/low file size. Should I be using something different?

In fact, both questions basically come down to the usual answer: We don’t know. You need to run a personal double-blind test to assess what is or is not perceptually transparent to you. Any other advice is based upon generalised conclusions or even assumptions. From there, you’re the one who must choose whether to assume that your auditory system is roughly equivalent to those of the ‘average’ participant in any given listening test – or to decide that you want more confirmation of the settings you use, perhaps to save space if something further down the scale than -V2 could still provide transparency.

But I could add in response to the first question that (i) LAME’s default settings are default for a reason, and there there is almost never any need to change them unless you have a technically valid reason, and (ii) talk of “presets” is unnecessary as there are no longer any one-switch routes to optimised settings; it has been many years since the --alt-presets and --presets were all folded into mere aliases to normal parameters or combinations thereof.
Go to the top of the page
+Quote Post
lvqcl
post Apr 21 2013, 13:15
Post #5





Group: Developer
Posts: 3325
Joined: 2-December 07
Member No.: 49183



QUOTE (bernhold @ Apr 21 2013, 14:58) *
Not really, that should be the best way. You could have a look at the quality setting of LAME (the -q switch), if I remember correctly, it defaults to -q 2 which is not the highest quality setting (the best is -q 0)

VBR encoding defaults to -q 0.

QUOTE (bernhold @ Apr 21 2013, 14:58) *
Check out http://lame.cvs.sourceforge.net/viewvc/lame/lame/USAGE for the additional settings.

It's very outdated.
Go to the top of the page
+Quote Post
db1989
post Apr 21 2013, 13:18
Post #6





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



QUOTE (lvqcl @ Apr 21 2013, 13:15) *
QUOTE (bernhold @ Apr 21 2013, 14:58) *
if I remember correctly, it defaults to -q 2 which is not the highest quality setting (the best is -q 0)
VBR encoding defaults to -q 0.
And CBR to -q3.
Go to the top of the page
+Quote Post
greynol
post Apr 21 2013, 18:06
Post #7





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (bernhold @ Apr 21 2013, 03:58) *
it depends on your equipment

Do you have any ability to support or elaborate on this that is based on hard evidence?


--------------------
YOUR EYES CANNOT HEAR!!!!!!!!!!!
Go to the top of the page
+Quote Post
bernhold
post Apr 21 2013, 19:40
Post #8





Group: Members
Posts: 41
Joined: 22-March 13
Member No.: 107330



If you have low quality equipment or environmental interference, you are less likely to notice compression artifacts, so you can encode with a higher compression ratio.
Go to the top of the page
+Quote Post
greynol
post Apr 21 2013, 19:46
Post #9





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (bernhold @ Apr 21 2013, 11:40) *
low quality equipment

Repeating yourself does not constitiute evidence! I'll ask again, do you have any evidence to support this all-too-often parroted assertion?

This post has been edited by greynol: Apr 21 2013, 19:50


--------------------
YOUR EYES CANNOT HEAR!!!!!!!!!!!
Go to the top of the page
+Quote Post
bernhold
post Apr 21 2013, 20:22
Post #10





Group: Members
Posts: 41
Joined: 22-March 13
Member No.: 107330



No, I don't have any evidence as I think this claim is trivial. For example, a single speaker will not be able to reproduce stereo sound, and a laptop speaker will not be able to reproduce low frequencies. In both cases you could remove that information in the compression process to save space without actually hearing a difference.
Go to the top of the page
+Quote Post
greynol
post Apr 21 2013, 20:47
Post #11





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



As soon as you're ready to discuss the effects of masking the assumptions made about it by the encoder and how it can be broken let me know.

This includes taking a deeper look into your crude examples, as they too do not make your case.

This post has been edited by greynol: Apr 21 2013, 20:49


--------------------
YOUR EYES CANNOT HEAR!!!!!!!!!!!
Go to the top of the page
+Quote Post
bernhold
post Apr 21 2013, 21:28
Post #12





Group: Members
Posts: 41
Joined: 22-March 13
Member No.: 107330



What I said is based on my own listening tests and my own equipment. My examples are not crude, they're just simple because simple examples are easier to explain. Are we going to debate the fact that low quality equipment isn't able to reproduce the same frequency range as high quality equipment?
Go to the top of the page
+Quote Post
IDefyAxioms
post Apr 21 2013, 21:36
Post #13





Group: Members
Posts: 2
Joined: 21-April 13
Member No.: 107771



QUOTE (db1989 @ Apr 21 2013, 03:25) *
“Basic encoding advice” is not the province of MP3 Tech.

In fact, both questions basically come down to the usual answer: We don’t know. You need to run a personal double-blind test to assess what is or is not perceptually transparent to you. Any other advice is based upon generalised conclusions or even assumptions.

But I could add in response to the first question that (i) LAME’s default settings are default for a reason, and there there is almost never any need to change them unless you have a technically valid reason, and (ii) talk of “presets” is unnecessary as there are no longer any one-switch routes to optimised settings; it has been many years since the --alt-presets and --presets were all folded into mere aliases to normal parameters or combinations thereof.

First, sorry about my improper post placement. I was using the (now obviously incorrect) logic of "I'm using mp3, so I'll just put it here." Again, sorry.

Thanks for the reinforcement of the testing. I was going to hold off on it for a while (just shameful laziness, to be truthful), and figured that v2 would be more of a subjective "safe" bet - for now at least.

As for settings, I was more or less talking about the extra settings under Foobar2k's "processing" options - stereo configs/conversions, EQ, PPHS, et al. Being portable I would've assumed extra settings wouldn't truly be necessary, given that (as you've stated and as I've read before) presets are generally the best bet.
Go to the top of the page
+Quote Post
lvqcl
post Apr 21 2013, 21:37
Post #14





Group: Developer
Posts: 3325
Joined: 2-December 07
Member No.: 49183



QUOTE (guruboolez @ Nov 19 2005, 22:43) *
QUOTE (Halcyon @ Nov 19 2005, 07:16 PM)
I've noticed myself, that just upgrading headphones made previously completely unaudible distortions audible to me.

If I replace my Beyerdynamic with a set of cheap erbuds (Sennheiser MX550), several distortion cease to be audible (but not pre-echo, still perceptible). But I'm not fully convinced that a very expensive of headphone (like Stax ones) would really really help me to catch additionnal encoding issues.

Go to the top of the page
+Quote Post
greynol
post Apr 21 2013, 22:39
Post #15





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (bernhold @ Apr 21 2013, 13:28) *
Are we going to debate the fact that low quality equipment isn't able to reproduce the same frequency range as high quality equipment?

Sigh. Again(!), you are failing to address masking, the assumptions the encoder makes about it and how is can be broken. Frequency response falls well short of telling the whole story. If it did TOS8 would likely read quite differently.

I recommend you read some of the discussions about it that exist on the forum.

This post has been edited by greynol: Apr 21 2013, 23:08


--------------------
YOUR EYES CANNOT HEAR!!!!!!!!!!!
Go to the top of the page
+Quote Post
mjb2006
post Apr 21 2013, 22:42
Post #16





Group: Members
Posts: 755
Joined: 12-May 06
From: Colorado, USA
Member No.: 30694



"I hear a difference" when playing music through different kinds of equipment should really not be a controversial statement.

The problem lies in the unsubstantiated assumptions:
"...and one is higher quality than the other"
"...and this is due to encoding artifacts"
"...therefore these other encoding settings are better / should make the problem go away"

This post has been edited by mjb2006: Apr 21 2013, 22:44
Go to the top of the page
+Quote Post
Heliologue
post Apr 22 2013, 03:15
Post #17





Group: Members
Posts: 103
Joined: 17-September 06
Member No.: 35303



QUOTE (mjb2006 @ Apr 21 2013, 16:42) *
"I hear a difference" when playing music through different kinds of equipment should really not be a controversial statement.


Oh, but it is: first it has to be proven that one hears the difference at all, irrespective of the assumed cause. Only way to do that is with double-blind testing, but believing something is much easier than measuring something; ergo, TOS8.
Go to the top of the page
+Quote Post
greynol
post Apr 22 2013, 03:39
Post #18





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



We get the expensive equipment better reveals lossy artifacts all the time. This is not universally true and those who suggest that it is never offer support, generally because they don't think they have to and/or don't understand that uneven frequency response can reveal artifacts that may otherwise remain hidden with a less uneven frequency response.

That speakers and headphones may/can sound different from one another isn't the issue.

We already have open discussions on the matter so they needn't be had here. Just know that if you parrot the claim without finesse you'll get "greynol'd" if I read it.

This post has been edited by greynol: Apr 22 2013, 04:30


--------------------
YOUR EYES CANNOT HEAR!!!!!!!!!!!
Go to the top of the page
+Quote Post
mjb2006
post Apr 22 2013, 05:25
Post #19





Group: Members
Posts: 755
Joined: 12-May 06
From: Colorado, USA
Member No.: 30694



QUOTE (Heliologue @ Apr 21 2013, 20:15) *
Oh, but it is


I have to backpedal; of course you're right, though I didn't say quite what I meant.

Some claims are meaningless if not substantiated by double-blind testing. Does TOS#8 encompass all such claims, or only some of them? I think only some of them.

Claiming a difference might be heard, given adjustment of some parameter (e.g., speakers, Monster Cable, WAV instead of FLAC, alignment of planets) may or may not be controversial. But is it a TOS#8 violation? Or is it a violation only when it's said a difference is heard? Both require double-blind testing to be meaningful. Not that it matters in bernhold's case.

Edit: (greynol essentially just said the same thing as I was writing this reply.)

QUOTE (greynol @ Jan 5 2012, 17:40) *
simply turning your head can create a vastly different frequency response, unless you're using headphones.


Hmm. smile.gif

This post has been edited by mjb2006: Apr 22 2013, 05:33
Go to the top of the page
+Quote Post
Nessuno
post Apr 22 2013, 08:16
Post #20





Group: Members
Posts: 422
Joined: 16-December 10
From: Palermo
Member No.: 86562



Back on topic, a more meaningful question to pose to the ones asking for advices on lossy codec is not about the devices, but about the usage they're going to make of them: it makes a lot of difference if one is going to listen in environments with high noise floor like public transports or gyms and, most of all, the degree of attention and concentration he/she is going to place on music.

Being able to spot artefacts chasing for them in the quiet of a living room doesn't mean you'll be bothered by them on a train or while exercising, no matter the device! wink.gif

This post has been edited by Nessuno: Apr 22 2013, 08:17


--------------------
... I live by long distance.
Go to the top of the page
+Quote Post
db1989
post Apr 22 2013, 13:56
Post #21





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



QUOTE (IDefyAxioms @ Apr 21 2013, 21:36) *
First, sorry about my improper post placement. I was using the (now obviously incorrect) logic of "I'm using mp3, so I'll just put it here.".
Your logic would have been ideal for MP3 - General, whereas Tech is only for genuinely technical stuff. tongue.gif

QUOTE
Thanks for the reinforcement of the testing. I was going to hold off on it for a while (just shameful laziness, to be truthful), and figured that v2 would be more of a subjective "safe" bet - for now at least.
You’re probably right there. Most people find it to be transparent on the majority of normal listening material. In a lot of cases, it’s probably excessive, even. The key point is that others can’t guarantee that to you. But if you’re willing to trust that accumulated averages and anecdotes, by all means do!

QUOTE
As for settings, I was more or less talking about the extra settings under Foobar2k's "processing" options - stereo configs/conversions, EQ, PPHS, et al. Being portable I would've assumed extra settings wouldn't truly be necessary, given that (as you've stated and as I've read before) presets are generally the best bet.
Your assumption is correct. I don’t really understand why the other options would seem appealing. In terms of “stereo configs”, if you were talking about those offered by foobar2000 as DSPs separate from the Converter, they’re not necessary unless you have a reason, as neither are resampling or the EQ. If you meant LAME’s own type of stereo, joint is the best option and actually means the flexibility to choose between mid/side and simple stereo on a frame-by-frame basis.
Go to the top of the page
+Quote Post
bernhold
post Apr 26 2013, 21:43
Post #22





Group: Members
Posts: 41
Joined: 22-March 13
Member No.: 107330



QUOTE (greynol @ Apr 21 2013, 23:39) *
QUOTE (bernhold @ Apr 21 2013, 13:28) *
Are we going to debate the fact that low quality equipment isn't able to reproduce the same frequency range as high quality equipment?

Sigh. Again(!), you are failing to address masking, the assumptions the encoder makes about it and how is can be broken. Frequency response falls well short of telling the whole story. If it did TOS8 would likely read quite differently.

I recommend you read some of the discussions about it that exist on the forum.


I fail to address address masking because I don't even know what it is. I'm not familiar with the internals of the LAME encoder (or encoders in general). Why would I need to be? All I need to know is: If two differently compressed samples sound the same on equipment A, but sound different on equipment B, it proves my point that the impact of different compression settings varies by equipment. How does it not?
Go to the top of the page
+Quote Post
greynol
post Apr 26 2013, 21:50
Post #23





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



With all the discussions I've read on the matter, I don't find your single data point all that compelling.

http://www.hydrogenaudio.org/forums/index.php?showtopic=8812

This post has been edited by greynol: Apr 26 2013, 21:51


--------------------
YOUR EYES CANNOT HEAR!!!!!!!!!!!
Go to the top of the page
+Quote Post
bernhold
post Apr 26 2013, 22:22
Post #24





Group: Members
Posts: 41
Joined: 22-March 13
Member No.: 107330



Then let's be more specific. Take some low frequency sample, a 20Hz sine wave for example:

http://www.audiocheck.net/audiofrequencysi...or_sinetone.php

(Type in a frequency of 20 Hz and click "Download WAV file")

Compress it with V2 and V9. Play both samples using laptop speakers. You won't hear a difference, because you won't hear anything at all - the laptop speakers are incapable of reproducing low frequencies. Listen again with better equipment and you'll hear a very obvious difference between both samples. I think 100% of the people doing this test will come to the same conclusion. Of course the difference will be a lot smaller in less isolated examples, but that doesn't hurt the thesis in general.

This post has been edited by bernhold: Apr 26 2013, 22:24
Go to the top of the page
+Quote Post
greynol
post Apr 26 2013, 22:48
Post #25





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



I'm sorry, but I don't listen to test tones, nor do I evaluate lossy codecs with them. Furthermore, I don't find this crude first order analysis (that makes absolutely no attempt to explore the concept of how hardware impacts the effectiveness of perceptual coding, you know that stuff that makes most worthwhile lossy encoders work?!?) even remotely interesting.

This post has been edited by greynol: Apr 26 2013, 23:23


--------------------
YOUR EYES CANNOT HEAR!!!!!!!!!!!
Go to the top of the page
+Quote Post

2 Pages V   1 2 >
Closed TopicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 23rd July 2014 - 09:42