IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
LAME -V0 - is it constrained to hit 245 kbps avg?
VBRecon
post Jan 13 2014, 00:05
Post #1





Group: Members
Posts: 4
Joined: 12-January 14
Member No.: 113978



Hello everyone. I'm trying to understand if the LAME -V0 preset is constrained in the amount of bitrate it can use.

For instance, if it was tasked with compressing some hypothetical "most challenging audio track ever", is it capable of ramping up the bitrate to 320 kbps and holding that for the duration of the track if it was deemed necessary to achieve the maximum quality?

Another way to put it... Is the -V0 preset a choice for maximum constant quality with bitrate and file size being unconstrained? Or is it a choice for maximum quality within a bitrate range that averages around 245 kbps? I'm guessing it's the former but I haven't been able to find a clear statement to that effect.

I ask because the encodings I do with that preset vary in bitrate allocation, but they always cluster around 224/256 (as shown in the command line window graph). Which led me to question whether it was constrained or if that's all the bitrate that LAME needs to achieve optimum results.

Thanks.

Go to the top of the page
+Quote Post
saratoga
post Jan 13 2014, 00:07
Post #2





Group: Members
Posts: 4975
Joined: 2-September 02
Member No.: 3264



Its not constrained, that would be ABR.

VBR in lame picks whatever bitrate it needs to hit a given quality level.
Go to the top of the page
+Quote Post
VBRecon
post Jan 13 2014, 00:57
Post #3





Group: Members
Posts: 4
Joined: 12-January 14
Member No.: 113978



QUOTE (saratoga @ Jan 12 2014, 15:07) *
Its not constrained, that would be ABR.

VBR in lame picks whatever bitrate it needs to hit a given quality level.

So in LAME's current state the -V0 preset achieves the maximum quality the encoder is capable of, while using no more bitrate than necessary to achieve it? In theory a more efficient equivalent to the -b 320 preset? Generally speaking that is. I wouldn't expect the differing approaches to result in truly identical files, I'm just trying to a get a better grip on how these 2 presets compare.
Go to the top of the page
+Quote Post
saratoga
post Jan 13 2014, 01:22
Post #4





Group: Members
Posts: 4975
Joined: 2-September 02
Member No.: 3264



QUOTE (VBRecon @ Jan 12 2014, 18:57) *
QUOTE (saratoga @ Jan 12 2014, 15:07) *
Its not constrained, that would be ABR.

VBR in lame picks whatever bitrate it needs to hit a given quality level.

So in LAME's current state the -V0 preset achieves the maximum quality the encoder is capable of, while using no more bitrate than necessary to achieve it? In theory a more efficient equivalent to the -b 320 preset? Generally speaking that is. I wouldn't expect the differing approaches to result in truly identical files, I'm just trying to a get a better grip on how these 2 presets compare.


In 320 CBR, all frames are 320, and any excess space is simply padded with zeros. In -V0, rather than zero padding, the frame size is reduced if the encoder judges that there are excess bits available.
Go to the top of the page
+Quote Post
mjb2006
post Jan 13 2014, 01:33
Post #5





Group: Members
Posts: 811
Joined: 12-May 06
From: Colorado, USA
Member No.: 30694



VBR: consistent quality, variable bitrate.
CBR: variable quality, consistent bitrate.
Bitrate for maximum quality = the minimum bitrate that achieves transparency, for that audio, for you—or any higher bitrate; once transparency is achieved, they're all the same, unless something's wrong with the encoder.

If you worry that V0 isn't good enough for you, you can try 320 CBR, but don't say it's higher quality without posting ABX test results showing you can hear a difference. (I don't think anyone ever has, to date.) It would be better to just switch to lossless if you want to eradicate the risk, however slight, of ever hearing a difference.

This post has been edited by mjb2006: Jan 13 2014, 01:34
Go to the top of the page
+Quote Post
VBRecon
post Jan 13 2014, 02:41
Post #6





Group: Members
Posts: 4
Joined: 12-January 14
Member No.: 113978



QUOTE (saratoga @ Jan 12 2014, 16:22) *
In 320 CBR, all frames are 320, and any excess space is simply padded with zeros. In -V0, rather than zero padding, the frame size is reduced if the encoder judges that there are excess bits available.

That's very interesting. So the data content is the same, it's just the padding/bitrate format that differs.


QUOTE (mjb2006 @ Jan 12 2014, 16:33) *
VBR: consistent quality, variable bitrate.
CBR: variable quality, consistent bitrate.
Bitrate for maximum quality = the minimum bitrate that achieves transparency, for that audio, for you—or any higher bitrate; once transparency is achieved, they're all the same, unless something's wrong with the encoder.

If you worry that V0 isn't good enough for you, you can try 320 CBR, but don't say it's higher quality without posting ABX test results showing you can hear a difference. (I don't think anyone ever has, to date.) It would be better to just switch to lossless if you want to eradicate the risk, however slight, of ever hearing a difference.

As I suspected the VBR mode could alternately be referred to as Constant Quality, with -V0 being the best LAME is capable of.

I understand about the transparency threshold, but for me it's been something of a moving target as it varies by playback equipment and environment. My equpment seem to be improving over time so I want to consider that when choosing an encoding technique. Knowing what I learned today I see no benefit in choosing 320 CBR over -V0 VBR for my current and near-future playback needs. By extension lossless encoding would also be of no value, but it would certainly be my choice if the goal was to preserve original source media in a digital/backup format.
Go to the top of the page
+Quote Post
saratoga
post Jan 13 2014, 03:10
Post #7





Group: Members
Posts: 4975
Joined: 2-September 02
Member No.: 3264



QUOTE (VBRecon @ Jan 12 2014, 20:41) *
QUOTE (saratoga @ Jan 12 2014, 16:22) *
In 320 CBR, all frames are 320, and any excess space is simply padded with zeros. In -V0, rather than zero padding, the frame size is reduced if the encoder judges that there are excess bits available.

That's very interesting. So the data content is the same, it's just the padding/bitrate format that differs.


Its not necessarily the same data.

QUOTE (VBRecon @ Jan 12 2014, 20:41) *
I understand about the transparency threshold, but for me it's been something of a moving target as it varies by playback equipment and environment.


The best you can do for spotting compression artifacts is a typical pair of headphones, so I wouldn't worry about that.
Go to the top of the page
+Quote Post
VBRecon
post Jan 13 2014, 05:39
Post #8





Group: Members
Posts: 4
Joined: 12-January 14
Member No.: 113978



QUOTE (saratoga @ Jan 12 2014, 18:10) *
Its not necessarily the same data.

OK, so like I said earlier, the differing approaches will not result in truly identical results. Which would support the LAME "recommended switches" help page comment on 320 CBR: "With the exception of a few situations, quality is rarely better than the highest VBR profiles..."

Well that doesn't change the conclusion I reached in my last post, of 320 not offering a benefit over -V0 for my daily needs. Thank you both for helping get rid of that nagging doubt I had each time I'd compress something.

This post has been edited by VBRecon: Jan 13 2014, 05:41
Go to the top of the page
+Quote Post
Kohlrabi
post Jan 13 2014, 07:25
Post #9





Group: Super Moderator
Posts: 1050
Joined: 12-March 05
From: Kiel, Germany
Member No.: 20561



FWIW, even -V0 is far above my transparency threshold, so -V0 might also be total overkill for you. Don't fear to test that.

This post has been edited by Kohlrabi: Jan 13 2014, 07:26


--------------------
Ceterum censeo Masterdiskem esse delendam.
Go to the top of the page
+Quote Post
ash92
post Jan 13 2014, 09:57
Post #10





Group: Members
Posts: 9
Joined: 27-July 13
Member No.: 109325







My songs. At many bitrates
Go to the top of the page
+Quote Post
includemeout
post Jan 13 2014, 10:36
Post #11





Group: Members
Posts: 282
Joined: 16-December 09
From: Maringá, Brazil
Member No.: 76067



@ash92 Your point being...?

This post has been edited by includemeout: Jan 13 2014, 10:39


--------------------
Listen to the music, not the media.
Go to the top of the page
+Quote Post
ash92
post Jan 13 2014, 12:09
Post #12





Group: Members
Posts: 9
Joined: 27-July 13
Member No.: 109325



I think VBR is not limited and if the song is complex enough can reach a higher bitrate. In the picture above I have given an example of complex music.
I apologize for the inaccuracy use translator.
Go to the top of the page
+Quote Post
includemeout
post Jan 13 2014, 12:24
Post #13





Group: Members
Posts: 282
Joined: 16-December 09
From: Maringá, Brazil
Member No.: 76067



Yeah, but from what you could gather, don't you think the previous posts had already made that quite clear?


--------------------
Listen to the music, not the media.
Go to the top of the page
+Quote Post
ash92
post Jan 13 2014, 12:59
Post #14





Group: Members
Posts: 9
Joined: 27-July 13
Member No.: 109325



Sorry smile.gif
Go to the top of the page
+Quote Post
TomasPin
post Jan 13 2014, 16:37
Post #15





Group: Members
Posts: 204
Joined: 5-June 13
From: Argentina
Member No.: 108508



QUOTE (Kohlrabi @ Jan 13 2014, 03:25) *
FWIW, even -V0 is far above my transparency threshold

I'd say it's far above most mortals' transparency thresholds... I pity those for whom it's "not enough" (whether they are certain of it, or not and just wasting space).

Anyway, I think the OP was given a satisfactory answer.

This post has been edited by TomasPin: Jan 13 2014, 16:40


--------------------
A man and his music: http://tubular.net/
Go to the top of the page
+Quote Post
naturfreak
post Jan 13 2014, 16:52
Post #16





Group: Members
Posts: 176
Joined: 16-October 03
Member No.: 9338



A wild guess: I get MP3 files with that kind of bitrate distribution when I encode music to 320 kbps and run them through mp3packer afterwards. Then most of the mp3 files are in the range of 290 to 315 kbps.
Go to the top of the page
+Quote Post
JJZolx
post Jan 13 2014, 18:28
Post #17





Group: Members
Posts: 396
Joined: 26-November 04
Member No.: 18345



QUOTE (VBRecon @ Jan 12 2014, 18:41) *
I understand about the transparency threshold, but for me it's been something of a moving target as it varies by playback equipment and environment. My equpment seem to be improving over time so I want to consider that when choosing an encoding technique. Knowing what I learned today I see no benefit in choosing 320 CBR over -V0 VBR for my current and near-future playback needs. By extension lossless encoding would also be of no value, but it would certainly be my choice if the goal was to preserve original source media in a digital/backup format.


Why not just go lossless and be done with it? Hard disk space is so cheap today that unless you're ripping and encoding tens of thousands of CDs, the difference between using Mp3 and FLAC is going to be a few dollars worth of drive space.
Go to the top of the page
+Quote Post

Reply to this 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: 24th September 2014 - 03:47