Topic: VBR(0) vs. 320kbps MP3

[acidjack]

I found the other FLAC>MP3 topic so interesting I decided I would repost my question here. 

Archive.org uses VBR for their MP3 standard.  I assume theirs are at 0 (highest quality) but I suppose I can't confirm that. 

320kbps MP3 files are larger, I believe, than VBR0 files.  But are they better?  Obviously one has more "information" still there, but I would think, intuitively, that VBR0 is designed to be "smarter" about when it compresses information. 

I'm interested in hearing both anecdotal and scientific explanations - both what people prefer, but also what the actual technical differences are.  I personally haven't spectrum analyzed files side by side to look at them.

[greenone]

Archive.org uses VBR for their MP3 standard.  I assume theirs are at 0 (highest quality) but I suppose I can't confirm that. 

If it helps, here's a log from a recent derive of one track so you can see what settings are used by ffmpeg:

EXECUTING PLAN
  Deriving guster2000-04-12d1t06.mp3 from guster2000-04-12d1t06.flac
    Class: MP3
    Class parameters: --preset standard
    Source file: /tmp/derive/guster2000-04-12.flac16/guster2000-04-12d1t06.flac
    Target file: /tmp/derive/guster2000-04-12.flac16/guster2000-04-12d1t06.mp3


<--------------- Module MP3 (v27464 2010May06 06:56) Starting  PDT: 2010-05-06 06:56:21 ------------------

[ PDT: 2010-05-06 06:56:21 ] Executing: rm -rf /tmp/derive-guster2000-04-12.flac16-MP3/
[ PDT: 2010-05-06 06:56:21 ] Executing: mkdir /tmp/derive-guster2000-04-12.flac16-MP3/
[ PDT: 2010-05-06 06:56:21 ] Executing: /petabox/sw/bin/timeout -s 9 14400 /usr/bin/ffmpeg -v 0 -y -vn -i /tmp/derive/guster2000-04-12.flac16/guster2000-04-12d1t06.flac -ac 2 -ar 44100 -f wav -|/usr/bin/lame --quiet --disptime 10 --preset standard --ta 'Guster' --tt 'Demons' --tn '06' --tl '2000-04-12 - Conte Forum, Boston College' --add-id3v2  --tc 'http://www.archive.org/details/guster2000-04-12.flac16' --tv 'LINK=http://www.archive.org/details/guster2000-04-12.flac16' - /tmp/derive-guster2000-04-12.flac16-MP3/tmp.mp3 > /dev/null 2>&1
FFmpeg version SVN-r22939, Copyright (c) 2000-2010 the FFmpeg developers
  built on Apr 22 2010 07:48:47 with gcc 4.3.3
  configuration: --disable-shared --enable-avfilter --enable-avfilter-lavf --enable-gpl --enable-libdc1394 --enable-libdirac --enable-libfaac --enable-libfaad --enable-libgsm --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libschroedinger --enable-libspeex --enable-libvorbis --enable-libxvid --enable-nonfree --enable-postproc --enable-pthreads --enable-static --disable-vdpau --enable-x11grab --prefix=/usr --enable-version3
  libavutil     50.14. 0 / 50.14. 0
  libavcodec    52.66. 0 / 52.66. 0
  libavformat   52.61. 0 / 52.61. 0
  libavdevice   52. 2. 0 / 52. 2. 0
  libavfilter    1.19. 0 /  1.19. 0
  libswscale     0.10. 0 /  0.10. 0
  libpostproc   51. 2. 0 / 51. 2. 0
[flac @ 0x1e680b0]MAX_READ_SIZE:5000000 reached
Input #0, flac, from '/tmp/derive/guster2000-04-12.flac16/guster2000-04-12d1t06.flac':
  Duration: 00:05:47.09, bitrate: 723 kb/s
    Stream #0.0: Audio: flac, 44100 Hz, 2 channels, s16
Output #0, wav, to 'pipe:':
  Metadata:
    encoder         : Lavf52.61.0
    Stream #0.0: Audio: pcm_s16le, 44100 Hz, 2 channels, s16, 1411 kb/s
Stream mapping:
  Stream #0.0 -> #0.0
Press [q] to stop encoding
size=   59792kB time=347.09 bitrate=1411.2kbits/s   
video:0kB audio:59792kB global headers:0kB muxing overhead 0.000072%
[ PDT: 2010-05-06 06:56:55 ] Executing: mv /tmp/derive-guster2000-04-12.flac16-MP3/tmp.mp3 /tmp/derive/guster2000-04-12.flac16/guster2000-04-12d1t06.mp3

Cleaning up temporary dir:
[ PDT: 2010-05-06 06:56:56 ] Executing: rm -rf /tmp/derive-guster2000-04-12.flac16-MP3/

--------------- Module MP3 Finished  PDT: 2010-05-06 06:56:56 (Took 35.0 seconds) ------------------->

[boojum]

To get a good handle on the current state of MP3 engineering go to http://hydrogenaudio.org/.   These fellows have a separate forum on the subject, constant testing which all and any can participate in and "how to" sections.  They are the most current board that I know of on this.   Check it out if you want the skinny.


Cheers

[OFOTD]

Keep in mind though that the hydrogenaudio forums while a very nice resource has biases that direct its conversations on many subjects as defined in large part by its TOS.   Very much an our way or the highway type of conversation.     

[rhinowing]

pretty sure it's long since been established that V0 > 320 CBR

[DMBprez]

pretty sure it's long since been established that V0 > 320 CBR

In terms of file space, yes.  Quality? No. They are stupidly close, but 320 CBR is slightly better.

[boojum]

Keep in mind though that the hydrogenaudio forums while a very nice resource has biases that direct its conversations on many subjects as defined in large part by its TOS.   Very much an our way or the highway type of conversation.   

While this has not been my experience over the course of eight years it apparently has been yours. I have always found it more welcoming and collegial than TS. And they do testing!

[OFOTD]

Keep in mind though that the hydrogenaudio forums while a very nice resource has biases that direct its conversations on many subjects as defined in large part by its TOS.   Very much an our way or the highway type of conversation.   

While this has not been my experience over the course of eight years it apparently has been yours. I have always found it more welcoming and collegial than TS. And they do testing!

I'm not disputing that board being welcoming, great members as well as having a a wealth of knowledge.  But just in fairness if you read the sites TOS its quite clear that they have an established bias on what they deem to be the 'right' way to do things.   Your posts here seem to coincide alot with the opinions of some on that forum so i'd expect you to be sensitive to the site. .   No disrespect to the site at all just a head's up on the site having a pretty open bias to start.

[Scooter123]

At 320kbs, I doubt I could tell the difference. 

But I don't have many mp3s.  All of my stuff is on flac, save and except iPod in which I used fairly cheap headphones/earbuds. 

[boojum]

Keep in mind though that the hydrogenaudio forums while a very nice resource has biases that direct its conversations on many subjects as defined in large part by its TOS.   Very much an our way or the highway type of conversation.   

While this has not been my experience over the course of eight years it apparently has been yours. I have always found it more welcoming and collegial than TS. And they do testing!

I'm not disputing that board being welcoming, great members as well as having a a wealth of knowledge.  But just in fairness if you read the sites TOS its quite clear that they have an established bias on what they deem to be the 'right' way to do things.   Your posts here seem to coincide alot with the opinions of some on that forum so i'd expect you to be sensitive to the site. .   No disrespect to the site at all just a head's up on the site having a pretty open bias to start.

Hydrogenaudio Terms of Service / Notices

1. All messages express the views of the author, and are not and should not be taken as representative of the official policies of the forum or the views of its staff, unless explicitly stated as such by the authoritative members of the staff.

2. All members, at the staff's discretion, must converse in an acceptable fashion to be allowed the privilege of continued participation.

3. All members must acknowledge and accept that data integrity is not the liability of the Hydrogenaudio staff. Those who do not accept this should not contribute any content whatsoever to the Hydrogenaudio community.

4. The Hydrogenaudio community requires that all members provide a valid email address. This email address will only be used for user account validation and activities directly related to the features of this community, and will never be given to any third party.

5. All members that post to the Hydrogenaudio community must acknowledge and discuss only acceptable topics when starting a new thread. If posting to an already existing thread, they must continue in the vein of discussion that the thread has already manifested; if they wish to change topics, they must start a new thread.

6. All members that create new threads must use a topic title which is understandable and describes the content of the thread properly. Furthermore, all new threads must be started in the correct forum categories, and posted only once -- in the appropriate category -- throughout the entire forum. Double posts will be removed, and threads with inappropriate topics will be moved. A member that continue violates this rule may be subject to administrative action.

7. All members must acknowledge the authority of the staff insofar as administrative actions are concerned. These issues are no longer open for discussion once the staff member has indicated that the matter is closed. Further concerns regarding an action should be communicated via a private message to the staff member, and not repeatedly discussed publicly. Furthermore, once a thread on a particular topic has been closed or removed and indicated as being unacceptable by a staff member, starting a new discussion on the exact same topic will result in administrative action being taken against the poster.

8. All members that put forth a statement concerning subjective sound quality, must -- to the best of their ability -- provide objective support for their claims. Acceptable means of support are double blind listening tests (ABX or ABC/HR) demonstrating that the member can discern a difference perceptually, together with a test sample to allow others to reproduce their findings. Graphs, non-blind listening tests, waveform difference comparisons, and so on, are not acceptable means of providing support.

9. All members must refrain from posting links to -- or information regarding how to obtain -- copyrighted or illegal material. Discussion containing information of how to obtain such material, how to bypass protection methodologies of such material, or how to otherwise violate laws pertaining to such matters will not be tolerated, and participating members may be subject to administrative action.

10. All members must post in English. Content in other languages are allowed as long as full English translations are provided, or otherwise at the discretion of the staff. User profile information (including sigs) in other languages is allowed.

11. All member sigs must be kept brief -- 3 to 4 lines at most. Signatures and avatars must also be non-obtrusive, non-offensive and work-safe.

12. All members are only allowed to have a single user account with the forums. Members who are found to violate this rule may be subject to immediate banishment or other administrative action.

13. Updates posted in the "News Submissions" forum should be of importance for the community.

14. All members must refrain from using the forum solely for advertising purposes.



Please point out the bias in the HA TOS.  Is number 8 a problem?  The one which requires testing?     

[Todd R]

pretty sure it's long since been established that V0 > 320 CBR

In terms of file space, yes.  Quality? No. They are stupidly close, but 320 CBR is slightly better.

This site seems to be a good one for testing of lossy codecs:
http://soundexpert.org/ratings

If you look at the ratings for the 320 kb/s codecs and compare them to the 224 kb/s grouping (which includes VBR 0), it looks like the 320kb/s codecs have been tested to be better, though both provide a compressed audio stream that is above the threshold of human perception on average.  (Not to say that some people can't detect the difference from the lossless version, but most do not.  I need to do some of these tests myself on my listening system to see where I come out on detecting the difference, but I know I'll keep using mp3's for my car and for my ipod listening.)

Glad I came across this thread, I'm just getting started re-transferring spindles of shn'd and flac'd masters and fixing and SBEs, tagging the flacs (or making flacs from shn's), and making/tagging mp3's of them while I'm at it.  I've been converting via VBR 0, but since smallest size isn't that important, I think I'll switch to 320 CBR.

[Gutbucket]

This site seems to be a good one for testing of lossy codecs:
http://soundexpert.org/ratings

Good resource, thanks for the link!