Vorbis - Aug 2001 - mp3-wav-cd-audio "acoustically equivalent" to wav-cd-audio ?

A friend of mine made the following comment in a discussion I had with
him that on a website we adminster we should offer

a) WAV or maybe shorten files
b) Ogg as a decent reference lossy encoded version

He's been trying to convince me that we should offer MP3 (in lieu of
WAV) and possibly Ogg.

The audio files are primarily vocals

I am not a physics guy but his statements don't intuitively feel right. 
Maybe he has a misunderstanding which can be clarified

Regards, Yusuf
----------------------------------------------------------------------
regarding WAV, we used to store the masters on the disk, but moved them
offline due to space concerns. If a user downloads the MP3 and creates a WAV
from it, that is going to be precisely equivalent on the audio CD in
quality. I have tested it myself in MATLAB using spectral and phase analysis
(I am a physics guy, after all).

recite WAV -> MP3 -> WAV -> CD Audio is *acoustically equivalent* to
recite
WAV -> CD audio.

to confirm, Perform this experiment at home (I have done so)

1. recite to WAV. save as A.wav
2. encode A.mp3 from A.wav
3. create B.wav from A.mp3.
4. ask someone to burn A.wav and B.wav to an audio CD (double blind -
neither you, nor they, shoudl know which is which).
5. listen to the CD in a standard player. You will hear zero differrence in
sound quality.

I am NOT saying,  not to make WAV available. But I am pointing out that the
vast majority of repackagers only want to create CDs. And that they will not
be able to improve upon the sound quality , though it is quite possible they
may degrade it, during their own post-processing stages. We can safely
assume that most repackagers will choose to go the easy route, and we can
easily have a note saying "true WAV masters for many audio files are
available for more advanced processing. Please contact Webmaster etc etc".

as an aside - there are already several excellent freeware tools which allow
you to create playlists of MP3 files and burn them straight to audio CD,
doing the WAV interpolation for you directly. No such tools yet exist for
OGG of course, but they too will come in a year or so (the ones I have seen
for OGG are still quite unstable - if it doesnt come from vorbis.com, I dont
trust it).

----------------------------------------------------------------------

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

>A friend of mine made the following comment in a discussion I had with
>him that on a website we adminster we should offer
>
>a) WAV or maybe shorten files
>b) Ogg as a decent reference lossy encoded version
>
>He's been trying to convince me that we should offer MP3 (in lieu of
>WAV) and possibly Ogg.
>
>The audio files are primarily vocals
>
>I am not a physics guy but his statements don't intuitively feel right. 
>Maybe he has a misunderstanding which can be clarified
>
>Regards, Yusuf
>----------------------------------------------------------------------
>regarding WAV, we used to store the masters on the disk, but moved them
>offline due to space concerns. If a user downloads the MP3 and creates a WAV
>from it, that is going to be precisely equivalent on the audio CD in
>quality. I have tested it myself in MATLAB using spectral and phase analysis
>(I am a physics guy, after all).
IMHO, as a professional physicist, he has no clue.

MP3 is a lossy encoding. That means you lose something.
At a low bit rate, you tend to lose a lot.
Just try doing this at 32 kbps stereo and you _will_
notice the acoustic nonequivalence.

Regards
Marshall Eubanks
>
>recite WAV -> MP3 -> WAV -> CD Audio is *acoustically equivalent*
to recite
>WAV -> CD audio.
>
>to confirm, Perform this experiment at home (I have done so)
>
>1. recite to WAV. save as A.wav
>2. encode A.mp3 from A.wav
>3. create B.wav from A.mp3.
>4. ask someone to burn A.wav and B.wav to an audio CD (double blind -
>neither you, nor they, shoudl know which is which).
>5. listen to the CD in a standard player. You will hear zero differrence in
>sound quality.
>
>I am NOT saying,  not to make WAV available. But I am pointing out that the
>vast majority of repackagers only want to create CDs. And that they will not
>be able to improve upon the sound quality , though it is quite possible they
>may degrade it, during their own post-processing stages. We can safely
>assume that most repackagers will choose to go the easy route, and we can
>easily have a note saying "true WAV masters for many audio files are
>available for more advanced processing. Please contact Webmaster etc
etc".
>
>as an aside - there are already several excellent freeware tools which allow
>you to create playlists of MP3 files and burn them straight to audio CD,
>doing the WAV interpolation for you directly. No such tools yet exist for
>OGG of course, but they too will come in a year or so (the ones I have seen
>for OGG are still quite unstable - if it doesnt come from vorbis.com, I dont
>trust it).
>
>----------------------------------------------------------------------
>
>--- >8 ----
>List archives:  http://www.xiph.org/archives/
>Ogg project homepage: http://www.xiph.org/ogg/
>To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
>containing only the word 'unsubscribe' in the body.  No subject is
needed.
>Unsubscribe messages sent to the list will be ignored/filtered.
>
>
Marshall Eubanks

tme@21rst-century.com

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

Hello Yusuf,

31. avgust 2001, 9:54:38, you wrote:

YG> 1. recite to WAV. save as A.wav
YG> 2. encode A.mp3 from A.wav
YG> 3. create B.wav from A.mp3.
YG> 4. ask someone to burn A.wav and B.wav to an audio CD (double blind -
YG> neither you, nor they, shoudl know which is which).
YG> 5. listen to the CD in a standard player. You will hear zero differrence
in
YG> sound quality.

This depends, on which bitrate you select for MP3.

A lot of people won't hear any difference between 128kbit MP3 and
original, however many will (and don't even try to play 128kbps mp3 to
an audiophile - he'll probably get heart attack:). Anyway, minimum
bitrate for what you're doing is 192kbps, but some people will hear
difference there, too...


-- 
Jernej Simoncic, jernej.simoncic@guest.arnes.si
http://www2.arnes.si/~sopjsimo/
ICQ: 26266467

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

----- Original Message -----
From: "Yusuf Goolamabbas" <yusufg@outblaze.com>
To: <vorbis@xiph.org>
Sent: Friday, August 31, 2001 9:54 AM
Subject: [vorbis] mp3-wav-cd-audio "acoustically equivalent" to
wav-cd-audio
?
> regarding WAV, we used to store the masters on the disk, but moved them
> offline due to space concerns. If a user downloads the MP3 and creates a
WAV
> from it, that is going to be precisely equivalent on the audio CD in
> quality.
An MP3 will _never_ reach the precise equivalent of an audio CD
in quality (for normal music).
> I have tested it myself in MATLAB using spectral and phase analysis
> (I am a physics guy, after all).
A pretty bad one I guess. You should know that looking at the graphs
is NOT going to tell you much about the subjective audio quality.
> 5. listen to the CD in a standard player. You will hear zero differrence
in
> sound quality.
Many of us have done the same and the conclusion is:

a) there are always _audible_ differences in _some_ samples
b) how many people will be able to tell the difference depends vastly
on the used MP3 encoder, used settings, and listener experience
> And that they will not
> be able to improve upon the sound quality , though it is quite possible
they
> may degrade it, during their own post-processing stages.
Here is something to think about. What if I want to equalize the
sound before listening? With the WAV that would be no problem.
But if I try to equalize the WAV produced by the decoded MP3 I
am going to violate some of the assumptions the MP3 encoder made
and suddenly some artifacts may become very audible.

What if I someone gets a CD made from the MP3 and tries to rip it
and encode to Vorbis? It will sound like shit. If the CD had been
made from the original WAV it would have been no problem.


--
GCP

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.