I have many mp3's of talk radio shows which were recorded as stereo mp3
files. The audio content is really the same on both channels, so the
encoding is very inefficient. These files are twice the data size that
they could have been if they had been originally recorded as mono
files. The mp3's are the only source available, so I can't go back to
a higher quality source.

I have been looking around for weeks to find an mp3 converter than can
"extract" one of the channels from a stereo mp3 file. I want to make
new mono files with the same sound quality but half the data size. I
really don't want to re-encode or re-sample anything! So far I have
had no luck finding what I need. Is it possible to do this?

Thanks, R2

On 9/15/05 12:01 AM, in article
, "
wrote:

I have many mp3's of talk radio shows which were recorded as stereo mp3
files. The audio content is really the same on both channels, so the
encoding is very inefficient. These files are twice the data size that
they could have been if they had been originally recorded as mono
files. The mp3's are the only source available, so I can't go back to
a higher quality source.

I have been looking around for weeks to find an mp3 converter than can
"extract" one of the channels from a stereo mp3 file. I want to make
new mono files with the same sound quality but half the data size. I
really don't want to re-encode or re-sample anything! So far I have
had no luck finding what I need. Is it possible to do this?

Thanks, R2

Don;t you just re-construct the file as a MONO mp3?

wrote in message
ups.com
I have many mp3's of talk radio shows which were recorded
as stereo mp3 files. The audio content is really the
same on both channels, so the encoding is very
inefficient. These files are twice the data size that
they could have been if they had been originally recorded
as mono files. The mp3's are the only source available,
so I can't go back to a higher quality source.

I have been looking around for weeks to find an mp3
converter than can "extract" one of the channels from a
stereo mp3 file. I want to make new mono files with the
same sound quality but half the data size. I really
don't want to re-encode or re-sample anything! So far I
have had no luck finding what I need. Is it possible to
do this?

Your point seems to be well-taken, but I don't know how much
waste there really is if the recordings are done in joint
stereo.

The files are twice as big as they need to be. That means the amount
of wasted data is 50% of the total. That is A LOT of wasted space as
far as I am concerned. Don't you agree?

The only way I know how to "reconstruct" a file as mono is to
"re-sample" or "re-encode" it. As I said, I don't want to lose sound
quality through the re-sampling process.

I am a bit disapointed that no one seems to know the answer to my
question, so please let me state the question again. Is it possible to
extract one channel into a mono mp3 file without losing any sound
quality?

Thanks, R2

On 16 Sep 2005 16:37:57 -0700, "R2" wrote:

The files are twice as big as they need to be. That means the amount
of wasted data is 50% of the total. That is A LOT of wasted space as
far as I am concerned. Don't you agree?

Arny can speak for himself, but part of his point
is that your 50% assumption is incorrect for MP3's.

Good fortune,

Chris Hornbeck
Irony and sarcasm emoticons available by request.
You place 'em.

Well I would also like to hear back from Arny on this one. But to be
sure just took a CD quality audio file of 2:35 and 26MB. I turned it
into a stereo 128 kpbs file of 2.36MB and also a corresponding mono
file at 64 kbps containing 1.18MB. The actual bit rate per channel is
the same but the file is exactly one half the size. Am I missing
something here or is it not true that mono mp3 files are exactly one
half the size of a stereo mp3 at the same bit rate?

On 9/16/05 7:37 PM, in article
, "R2"
wrote:

The files are twice as big as they need to be. That means the amount
of wasted data is 50% of the total. That is A LOT of wasted space as
far as I am concerned. Don't you agree?

The only way I know how to "reconstruct" a file as mono is to
"re-sample" or "re-encode" it. As I said, I don't want to lose sound
quality through the re-sampling process.

I am a bit disapointed that no one seems to know the answer to my
question, so please let me state the question again. Is it possible to
extract one channel into a mono mp3 file without losing any sound
quality?

Look, you;re ALREADY into lossy compression here. Without access to the
original file, the best you can do is a real clean re-compress-to-mono.
I don;t know if there's anything better than that. The damage has been done.
Just use a GOOD encoder to do the new process.

I fully understand what you are saying about LOSSY compression. That
is exactly why I was hoping for a way to extract a channel without
re-sampling it. The original file sounds OK. I just don't want to
cause any more loss.

On 16 Sep 2005 17:39:31 -0700, "R2" wrote:

Well I would also like to hear back from Arny on this one. But to be
sure just took a CD quality audio file of 2:35 and 26MB. I turned it
into a stereo 128 kpbs file of 2.36MB and also a corresponding mono
file at 64 kbps containing 1.18MB. The actual bit rate per channel is
the same but the file is exactly one half the size. Am I missing
something here or is it not true that mono mp3 files are exactly one
half the size of a stereo mp3 at the same bit rate?

Do a google search on "joint stereo" + "MP3" or some such
on this newsgroup, and you'll get some appropriate info.
It's arisen fairly recently.

Sorry, it's not a topic I'm personally interested in, so
can't quote authoritatively. The gist is that some (all?)
MP3 encoders operate on the sum and difference separately,
with different audibility weightings. I damn sure would, if I
had to design one.

Good fortune,

Chris Hornbeck
Irony and sarcasm emoticons available by request.
You place 'em.

On 16 Sep 2005 17:57:19 -0700, "R2" wrote:

I fully understand what you are saying about LOSSY compression. That
is exactly why I was hoping for a way to extract a channel without
re-sampling it. The original file sounds OK. I just don't want to
cause any more loss.

I guess it's just a sign of the national emotional
numbness in the wake of the disater that you haven't
received a half dozen posts to the effect of:

"Then leave it the hell alone. It's already an MP3,
fercrissakes" complete with exclamation marks.

But nobody has the energy. Lucky you,

Chris Hornbeck
Irony and sarcasm emoticons available by request.
You place 'em.

R2 wrote:
I fully understand what you are saying about LOSSY compression. That
is exactly why I was hoping for a way to extract a channel without
re-sampling it. The original file sounds OK. I just don't want to
cause any more loss.

What you are missing is that the lossy compression scheme basically
throws away information that isn't important... and if both channels
are the same and the encoder is set up for coupled stereo, it will
not bother storing any L-R stereo information in the file because
there will be nothing to store.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

On 9/16/05 8:39 PM, in article
, "R2"
wrote:

Well I would also like to hear back from Arny on this one. But to be
sure just took a CD quality audio file of 2:35 and 26MB. I turned it
into a stereo 128 kpbs file of 2.36MB and also a corresponding mono
file at 64 kbps containing 1.18MB. The actual bit rate per channel is
the same but the file is exactly one half the size. Am I missing
something here or is it not true that mono mp3 files are exactly one
half the size of a stereo mp3 at the same bit rate?


Rememeber, the COMPRESSED DATA FILE is NOT a real representation of the
resultant listenable music file.

Now I have NO idea how these thigns work, but going on the idea that if -I-
can think of a half-clever way to solve this problem, the real
number-cruncher and statistical-analysis guys must've already come up with a
WAYYYYY more effective and efficient approach...
With that in mind:
Considering how smart the compression program at hand might be, it depends
on the music at any moment and how much DIFFERENCE there actually is in the
mix L vs R. Much music is pretty darned mono (especially in the bass) in
general and a good program may well not actually use up anything NEAR
twice-as-much storage for a stereo-vs-mono version...

I could see it looking at the mix NOT as LEFT-ch and RIGHT-ch,
but as a
L+R channel
and a
L-R channel...
And if the stereo info was minimal, the L-R channel wouldn't have all that
much data over what's already in the L+R channel... The STEREO encoded file
might only have, say 10% more data in the file than the MONO version.

In your case, with LEFT being the SAME as RIGHT, there would be the data
amount from whatever it did dealing with the L+R channel,
and then, since there would be NOTHING in the L-R channel, your resultant
data file would be pretty much identical for a mono or a stereo version.
For all you know, you already HAVE it as far down as it can go.

R2 wrote:

Am I missing something here or is it not true that mono mp3 files are
exactly one
half the size of a stereo mp3 at the same bit rate?

What you are missing is that the majority of mp3 stereo encodings at
128k or less are done using "joint stereo", a technique that takes
advantage of the fact that the left and right channels of a stereo music
signal are very similar. Joint stereo gives far better sound for a given
bit rate than encoding the two channels independently at half the bit
rate each.

So it's likely that your 128k stereo MP3 is joint stereo. if you convert
to mono at 64k, you will get half the file size all right, but you'll
get a drop in sound quality. You'd need to re-encode at about 96k to
keep equivalent quality, resulting in a file size reduction of only 25%,
for example.

By the way, the LAME encoder http://lame.sf.net will take an MP3 as
input, for re-ecoding in a different format. That what I use; I can't
speak for other software.

--
Anahata
-+- http://www.treewind.co.uk
Home: 01638 720444 Mob: 07976 263827

R2 wrote:
The files are twice as big as they need to be. That means the amount
of wasted data is 50% of the total. That is A LOT of wasted space as
far as I am concerned. Don't you agree?

The only way I know how to "reconstruct" a file as mono is to
"re-sample" or "re-encode" it. As I said, I don't want to lose sound
quality through the re-sampling process.

I am a bit disapointed that no one seems to know the answer to my
question, so please let me state the question again. Is it possible to
extract one channel into a mono mp3 file without losing any sound
quality?

Thanks, R2


Convert it to stereo .wav, open it in an audio editor such as Cubase or
PT that will allow you to split the .wav into two seperate mono
channels, then export one, or both of the channels to a mono .wav or
..mp3....a lot of radio mixes are either done in mono, or mixed in such a
way that it doesn't entirely matter what channel you take....pick, L, or
R, or both, export it, and you'll be fine........Best of luck!

Jonny Durango

R2 wrote:
I fully understand what you are saying about LOSSY compression. That
is exactly why I was hoping for a way to extract a channel without
re-sampling it. The original file sounds OK. I just don't want to
cause any more loss.

The least intrusive thing you can do is play it back through the best
(analog) player you have, and re-record it through the best (analog)
recorder you have. I suppose you could call the "re-sampling" but it's
a real process, not a mathematical one. With today's equipment, even
inexpensive equipment, while the result may not be mathematically
equivalent to the source, it isn't going to sound enough different to
matter.

Screw programs. Just use a cable. In less that the time it takes to
find the perfect program, download and install it, figure out how to
work it, and do the job, you could be listening to your half-sized
recording in glorious mono.

"R2" wrote in message
oups.com

The files are twice as big as they need to be. That
means the amount of wasted data is 50% of the total.
That is A LOT of wasted space as far as I am concerned.
Don't you agree?


I don't know if I agree or disagree because I don't know all
the facts.

For example, if the file is coded in joint stereo mode, the
bandwidth that is allocated to the L+R data is usually far
greater than the bandwidth allocated to the L-R data. This
can work because there may not be a lot of useful
information in the L-R domain. IOW, you have a single person
speaking close to a microphone and/or in a dead space, so
the information is essentially mono.

What you are suggesting is do away with the L-R channel
which makes a lot of sense for a single person speaking.

But, if your MP3 coder is already allocating most of the
available bandwidth to the L+R data, then downconverting to
mono won't save a lot of bandwidth.


The only way I know how to "reconstruct" a file as mono
is to "re-sample" or "re-encode" it. As I said, I don't
want to lose sound quality through the re-sampling
process.

Yes, transcodeing MP3 is far from a slam dunk.

I am a bit disapointed that no one seems to know the
answer to my question, so please let me state the
question again. Is it possible to extract one channel
into a mono mp3 file without losing any sound quality?

Not necessarily. Some MP3 coders are *smart* enough to
figure out that the audio is essentially mono, and code it
accordingly.

Thanks for all the input. I now have a few ideas about where to go
with this. But for now I will probably just keep the files as they
are.

I knew that mp3 encoding was complicated but I found out that it was
even more complicated than I thought.

R2

"R2" wrote in message
oups.com
Thanks for all the input. I now have a few ideas about
where to go with this. But for now I will probably just
keep the files as they are.

I knew that mp3 encoding was complicated but I found out
that it was even more complicated than I thought.


Huzzah! That was the message that several of us were trying
to communicate.

Enjoy!

On 14 Sep 2005 21:01:28 -0700, wrote:

I have been looking around for weeks to find an mp3 converter than can
"extract" one of the channels from a stereo mp3 file. I want to make
new mono files with the same sound quality but half the data size. I
really don't want to re-encode or re-sample anything! So far I have
had no luck finding what I need. Is it possible to do this?

The process you describe is called "Re-encoding as a mono MP3". If
you are dead set against re-encoding, you're going to find it
difficult :-)

Open the MP3 in an audio editor. Save As, choosing appropriate MP3
settings.

On 16 Sep 2005 16:37:57 -0700, "R2" wrote:

I am a bit disapointed that no one seems to know the answer to my
question, so please let me state the question again. Is it possible to
extract one channel into a mono mp3 file without losing any sound
quality?

If you re-encode at the same bit-rate you'll lose very little if any
quality. It's only a MP3 for goodness' sake! If you'd wanted
quality you'd have recorded a wav.

On 16 Sep 2005 17:57:19 -0700, "R2" wrote:

I fully understand what you are saying about LOSSY compression. That
is exactly why I was hoping for a way to extract a channel without
re-sampling it. The original file sounds OK. I just don't want to
cause any more loss.

Have you tried? I think you'll discover your fears are unfounded.

On Sat, 17 Sep 2005 01:16:02 GMT, Chris Hornbeck
wrote:

I guess it's just a sign of the national emotional
numbness in the wake of the disater that you haven't
received a half dozen posts to the effect of:

What nation? :-)

On 16 Sep 2005 22:04:30 -0400, (Scott Dorsey) wrote:

What you are missing is that the lossy compression scheme basically
throws away information that isn't important... and if both channels
are the same and the encoder is set up for coupled stereo, it will
not bother storing any L-R stereo information in the file because
there will be nothing to store.

Does MP3 compression analyse for stereo content before compressing?
Which encoder are you using?

On 18 Sep 2005 12:47:19 -0700, "R2" wrote:

Thanks for all the input. I now have a few ideas about where to go
with this. But for now I will probably just keep the files as they
are.

I knew that mp3 encoding was complicated but I found out that it was
even more complicated than I thought.

Except that, for your application regarding a mono speech broadcast,it
really isn't complicated at all. Just do it. You started out with
adequate, but not wonderful quality. You'll end up with the same.

If this was a quality music recording we'd first be slamming you for
going anywhere near MP3 in the first place, then suggesting ways of
minimising further damage. But speech is very undemanding.

Laurence Payne wrote:
Does MP3 compression analyse for stereo content before compressing?

It does what you ask it to!

If the input is stereo, MP3 compression can do one of three things,
usually all selectable by the user:
(1) make a mono mp3 file by encoding L+R
(2) make a true stereo MP3 by encoding L and R independently
(3) Joint stereo, which capitalizes on the inherent redundancy of stereo
audio, i.e. the L and R channels have similar content.

The "Lame" encoder by default selects joint stereo for any bit rate
below some threshold (160k I think) and normal stereo otherwise.

Anahata

"R2" r2mhf@
I am a bit disapointed that no one seems to know the answer to my
question, so please let me state the question again. Is it possible to
extract one channel into a mono mp3 file without losing any sound
quality?

In Cool Edit/Adobe Audition you can load a stereo file and select
EditEdit Channel(Left/Right) Channel. After that, Copy and Paste to a New
file.

Job done, One channel of your original stereo, as a seperate Mono track.

What software do you have? I've not followed this one closely as I can't
remember the details.

Rv!

Just for the record I have tried using both Cool Edit and Adobe
Audition on this task. Both of them automatically convert an mp3 to
wav when you load the file. Therefore the only way to edit an mp3 is
to convert to wav and back to mp3. I could find no way of changing
this setting. I can't really complain about that though, both are sold
was wav editors and not mp3 editors.

R2

On 20 Sep 2005 21:55:53 -0700, "R2" wrote:

Just for the record I have tried using both Cool Edit and Adobe
Audition on this task. Both of them automatically convert an mp3 to
wav when you load the file. Therefore the only way to edit an mp3 is
to convert to wav and back to mp3. I could find no way of changing
this setting. I can't really complain about that though, both are sold
was wav editors and not mp3 editors.

You may be interested in the editors on mp3fe.com .

Their freeware .mp3 to .wav convertor works better than
most although it balks at flakey sources. (The temp
file it leaves is still often usable.)

Good fortune,

Chris Hornbeck

R2 wrote:
Just for the record I have tried using both Cool Edit and Adobe
Audition on this task. Both of them automatically convert an mp3 to
wav when you load the file. Therefore the only way to edit an mp3 is
to convert to wav and back to mp3. I could find no way of changing
this setting.

Probably because there isn't any better way of doing it. :-)


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

On 20 Sep 2005 21:55:53 -0700, "R2" wrote:

Just for the record I have tried using both Cool Edit and Adobe
Audition on this task. Both of them automatically convert an mp3 to
wav when you load the file. Therefore the only way to edit an mp3 is
to convert to wav and back to mp3. I could find no way of changing
this setting. I can't really complain about that though, both are sold
was wav editors and not mp3 editors.

ARE there editors that work directly on MP3 files, for any process
beyond a bit of simple cutting and splicing perhaps? It's like
asking Word to work on a file in a ZIP archive without extracting it.
The extraction and re-compression may be transparent to the user, but
it still has to happen.

Laurence Payne wrote:

ARE there editors that work directly on MP3 files, for any process
beyond a bit of simple cutting and splicing perhaps?

I believe it's possible to change the volume level of an MP3 by editing
a scale factor that's part of each encoding block, with some
limitations, and there are numerous utilities for editing ID tags.

As far as I know, anything else requires decoding and re-encoding.

Anahata