good mp3 batch encoder for true stereo separation?

[Mad Scientist]

I MP3'd a recording that was originally fully stereo separated, using
MusicMatch Jukebox 6. Somehow in the encoding process, it got turned
into semi-stereo (not quite mono), which can be heard listening to the
original WAV file and then the mp3. I think it is called "joint
stereo" which is horrifying to an audiophile like me, because now I
realize that all my MP3'd recordings may have LOST their STEREO
SEPARATION!!! ARGH!!! Am I going to have to re-mp3 them? Or is this
the player doing this?

What would be the one or two best (affordable) MP3 encoders on the
market that result in great audio quality (archive quality?) and can
encode entire batches of files.

Any advice welcome.

[Lionel]

Mad Scientist wrote:
> I MP3'd a recording that was originally fully stereo separated, using
> MusicMatch Jukebox 6. Somehow in the encoding process, it got turned
> into semi-stereo (not quite mono), which can be heard listening to the
> original WAV file and then the mp3. I think it is called "joint
> stereo" which is horrifying to an audiophile like me, because now I
> realize that all my MP3'd recordings may have LOST their STEREO
> SEPARATION!!! ARGH!!! Am I going to have to re-mp3 them? Or is this
> the player doing this?
>
> What would be the one or two best (affordable) MP3 encoders on the
> market that result in great audio quality (archive quality?) and can
> encode entire batches of files.
>
> Any advice welcome.

L.A.M.E.
http://lame.sourceforge.net/

You're welcome

[Arny Krueger]

"Mad Scientist" > wrote in message
om...

> I MP3'd a recording that was originally fully stereo separated, using
> MusicMatch Jukebox 6. Somehow in the encoding process, it got turned
> into semi-stereo (not quite mono), which can be heard listening to the
> original WAV file and then the mp3. I think it is called "joint
> stereo" which is horrifying to an audiophile like me, because now I
> realize that all my MP3'd recordings may have LOST their STEREO
> SEPARATION!!! ARGH!!! Am I going to have to re-mp3 them?

Probably, yes.

>Or is this the player doing this?

Probably, no.

> What would be the one or two best (affordable) MP3 encoders on the
> market that result in great audio quality (archive quality?) and can
> encode entire batches of files.

You just might already have it. Review your MusicMatch MP3 encoding
parameters and change as required.

"Joint Stereo" doesn't totally destroy separation if you're also using a
respectable bit rate. However the defaults with programs like MusicMatch are
generally chosen to save disk space. The defaults don't do the best
possible job of preserving the sound quality of music, by audiophile
standards. Remember that about 95%+ of the people who download this software
use computer speakers and background music as their preferred listening
environment. Not my cup of tea and probably not yours.

The *minimum* acceptable bit rate for MP3s is 112 bps while 128 is the most
common, but many audiophiles report going up to 192 bps or higher to get
what they find to be acceptable results. I have personally yet to find a MP3
coder that I consider to be transparent enough for the highest quality use.
But, the utilitarian aspects of MP3 processing can't be denied.

I've put my money where my mouth is - I'm building a collection of .wav
files which are totally uncompressed bit-perfect representations of CDs I
own, on relatively large hard drive arrays. I prefer using a portable player
(20 GB Nomad Jukebox 3) that can effectively load and play .wav files.

OTOH when top quality is not of the essence, I've been known to listen to a
few MP3s and concentrate on the music, not just audio values. Therefore, I
also have a 128 meg Nomad 2 flash memory player and also an portable CD
player that can play MP3 CDs.

Convenience is fine in its place, but I very much prefer my Nomad Jukebox
and its rich selection of .wav files.

[ff123]

On 6 Nov 2003 19:27:32 -0800, (Mad Scientist)
wrote:

>I MP3'd a recording that was originally fully stereo separated, using
>MusicMatch Jukebox 6. Somehow in the encoding process, it got turned
>into semi-stereo (not quite mono), which can be heard listening to the
>original WAV file and then the mp3. I think it is called "joint
>stereo" which is horrifying to an audiophile like me, because now I
>realize that all my MP3'd recordings may have LOST their STEREO
>SEPARATION!!! ARGH!!! Am I going to have to re-mp3 them? Or is this
>the player doing this?
>
>What would be the one or two best (affordable) MP3 encoders on the
>market that result in great audio quality (archive quality?) and can
>encode entire batches of files.
>
>Any advice welcome.

Upgrade your MMJB to at least 6.1. Version 6.0 has a stereo
separation bug. See:

http://ff123.net/training/training.html#mc_sich_short

for an example of what this bug can sound like.

But for best quality mp3, use the lame mp3 encoder with the
--alt-preset standard setting.

http://doc.hydrogenaudio.org/wikis/hydrogenaudio/RecommendedLAME

http://doc.hydrogenaudio.org/wikis/hydrogenaudio/LameCompiles

ff123

[Browntimdc]

ff123 > wrote in news:rhtnqvs4cqvcspsb1ffbpi423tp2bh83v7
@4ax.com:

> On 6 Nov 2003 19:27:32 -0800, (Mad Scientist)
> wrote:
>
>>I MP3'd a recording that was originally fully stereo separated, using
>>MusicMatch Jukebox 6. Somehow in the encoding process, it got turned
>>into semi-stereo (not quite mono), which can be heard listening to the
>>original WAV file and then the mp3. I think it is called "joint
>>stereo" which is horrifying to an audiophile like me, because now I
>>realize that all my MP3'd recordings may have LOST their STEREO
>>SEPARATION!!! ARGH!!! Am I going to have to re-mp3 them? Or is this
>>the player doing this?

Joint stereo doesn't discernably reduce seperation once you eliminate the
bug mentioned below.

>>
>>What would be the one or two best (affordable) MP3 encoders on the
>>market that result in great audio quality (archive quality?) and can
>>encode entire batches of files.
>>
>>Any advice welcome.
>
> Upgrade your MMJB to at least 6.1. Version 6.0 has a stereo
> separation bug.
[snip]
> But for best quality mp3, use the lame mp3 encoder with the
> --alt-preset standard setting.
>
> http://doc.hydrogenaudio.org/wikis/hydrogenaudio/RecommendedLAME
>
> http://doc.hydrogenaudio.org/wikis/hydrogenaudio/LameCompiles
>
> ff123
>

Also, look for Exact Audio Copy (EAC), which puts a nice interface on
LAME.

Tim

--

"The strongest human instinct is to impart information,
and the second strongest is to resist it."

Kenneth Graham

[Lionel]

Browntimdc wrote:
> ff123 > wrote in news:rhtnqvs4cqvcspsb1ffbpi423tp2bh83v7
> @4ax.com:
>
>
>>On 6 Nov 2003 19:27:32 -0800, (Mad Scientist)
>>wrote:
>>
>>
>>>I MP3'd a recording that was originally fully stereo separated, using
>>>MusicMatch Jukebox 6. Somehow in the encoding process, it got turned
>>>into semi-stereo (not quite mono), which can be heard listening to the
>>>original WAV file and then the mp3. I think it is called "joint
>>>stereo" which is horrifying to an audiophile like me, because now I
>>>realize that all my MP3'd recordings may have LOST their STEREO
>>>SEPARATION!!! ARGH!!! Am I going to have to re-mp3 them? Or is this
>>>the player doing this?
>
>
> Joint stereo doesn't discernably reduce seperation once you eliminate the
> bug mentioned below.
>
>
>>>What would be the one or two best (affordable) MP3 encoders on the
>>>market that result in great audio quality (archive quality?) and can
>>>encode entire batches of files.
>>>
>>>Any advice welcome.
>>
>>Upgrade your MMJB to at least 6.1. Version 6.0 has a stereo
>>separation bug.
>
> [snip]
>
>>But for best quality mp3, use the lame mp3 encoder with the
>>--alt-preset standard setting.
>>
>>http://doc.hydrogenaudio.org/wikis/hydrogenaudio/RecommendedLAME
>>
>>http://doc.hydrogenaudio.org/wikis/hydrogenaudio/LameCompiles
>>
>>ff123
>>
>
>
> Also, look for Exact Audio Copy (EAC), which puts a nice interface on
> LAME.
>
> Tim
>
Cdex ?

[Lionel]

Lionel wrote:

> Browntimdc wrote:
>
>> ff123 > wrote in news:rhtnqvs4cqvcspsb1ffbpi423tp2bh83v7
>> @4ax.com:
>>
>>
>>> On 6 Nov 2003 19:27:32 -0800, (Mad Scientist)
>>> wrote:
>>>
>>>
>>>> I MP3'd a recording that was originally fully stereo separated, using
>>>> MusicMatch Jukebox 6. Somehow in the encoding process, it got turned
>>>> into semi-stereo (not quite mono), which can be heard listening to the
>>>> original WAV file and then the mp3. I think it is called "joint
>>>> stereo" which is horrifying to an audiophile like me, because now I
>>>> realize that all my MP3'd recordings may have LOST their STEREO
>>>> SEPARATION!!! ARGH!!! Am I going to have to re-mp3 them? Or is this
>>>> the player doing this?
>>
>>
>>
>> Joint stereo doesn't discernably reduce seperation once you eliminate
>> the bug mentioned below.
>>
>>
>>>> What would be the one or two best (affordable) MP3 encoders on the
>>>> market that result in great audio quality (archive quality?) and can
>>>> encode entire batches of files.
>>>>
>>>> Any advice welcome.
>>>
>>>
>>> Upgrade your MMJB to at least 6.1. Version 6.0 has a stereo
>>> separation bug.
>>
>>
>> [snip]
>>
>>> But for best quality mp3, use the lame mp3 encoder with the
>>> --alt-preset standard setting.
>>>
>>> http://doc.hydrogenaudio.org/wikis/hydrogenaudio/RecommendedLAME
>>>
>>> http://doc.hydrogenaudio.org/wikis/hydrogenaudio/LameCompiles
>>>
>>> ff123
>>>
>>
>>
>> Also, look for Exact Audio Copy (EAC), which puts a nice interface on
>> LAME.
>>
>> Tim
>>
> Cdex ?
>
Ooops !
http://www.cdex.n3.net/

[KikeG]

"Arny Krueger" > wrote in message >...

>I have personally yet to find a MP3
> coder that I consider to be transparent enough for the highest quality use.

The thing is, no psychoacoustic lossy compression scheme gives audibly
perfect results for everything.

For example, LAME 3.90.3 with the "--alt-preset standard" preset gives
transparent quality for most things, but there are still some
pathological cases that expose its flaws (being the castanets and
trumpets1 samples at your PCABX site some of these cases). Using the
"--alt-preset insane" preset cures some of these cases, but not all,
at the expense of raising the bitrate up to 320 kbps. Still, those
flaws are audibly quite subtle, and many times trained ears are
required to hear them.

Other newer and higher quality lossy audio codecs, such as AAC and
specially Musepack (MPC), have an even smaller repertoire of problem
cases (Musepack does fine with PCABX samples at around 190 kbps), but
still have a few of those, even when raising the bitrate to insane
levels. That's why I say that, at least currently, no psychoacoustic
lossy compresion scheme gives always audibly perfect results. However,
for most cases (all but quite rare cases with codecs such as
Musepack), quality is undistinguishable even for well trained ears.

[Arny Krueger]

"KikeG" > wrote in message
om

> "Arny Krueger" > wrote in message
> >...

>> I have personally yet to find a MP3
>> coder that I consider to be transparent enough for the highest
>> quality use.

> The thing is, no psychoacoustic lossy compression scheme gives audibly
> perfect results for everything.

That seems to be, quire regrettably, the current state of the art. There is
still room for progress.

> For example, LAME 3.90.3 with the "--alt-preset standard" preset gives
> transparent quality for most things, but there are still some
> pathological cases that expose its flaws (being the castanets and
> trumpets1 samples at your PCABX site some of these cases). Using the
> "--alt-preset insane" preset cures some of these cases, but not all,
> at the expense of raising the bitrate up to 320 kbps. Still, those
> flaws are audibly quite subtle, and many times trained ears are
> required to hear them.

Yes, on balance the more subtle flaws are hard to hear. Sometimes they can
only be heard in quick, side-by-side comparisons. Sometimes only certain
listeners can hear them and then only with difficulty. Sometimes
considerable training and familiarization is required before even that.

The PCABX castanets sample is really quite grueling. I don't think that
vinyl or high speed analog tape can handle it, either. In fact, I suspect
that vinyl would fail pretty obviously. It's quite a torture test involving
a rich collection of impulses and high harmonics. On the face of it, the
trumpets selection seems like it should be easy. Mostly midrange, and
largely steady-state. However to be reproduced transparently, frequency
response in the upper midrange and treble has to be "just right". Again I
doubt that vinyl or analog tape could handle it transparently. On balance,
high-performance perceptual coding might outperform both kinds of legacy
media in a ABX test.

OTOH, leaving files as .wav files and/or using lossless compression is a far
more viable option given that the costs and size considerations associated
with storing data continue to improve dramatically. I think I did my first
MP3 listening on modern for the day computers with 1 GB hard drives. Today
80 GB hard drives are more-or-less standard on new PCs. For the longest time
we were limited to what we could burn on a CD - 550 megs. Today we can burn
DVD with equal-or-better speed, convenience, and media costs - 8-9 times as
much data.

> Other newer and higher quality lossy audio codecs, such as AAC and
> specially Musepack (MPC), have an even smaller repertoire of problem
> cases (Musepack does fine with PCABX samples at around 190 kbps), but
> still have a few of those, even when raising the bitrate to insane
> levels. That's why I say that, at least currently, no psychoacoustic
> lossy compression scheme gives always audibly perfect results. However,
> for most cases (all but quite rare cases with codecs such as
> Musepack), quality is undistinguishable-le even for well trained ears.

Thus lossy compression while providing highly attractive practical
advantages, and de facto acceptable to millions of music lovers in the
mainstream, still does not meet the standards of the audio perfectionist.

[Arny Krueger]

"Lionel" > wrote in message

> Lionel wrote:
>
>> Browntimdc wrote:
>>
>>> ff123 > wrote in
>>> news:rhtnqvs4cqvcspsb1ffbpi423tp2bh83v7 @4ax.com:
>>>
>>>
>>>> On 6 Nov 2003 19:27:32 -0800, (Mad Scientist)
>>>> wrote:
>>>>
>>>>
>>>>> I MP3'd a recording that was originally fully stereo separated,
>>>>> using MusicMatch Jukebox 6. Somehow in the encoding process, it
>>>>> got turned into semi-stereo (not quite mono), which can be heard
>>>>> listening to the original WAV file and then the mp3. I think it
>>>>> is called "joint stereo" which is horrifying to an audiophile
>>>>> like me, because now I realize that all my MP3'd recordings may
>>>>> have LOST their STEREO SEPARATION!!! ARGH!!! Am I going to have
>>>>> to re-mp3 them? Or is this the player doing this?
>>>
>>>
>>>
>>> Joint stereo doesn't discernably reduce seperation once you
>>> eliminate the bug mentioned below.
>>>
>>>
>>>>> What would be the one or two best (affordable) MP3 encoders on the
>>>>> market that result in great audio quality (archive quality?) and
>>>>> can encode entire batches of files.
>>>>>
>>>>> Any advice welcome.
>>>>
>>>>
>>>> Upgrade your MMJB to at least 6.1. Version 6.0 has a stereo
>>>> separation bug.
>>>
>>>
>>> [snip]
>>>
>>>> But for best quality mp3, use the lame mp3 encoder with the
>>>> --alt-preset standard setting.
>>>>
>>>> http://doc.hydrogenaudio.org/wikis/hydrogenaudio/RecommendedLAME
>>>>
>>>> http://doc.hydrogenaudio.org/wikis/hydrogenaudio/LameCompiles
>>>>
>>>> ff123
>>>>
>>>
>>>
>>> Also, look for Exact Audio Copy (EAC), which puts a nice interface
>>> on LAME.
>>>
>>> Tim
>>>
>> Cdex ?
>>
> Ooops !
> http://www.cdex.n3.net/

CDEX and EAC are in my experience equally accurate and easy-to-use rippers.
However, there still seem to be a few machines where one works well, and the
other works not-so-well. Therefore, there is a clear place for both of them.

[Sockpuppet Yustabe]

"Arny Krueger" > wrote in message
...
>
> The PCABX castanets sample is really quite grueling.

I find it quite enjoyable. It's my favorite.




----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---

[ScottW]

"Arny Krueger" > wrote in message
...
>
> OTOH, leaving files as .wav files and/or using lossless compression is a
far
> more viable option given that the costs and size considerations
associated
> with storing data continue to improve dramatically. I think I did my
first
> MP3 listening on modern for the day computers with 1 GB hard drives.
Today
> 80 GB hard drives are more-or-less standard on new PCs. For the longest
time
> we were limited to what we could burn on a CD - 550 megs. Today we can
burn
> DVD with equal-or-better speed, convenience, and media costs - 8-9 times
as
> much data.

I believe the need to use compression in the home will be completely
eliminated
in a few years.

http://newscenter.verizon.com/proactive/newsroom/release.vtml?id=80434

Actually, the need for home media storage may also be eliminated.

>
> > Other newer and higher quality lossy audio codecs, such as AAC and
> > specially Musepack (MPC), have an even smaller repertoire of problem
> > cases (Musepack does fine with PCABX samples at around 190 kbps), but
> > still have a few of those, even when raising the bitrate to insane
> > levels. That's why I say that, at least currently, no psychoacoustic
> > lossy compression scheme gives always audibly perfect results. However,
> > for most cases (all but quite rare cases with codecs such as
> > Musepack), quality is undistinguishable-le even for well trained ears.
>
> Thus lossy compression while providing highly attractive practical
> advantages, and de facto acceptable to millions of music lovers in the
> mainstream, still does not meet the standards of the audio perfectionist.

Still, compression will continue to be required for the mobil wireless
users for many years to come.

ScottW

[Mad Scientist]

Thank you for you response.

I'll have to play with the musicmatch encoding settings.
Currently they are 160 bps, all other settings default,
but clicking Options > Settings > Advanced gives some
MP3 encoding settings such as Processing Level which
is currently at Normal. I think Very High might preserve
the stereo separation 100%, will have to see.
There may be other parameters to tweak as well,
that resulted in some loss of stereo separation.

Anyone have any ideas?

I have all my audio as WAV files also,
but the size makes it prohibitive to making multi backup copies.
I'll have to explore the settings and other codecs (such as LAME)
and go a little higher on the bps (prolly 192).

Thanks to everyone else also for your responses.