Poll: WMA vs. MP3

[Radium]

Hi group:

Lets say there is a 44.1 khz, monoaural, 32 kbps WMA file and a 44.1
khz, monoaural, 32 kbps MP3 file; which one would rather listen to?

I'd prefer the WMA. I readily notice the difference in WMA artifacts
and MP3 artifacts; the differences are very difficult for me to
describe, but they are signficant. Both resemble "digital tones" of old
video games and binary signals but they are noticeably different from
each other.

BTW, what is responsible for those differences in audio artifacts
resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?


Thanks,

Radium

[Don Pearce]

On 5 Nov 2006 00:46:06 -0800, "Radium" > wrote:

>Hi group:
>
>Lets say there is a 44.1 khz, monoaural, 32 kbps WMA file and a 44.1
>khz, monoaural, 32 kbps MP3 file; which one would rather listen to?
>
>I'd prefer the WMA. I readily notice the difference in WMA artifacts
>and MP3 artifacts; the differences are very difficult for me to
>describe, but they are signficant. Both resemble "digital tones" of old
>video games and binary signals but they are noticeably different from
>each other.
>
>BTW, what is responsible for those differences in audio artifacts
>resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?
>
>
>Thanks,
>
>Radium

Why would you listen to either when AAC+ is out there?

www.tuner2.com

Needs Winamp

d

--
Pearce Consulting
http://www.pearce.uk.com

[TT]

"Radium" > wrote in message
oups.com...
> Hi group:
>
> Lets say there is a 44.1 khz, monoaural, 32 kbps WMA file and a 44.1
> khz, monoaural, 32 kbps MP3 file; which one would rather listen to?
>
> I'd prefer the WMA. I readily notice the difference in WMA artifacts
> and MP3 artifacts; the differences are very difficult for me to
> describe, but they are signficant. Both resemble "digital tones" of old
> video games and binary signals but they are noticeably different from
> each other.
>
> BTW, what is responsible for those differences in audio artifacts
> resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?
>
>
> Thanks,
>
> Radium
>

A better question would be "Who cares?" Anything under 1411kbs you can keep
;-)

Regards TT

[Laurence Payne]

On 5 Nov 2006 00:46:06 -0800, "Radium" > wrote:

>Lets say there is a 44.1 khz, monoaural, 32 kbps WMA file and a 44.1
>khz, monoaural, 32 kbps MP3 file; which one would rather listen to?
>
>I'd prefer the WMA. I readily notice the difference in WMA artifacts
>and MP3 artifacts; the differences are very difficult for me to
>describe, but they are signficant. Both resemble "digital tones" of old
>video games and binary signals but they are noticeably different from
>each other.
>
>BTW, what is responsible for those differences in audio artifacts
>resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?


Both systems make a remarkably good attempt at the impossible -
squashing audio to a small file size while retaining quality. They
both do this by discarding "redundant" information, and both think
their definition of redundant is the better one!

At 32 kbps both are going to sound crap. But differently crap. You
would only use this sort of rate for an application where quality
didn't matter but small file-size did. Like a telephone answering
system.

Over 128 kbps both systems can sound acceptable, and the differences
will be less noticeable.

[Laurence Payne]

On Sun, 05 Nov 2006 08:55:22 GMT, (Don Pearce)
wrote:

>Why would you listen to either when AAC+ is out there?
>
>www.tuner2.com
>
>Needs Winamp


If it's staying on your system, why throw away quality by compressing
at all? If it needs to be portable, you must use a format everyone
can play without installing special software.

[Don Pearce]

On Sun, 05 Nov 2006 11:34:13 +0000, Laurence Payne
<lpayne1NOSPAM@dslDOTpipexDOTcom> wrote:

>On Sun, 05 Nov 2006 08:55:22 GMT, (Don Pearce)
>wrote:
>
>>Why would you listen to either when AAC+ is out there?
>>
>>www.tuner2.com
>>
>>Needs Winamp
>
>
>If it's staying on your system, why throw away quality by compressing
>at all? If it needs to be portable, you must use a format everyone
>can play without installing special software.

If it is staying on my system, as you say it stays in native format.
This is streaming, which is another matter entirely, and needs
compression. AAC+ is best of breed at this moment.

d

--
Pearce Consulting
http://www.pearce.uk.com

[Arny Krueger]

"Radium" > wrote in message
oups.com
> Hi group:
>
> Lets say there is a 44.1 khz, monoaural, 32 kbps WMA file
> and a 44.1 khz, monoaural, 32 kbps MP3 file; which one
> would rather listen to?

IME WMA holds together better at too-low low bitrates.

> I'd prefer the WMA. I readily notice the difference in
> WMA artifacts and MP3 artifacts; the differences are very
> difficult for me to describe, but they are signficant.

I've found that the MP3 decoder in the current Windows Media Player does a
bettter-than-average job on low-bitrate MP3s.

> BTW, what is responsible for those differences in audio
> artifacts resulting from a low-bit-rate WMAs vs.
> low-bit-rate MP3s?

It is my understanding that every MP3 file has to be decoded by the same
decoder, and that the common MP3 decoder implements a more-or-less common
set of strategies, no matter what the bitrate is. It is my understanding
that the WMA technology is more diverse and shifts strategies, depending on
bitrate.

[Radium]

Don Pearce wrote:
> On Sun, 05 Nov 2006 11:34:13 +0000, Laurence Payne
> <lpayne1NOSPAM@dslDOTpipexDOTcom> wrote:
>
> >On Sun, 05 Nov 2006 08:55:22 GMT, (Don Pearce)
> >wrote:
> >
> >>Why would you listen to either when AAC+ is out there?
> >>
> >>www.tuner2.com
> >>
> >>Needs Winamp
> >
> >
> >If it's staying on your system, why throw away quality by compressing
> >at all? If it needs to be portable, you must use a format everyone
> >can play without installing special software.
>
> If it is staying on my system, as you say it stays in native format.
> This is streaming, which is another matter entirely, and needs
> compression. AAC+ is best of breed at this moment.
>
> d
>
> --
> Pearce Consulting
> http://www.pearce.uk.com

IMHO, AAC+ is good only if its at least 44.1 khz, monoaural, and with a
bit-rate of 320 kbps or more. Otherwise, it reeks of disgusting
artifacts much like MP3 and other non-WMA compression schemes.

The only two audio format I prefer are uncompressed PCM [such as WAV
files and CD audio] with a sample-rate of at least 44.1 khz, monoaural,
and a bit-resolution or 16-bit or more and WMA-audio whose sample-rate
is at least 44.1 khz [and should be exactly the same sample-rate as it
was before it was compressed to WMA] and monoaural. I don't like
stereo.

Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
Adobe Audition to convert the stereo file to mono. But its more
complicated than just stereo-to-mono conversion. First, I make two
copies the stereo file. With one copy, I invert the phase of the left
channel and then combine it with the right channel -- this causes
whatever was identical the left and right channel to be cancelled.
Before I combine these two channels, I decrease the loudness of each
channel by 50%.

With the other copy of the stereo file, I don't invert either the left
or right channel. I simply decrease the loudness of both channels by
77.5% and then I combine the left and right together to make a mono
file.

I then combine the audio of both copies together in a new WAV file.
Voila! I get some nice monoaural audio with at least 44.1 khz and at
least 16-bit. If I want to place this audio file on my "Yahoo group's"
website [which has a 20 MB limit], I make a monoaural WMA copy of the
WAV file. I make sure to encode the WMA in 20 kbps and 44.1 khz. The
WMAs sound excellent for their awesomely small file size.

Now would I do the same with MP3, AAC+, or any non-WMA type of audio
compression? No way!

[Don Pearce]

On 5 Nov 2006 13:59:33 -0800, "Radium" > wrote:

>
>Don Pearce wrote:
>> On Sun, 05 Nov 2006 11:34:13 +0000, Laurence Payne
>> <lpayne1NOSPAM@dslDOTpipexDOTcom> wrote:
>>
>> >On Sun, 05 Nov 2006 08:55:22 GMT, (Don Pearce)
>> >wrote:
>> >
>> >>Why would you listen to either when AAC+ is out there?
>> >>
>> >>www.tuner2.com
>> >>
>> >>Needs Winamp
>> >
>> >
>> >If it's staying on your system, why throw away quality by compressing
>> >at all? If it needs to be portable, you must use a format everyone
>> >can play without installing special software.
>>
>> If it is staying on my system, as you say it stays in native format.
>> This is streaming, which is another matter entirely, and needs
>> compression. AAC+ is best of breed at this moment.
>>
>> d
>>
>> --
>> Pearce Consulting
>> http://www.pearce.uk.com
>
>IMHO, AAC+ is good only if its at least 44.1 khz, monoaural, and with a
>bit-rate of 320 kbps or more. Otherwise, it reeks of disgusting
>artifacts much like MP3 and other non-WMA compression schemes.
>
>The only two audio format I prefer are uncompressed PCM [such as WAV
>files and CD audio] with a sample-rate of at least 44.1 khz, monoaural,
>and a bit-resolution or 16-bit or more and WMA-audio whose sample-rate
>is at least 44.1 khz [and should be exactly the same sample-rate as it
>was before it was compressed to WMA] and monoaural. I don't like
>stereo.
>
>Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
>Adobe Audition to convert the stereo file to mono. But its more
>complicated than just stereo-to-mono conversion. First, I make two
>copies the stereo file. With one copy, I invert the phase of the left
>channel and then combine it with the right channel -- this causes
>whatever was identical the left and right channel to be cancelled.
>Before I combine these two channels, I decrease the loudness of each
>channel by 50%.
>
>With the other copy of the stereo file, I don't invert either the left
>or right channel. I simply decrease the loudness of both channels by
>77.5% and then I combine the left and right together to make a mono
>file.
>
>I then combine the audio of both copies together in a new WAV file.
>Voila! I get some nice monoaural audio with at least 44.1 khz and at
>least 16-bit. If I want to place this audio file on my "Yahoo group's"
>website [which has a 20 MB limit], I make a monoaural WMA copy of the
>WAV file. I make sure to encode the WMA in 20 kbps and 44.1 khz. The
>WMAs sound excellent for their awesomely small file size.
>
>Now would I do the same with MP3, AAC+, or any non-WMA type of audio
>compression? No way!

Anybody?

d

--
Pearce Consulting
http://www.pearce.uk.com

[George M. Middius]

Don Pearce said:

> >Now would I do the same with MP3, AAC+, or any non-WMA type of audio
> >compression? No way!

> Anybody?

Will it be on the test?





--

Krooscience: The antidote to education, experience, and excellence.

[Geoff]

Radium wrote:
>
> BTW, what is responsible for those differences in audio artifacts
> resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?


Just the different compression algorythmns. Similar (maybe lesser)
differences between different MP3 encoders.

geoff

[Radium]

Geoff wrote:
> Radium wrote:
> >
> > BTW, what is responsible for those differences in audio artifacts
> > resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?
>
>
> Just the different compression algorythmns. Similar (maybe lesser)
> differences between different MP3 encoders.
>
> geoff

Is there a website or book where I can find the mechanisms of what
causes WMAs and MP3s to differ in terms of their audio artifacts? I've
looked as hard as I can without any luck so far.

Also, does anyone have a valid email address so I can send them some
WMA and MP3 files to show them what I am talking about? I have the same
song with one file in WMA and the other in MP3. If anyone's interested.

[Eeyore]

Radium wrote:

> Geoff wrote:
> > Radium wrote:
> > >
> > > BTW, what is responsible for those differences in audio artifacts
> > > resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?
> >
> >
> > Just the different compression algorythmns. Similar (maybe lesser)
> > differences between different MP3 encoders.
> >
> > geoff
>
> Is there a website or book where I can find the mechanisms of what
> causes WMAs and MP3s to differ in terms of their audio artifacts? I've
> looked as hard as I can without any luck so far.
>
> Also, does anyone have a valid email address so I can send them some
> WMA and MP3 files to show them what I am talking about? I have the same
> song with one file in WMA and the other in MP3. If anyone's interested.

Check out sub-band codecs.
http://www.google.com/search?client=opera&rls=en&q=sub-band+codec&sourceid=opera&ie=utf-8&oe=utf-8

Both are examples of this as is the compression used by Sony in Mini-Disc and
Philips in their now deceased DCC.

Various types exist - often to avoid paying royalties.


Graham

[Arny Krueger]

"Radium" > wrote in message
ups.com
> Geoff wrote:
>> Radium wrote:
>>>
>>> BTW, what is responsible for those differences in audio
>>> artifacts resulting from a low-bit-rate WMAs vs.
>>> low-bit-rate MP3s?

>> Just the different compression algorythmns. Similar
>> (maybe lesser) differences between different MP3
>> encoders.

Agreed.

> Is there a website or book where I can find the
> mechanisms of what causes WMAs and MP3s to differ in
> terms of their audio artifacts? I've looked as hard as I
> can without any luck so far.

The details of the processes are proprietary. The details of the standard
MP3 decoder is known to all who want to know, but the encoders are all over
the map. The whole WMA process is proprietary, encoder, decoder, the whole
ball of wax.

> Also, does anyone have a valid email address so I can
> send them some WMA and MP3 files to show them what I am
> talking about? I have the same song with one file in WMA
> and the other in MP3. If anyone's interested.

arnyk at comcast.net is there for you.

[Richard Crowley]

"Don Pearce" wrote ...
> Anybody?

Do whatever sounds good for the content at hand
and fits within your bandwidth/space budget.

[Walt]

Don Pearce wrote:

> On 5 Nov 2006 13:59:33 -0800, "Radium" > wrote:

>>Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
>>Adobe Audition to convert the stereo file to mono. But its more
>>complicated than just stereo-to-mono conversion. First, I make two
>>copies the stereo file. With one copy, I invert the phase of the left
>>channel and then combine it with the right channel -- this causes
>>whatever was identical the left and right channel to be cancelled.
>>Before I combine these two channels, I decrease the loudness of each
>>channel by 50%.
>>
>>With the other copy of the stereo file, I don't invert either the left
>>or right channel. I simply decrease the loudness of both channels by
>>77.5% and then I combine the left and right together to make a mono
>>file.
>>
>>I then combine the audio of both copies together in a new WAV file.
>>Voila! I get some nice monoaural audio with at least 44.1 khz and at
>>least 16-bit.
>
> Anybody?

Interesting approach. It appears he's making a mono sum by

..5(R-L) + .775(L+R) = 1.25(R) + .25(L)

I guess he hates the leftt channel. More cello, more bass, less violin,
less percussion. Hey, whatever floats your boat.

//Walt

[Radium]

Walt wrote:
> Don Pearce wrote:
>
> > On 5 Nov 2006 13:59:33 -0800, "Radium" > wrote:
>
> >>Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
> >>Adobe Audition to convert the stereo file to mono. But its more
> >>complicated than just stereo-to-mono conversion. First, I make two
> >>copies the stereo file. With one copy, I invert the phase of the left
> >>channel and then combine it with the right channel -- this causes
> >>whatever was identical the left and right channel to be cancelled.
> >>Before I combine these two channels, I decrease the loudness of each
> >>channel by 50%.
> >>
> >>With the other copy of the stereo file, I don't invert either the left
> >>or right channel. I simply decrease the loudness of both channels by
> >>77.5% and then I combine the left and right together to make a mono
> >>file.
> >>
> >>I then combine the audio of both copies together in a new WAV file.
> >>Voila! I get some nice monoaural audio with at least 44.1 khz and at
> >>least 16-bit.
> >
> > Anybody?
>
> Interesting approach. It appears he's making a mono sum by
>
> .5(R-L) + .775(L+R) = 1.25(R) + .25(L)
>
> I guess he hates the leftt channel. More cello, more bass, less violin,
> less percussion. Hey, whatever floats your boat.
>
> //Walt

Actually, with most stereo music, if you invert the phase of one
channel and combine with the other non-inverted, you get a mono without
the main vocals, bass, and percussion. This is because -- usually --
the lead vocals, bass, and percussion are recorded identically in both
the left and right channels. The paino, chorus, guitar, and synth pads
are usually recorded differently in the left and right channel -- hence
you won't lose these sounds if you process the audio file in the
aformentioned manner. Most "voice cancellors" use this technique. Of
course, this method assumes that the voice is recorded in the center of
the stereo channels while non-vocals are not in phase for both
channels. No good for the old old songs where everything is recorded in
mono.

[Laurence Payne]

On 6 Nov 2006 10:42:37 -0800, "Radium" > wrote:

>Actually, with most stereo music, if you invert the phase of one
>channel and combine with the other non-inverted, you get a mono without
>the main vocals, bass, and percussion. This is because -- usually --
>the lead vocals, bass, and percussion are recorded identically in both
>the left and right channels.

Have you found this technique gives useful results? Or are you
quoting theory?

[Adam Sampson]

Laurence Payne <lpayne1NOSPAM@dslDOTpipexDOTcom> writes:

> Have you found this technique gives useful results?

Yes -- although which parts you lose depends on how the music was
mixed, of course (and you'll quite often lose the main vocal track but
keep the stereo reverb that was added to it, for example). It's
usually most dramatic on early stereo recordings; try The Beatles'
"Birthday" for a good example, where all the instruments are in the
middle and the vocal parts are panned hard left and right...

It's also a useful technique for showing up MP3 joint-stereo
compression artefacts; a track that sounds all right in stereo will
often turn into a bubbly mess when you listen to the difference
between the two channels.

--
Adam Sampson > <http://offog.org/>

[Walt]

Radium wrote:

>
> Actually, with most stereo music, if you invert the phase of one
> channel and combine with the other non-inverted, you get a mono without
> the main vocals, bass, and percussion. This is because -- usually --
> the lead vocals, bass, and percussion are recorded identically in both
> the left and right channels. The paino, chorus, guitar, and synth pads
> are usually recorded differently in the left and right channel -- hence
> you won't lose these sounds if you process the audio file in the
> aformentioned manner. Most "voice cancellors" use this technique. Of
> course, this method assumes that the voice is recorded in the center of
> the stereo channels while non-vocals are not in phase for both
> channels.


>No good for the old old songs where everything is recorded in
> mono.
>

Fascinating. A mono sum technique that is incompatible with mono source
material. Please tell us more.

//Walt

[Walt]

Adam Sampson wrote:

> Laurence Payne <lpayne1NOSPAM@dslDOTpipexDOTcom> writes:
>
>
>>Have you found this technique gives useful results?
>
>
> Yes -- although which parts you lose depends on how the music was
> mixed, of course (and you'll quite often lose the main vocal track but
> keep the stereo reverb that was added to it, for example).

I guess it all depends on what one means by "useful". If you are trying
to make the things panned to the center go away, as you say, it works
(more or less).

> It's
> usually most dramatic on early stereo recordings; try The Beatles'
> "Birthday" for a good example, where all the instruments are in the
> middle and the vocal parts are panned hard left and right...

I'm not sure about that particular cut, but most of the early Beatles
recordings were intended to be mono. They were recorded on a two-track
machine so they could do overdubs. These two-track masters were never
intended to be released as-is and are not "stereo" in the sense of
providing a 3 dimensional sound stage. But, due to the overwhelming
market force demanding "stereo" recordings they were released that way,
much to the dismay of George Martin. Frankly, they're bizarre to listen
to - everthing is panned hard left or right, and sometimes a voice will
ping pong from channel to channel if the overdubbing was happening that way.

>
> It's also a useful technique for showing up MP3 joint-stereo
> compression artefacts; a track that sounds all right in stereo will
> often turn into a bubbly mess when you listen to the difference
> between the two channels.

MP3 is magic. Do not look where the magician doesn't intend for you to
look, or you'll be severely disappointed with the cheapness and
shabbiness of the trick.

//Walt

[Laurence Payne]

On Mon, 06 Nov 2006 20:45:48 +0000, Adam Sampson >
wrote:

>Yes -- although which parts you lose depends on how the music was
>mixed, of course (and you'll quite often lose the main vocal track but
>keep the stereo reverb that was added to it, for example). It's
>usually most dramatic on early stereo recordings; try The Beatles'
>"Birthday" for a good example, where all the instruments are in the
>middle and the vocal parts are panned hard left and right...

Which was my point :-) Those were recorded on a console with three
panning options - R, L or centre. Most of the music we listen to now
wasn't.

[Radium]

Laurence Payne wrote:
> On 6 Nov 2006 10:42:37 -0800, "Radium" > wrote:
>
> >Actually, with most stereo music, if you invert the phase of one
> >channel and combine with the other non-inverted, you get a mono without
> >the main vocals, bass, and percussion. This is because -- usually --
> >the lead vocals, bass, and percussion are recorded identically in both
> >the left and right channels.
>
> Have you found this technique gives useful results? Or are you
> quoting theory?

Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
Adobe Audition to convert the stereo file to mono. But its more
complicated than just stereo-to-mono conversion. First, I make two
copies the stereo file. With one copy, I invert the phase of the left
channel and then combine it with the right channel -- this causes
whatever was identical the left and right channel to be cancelled.
Before I combine these two channels, I decrease the loudness of each
channel by 50%.

With the other copy of the stereo file, I don't invert either the left
or right channel. I simply decrease the loudness of both channels by
77.5% and then I combine the left and right together to make a mono
file.

I then combine the audio of both copies together in a new WAV file.
Voila! I get some nice monoaural audio with at least 44.1 khz and at
least 16-bit.

Sorry for the above repeat post.

AFAIK, the audio that is originally panned to the center is
significantly louder than the audio whose phase is different in the
left & right channels. This is why, I reduce the loudness of
non-inverted stereo audio file by 77.5% [before converting it to mono].
In the stereo file whose left channel has its phase inverted, I
decrease the loudness only by 50%.

I then use Wavelab or Adobe Audition to create new WAV file with 44.1
khz, 16-bit, stereo. Next, I cut and paste the audio from one of the
above files into the left channel of the new file, and the audio from
another one of the above files into the right channel. Then I convert
this new WAV file to mono. I just love moderation, I don't want either
the center or the surround to be too loud or too soft. I want
everything equal and this is how I do it.

Here are the steps:

1. Record audio from CD into Wavelab or Adobe Audition into a file. For
simplicity lets call this file "Track1.wav"

2. Make a copy of Track1.wav and save the copy as "Track1B.wav"

3. Open Track1.wav and reduce the gain of its audio by 77.5%

4. Convert Track1.wav to monoaural audio

5. Save Track.1

6. Open Track1B.wav and reduce its audio gain by 50%

7. Invert the phase of the left channel of Track1B.wav

8. Convert Track1B.wav to mono

9. Save Track1B.wav

10. Create a new stereo wave file whose bit-resolution is 16-bit and
sample rate is 44.1 khz. For simplicity lets call this file
"untitled.wav"

11. Copy and paste the audio of Track1.wav into the left channel of
untitled.wav

12. Copy and paste the audio of Track1B.wave into the right channel of
untitled.wav

13. Convert untitled.wav to mono

14. Save untitled.wav

15. Listen to the beauty of the song in untitled.wav!

If you want to place untitled.wav on the internet, chances are you'll
have to somehow compress it becayse WAV files tend to take up lots of
bandwidth. My idea is to convert this file to WMA. Make sure
untitled.wma is monoaural and has a sample-rate of 44.1 khz. To restore
bandwidth make sure untitled.wma has a bit-rate of 20kbps. As with any
compression, be sure to use mono and a sample rate that is at least
44.1 khz and make sure that the audio is the same sample rate as it was
before the compression -- your ears will thank you for it!

My advice is to avoid using MP3s and other non-WMA compression schemes
for audio. Yes, peer pressure to use MP3s maybe intense but just try
your best to steer clear.

[Laurence Payne]

On 6 Nov 2006 15:28:02 -0800, "Radium" > wrote:

>Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
>Adobe Audition to convert the stereo file to mono.

Why?

[[email protected]]

Radium wrote:
>
> Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
> Adobe Audition to convert the stereo file to mono.

Why?

> I then combine the audio of both copies together in a new WAV file.
> Voila! I get some nice monoaural audio with at least 44.1 khz and at
> least 16-bit.

The problem is that it's not true mono, but rather matrixed mono. :-)

> Sorry for the above repeat post.

Yeah, so are we.

> If you want to place untitled.wav on the internet, chances are you'll
> have to somehow compress it becayse WAV files tend to take up lots of
> bandwidth. My idea is to convert this file to WMA. Make sure
> untitled.wma is monoaural and has a sample-rate of 44.1 khz. To restore
> bandwidth make sure untitled.wma has a bit-rate of 20kbps.

Well, that explains a lot behind your question.

Dropping CD quality audio 20 an effective bit rate of
20 kbits/sec is VERY likely going to sound perfectly
LOUSY no matter what encoder you use. That's a
compression ratio of about 70:1.

SO your question "WMA vs MP3" really boils down
to which miserable piece of **** audio are you willing
to tolerate.

The answer is neither, they both suck.

[Radium]

Laurence Payne wrote:
> On 6 Nov 2006 15:28:02 -0800, "Radium" > wrote:
>
> >Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
> >Adobe Audition to convert the stereo file to mono.
>
> Why?

Because:

1. I want things to sound the same in all speakers

2. I do not want the original central channel to be louder than the
original signals that are different in both the left and right. Since
the central channel is normally recorded noticeably louder than the
signals that are not in phase, I like to decrease the volume of the
center by 77.5% while decreasing the periphery by only 50%

IOW, what in-phase and whats not in-phase should not have any
noticeably differences in volume and all audio channels should give out
the same signal.

[Richard Crowley]

"Radium" wrote ...
> Laurence Payne wrote:
>> "Radium" wrote:
>>
>> >Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
>> >Adobe Audition to convert the stereo file to mono.
>>
>> Why?
>
> Because:
>
> 1. I want things to sound the same in all speakers

That makes no sense at all.

> 2. I do not want the original central channel to be louder than the
> original signals that are different in both the left and right. Since
> the central channel is normally recorded noticeably louder than the
> signals that are not in phase, I like to decrease the volume of the
> center by 77.5% while decreasing the periphery by only 50%

Stereo is two channels by definition. There is no "central channel".

> IOW, what in-phase and whats not in-phase should not have any
> noticeably differences in volume and all audio channels should give out
> the same signal.

Does color television trouble you? Only have black & white?

[Radium]

wrote:
> Well, that explains a lot behind your question.
>
> Dropping CD quality audio 20 an effective bit rate of
> 20 kbits/sec is VERY likely going to sound perfectly
> LOUSY no matter what encoder you use. That's a
> compression ratio of about 70:1.

Okay.

> SO your question "WMA vs MP3" really boils down
> to which miserable piece of **** audio are you willing
> to tolerate.
>
> The answer is neither, they both suck.

Well, if my bandwidth is suffering I don't mind 20 kbps of WMA. Just
make sure that it is monoaural and at least 44.1 khz sample rate. Oh,
and make sure that it is 44.1 khz and monoaural before it is compressed
as well. Don't want aliasing or any artifacts associated with
sample-rate change. IOW, the sample rate of the digital audio should be
the same [and at least 44.1 khz] before compression and after
compression.

As for MP3s and non-WMA compression schemes. F--k those s--ts. If I am
going to listen to digital audio that is compressed with something
other than WMA, than it needs to be at least 44.1 khz, monoaural, and
no less than 320 kbps. Even in 128 kbps MP3s I notice the ear-foaming
degradation.

I really don't get how anyone can stand the noise that occurs in MP3
and other non-WMA compressions.

[Radium]

Richard Crowley wrote:
> "Radium" wrote ...
> > Laurence Payne wrote:
> >> "Radium" wrote:
> >>
> >> >Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
> >> >Adobe Audition to convert the stereo file to mono.
> >>
> >> Why?
> >
> > Because:
> >
> > 1. I want things to sound the same in all speakers
>
> That makes no sense at all.

How doesn't it?

> > 2. I do not want the original central channel to be louder than the
> > original signals that are different in both the left and right. Since
> > the central channel is normally recorded noticeably louder than the
> > signals that are not in phase, I like to decrease the volume of the
> > center by 77.5% while decreasing the periphery by only 50%
>
> Stereo is two channels by definition. There is no "central channel".

Yes. Stereo is two channels. However, most of today's music contains
audio is in-phase for both L and R channels -- usually the lead vocals,
basses, and percussions, and audio that phases differently in L and R
channels -- usually the pianos, guitars, choruses, and synth pads.

By the "central channel", I am reffering to the parts of the signal
stuff that is in-phase for both L and R channels.

Those so-called "voice cancellors" rely on the assumption that the lead
vocals are in the center. Voice-cancellors work by inverting the phase
of one stereo channel and combining it with the other. This results in
an end mono signal without the audio that was identical in the L and R
channels when the signals were stereo. Because of this method, most
"vocal eliminators" also end up removing the bass and percussive
instruments.

> > IOW, what in-phase and whats not in-phase should not have any
> > noticeably differences in volume and all audio channels should give out
> > the same signal.
>

> Does color television trouble you?

Not at all.

>Only have black & white?

Sorry. I don't understand your analogy. Please explain.

[Richard Crowley]

"Radium" > wrote in message
oups.com...
>
> Richard Crowley wrote:
>> "Radium" wrote ...
>> > Laurence Payne wrote:
>> >> "Radium" wrote:
>> >>
>> >> >Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab
>> >> >or
>> >> >Adobe Audition to convert the stereo file to mono.
>> >>
>> >> Why?
>> >
>> > Because:
>> >
>> > 1. I want things to sound the same in all speakers
>>
>> That makes no sense at all.
>
> How doesn't it?
>
>> > 2. I do not want the original central channel to be louder than the
>> > original signals that are different in both the left and right. Since
>> > the central channel is normally recorded noticeably louder than the
>> > signals that are not in phase, I like to decrease the volume of the
>> > center by 77.5% while decreasing the periphery by only 50%
>>
>> Stereo is two channels by definition. There is no "central channel".
>
> Yes. Stereo is two channels. However, most of today's music contains
> audio is in-phase for both L and R channels -- usually the lead vocals,
> basses, and percussions, and audio that phases differently in L and R
> channels -- usually the pianos, guitars, choruses, and synth pads.
>
> By the "central channel", I am reffering to the parts of the signal
> stuff that is in-phase for both L and R channels.

Sorry to hear that the only kind of "music" you listen to is over-
processed pan-pot multi-mono. Get out and hear some real music
sometime.

> Those so-called "voice cancellors" rely on the assumption that the lead
> vocals are in the center. Voice-cancellors work by inverting the phase
> of one stereo channel and combining it with the other. This results in
> an end mono signal without the audio that was identical in the L and R
> channels when the signals were stereo. Because of this method, most
> "vocal eliminators" also end up removing the bass and percussive
> instruments.

I don't see what this has to do with your preference for listening
to stereo program material in monaural? But never mind, I have
not understood the purpose or motivation of any other parts of
this thread, either, so I'll leave you to it.

>> > IOW, what in-phase and whats not in-phase should not have any
>> > noticeably differences in volume and all audio channels should give out
>> > the same signal.
>>
>
>> Does color television trouble you?
>
> Not at all.
>
>>Only have black & white?
>
> Sorry. I don't understand your analogy. Please explain.

Squashing stereo back to monaural seems very much like
removing the color from video. It takes technology in a "retro"
direction (i.e. backwards.)

[Radium]

Richard Crowley wrote:
> "Radium" > wrote in message
> oups.com...
> >
> > Richard Crowley wrote:
> >> "Radium" wrote ...
> >> > Laurence Payne wrote:
> >> >> "Radium" wrote:
> >> >>
> >> >> >Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab
> >> >> >or
> >> >> >Adobe Audition to convert the stereo file to mono.
> >> >>
> >> >> Why?
> >> >
> >> > Because:
> >> >
> >> > 1. I want things to sound the same in all speakers
> >>
> >> That makes no sense at all.
> >
> > How doesn't it?
> >
> >> > 2. I do not want the original central channel to be louder than the
> >> > original signals that are different in both the left and right. Since
> >> > the central channel is normally recorded noticeably louder than the
> >> > signals that are not in phase, I like to decrease the volume of the
> >> > center by 77.5% while decreasing the periphery by only 50%
> >>
> >> Stereo is two channels by definition. There is no "central channel".
> >
> > Yes. Stereo is two channels. However, most of today's music contains
> > audio is in-phase for both L and R channels -- usually the lead vocals,
> > basses, and percussions, and audio that phases differently in L and R
> > channels -- usually the pianos, guitars, choruses, and synth pads.
> >
> > By the "central channel", I am reffering to the parts of the signal
> > stuff that is in-phase for both L and R channels.
>
> Sorry to hear that the only kind of "music" you listen to is over-
> processed pan-pot multi-mono. Get out and hear some real music
> sometime.
>
> > Those so-called "voice cancellors" rely on the assumption that the lead
> > vocals are in the center. Voice-cancellors work by inverting the phase
> > of one stereo channel and combining it with the other. This results in
> > an end mono signal without the audio that was identical in the L and R
> > channels when the signals were stereo. Because of this method, most
> > "vocal eliminators" also end up removing the bass and percussive
> > instruments.
>
> I don't see what this has to do with your preference for listening
> to stereo program material in monaural? But never mind, I have
> not understood the purpose or motivation of any other parts of
> this thread, either, so I'll leave you to it.

I don't want too much lead vocal. That the main reason. The other
reason is I prefer both my ears to be equal when listening to music
from an electronic source.


>
> >> > IOW, what in-phase and whats not in-phase should not have any
> >> > noticeably differences in volume and all audio channels should give out
> >> > the same signal.
> >>
> >
> >> Does color television trouble you?
> >
> > Not at all.
> >
> >>Only have black & white?
> >
> > Sorry. I don't understand your analogy. Please explain.
>
> Squashing stereo back to monaural seems very much like
> removing the color from video. It takes technology in a "retro"
> direction (i.e. backwards.)

Not really. Initally, I need the recording to be stereo, so that I can
decrease the volume of the lead vocals.

Do one thing. Try listening to a non-WMA compression file -- such as
MP3 -- whose format is in stereo -- or better yet, 7.1 surround -- with
a sample-rate of 44.1 khz, and a bit-rate of 20 kbps. Tell me how you
like it -- if you can still keep your sanity.

[Laurence Payne]

On 6 Nov 2006 17:39:25 -0800, "Radium" > wrote:

>
>Because:
>
>1. I want things to sound the same in all speakers

Why? Stereo is good. Why discard it?

[[email protected]]

Radium wrote:
> wrote:
> > SO your question "WMA vs MP3" really boils down
> > to which miserable piece of **** audio are you willing
> > to tolerate.
> >
> > The answer is neither, they both suck.
>
> Well, if my bandwidth is suffering I don't mind 20 kbps of WMA.

Like I said. They both suck big time. It appears that
you like the miserably ****ty sound of 20 kbs WMA
better than the miserably ****ty sound of 20 kbps MP3.

Hey, more power to you. Just please refrain from making
your usual sweeping generalizations as you have in the
past, because it's clear you have ignored or missed some
rather important points, like the quality of ANY bit-rate audio
compression is almost totally dependent upon how well
the encoder is implemented, more so than on the actual
format, and that it's easy to get miserable results if you
don't use the encoder properly.

And, yeah, your bandwidth probably is suffering, but
not the way you think. Your earlier threads (e.g. "I am back,"
"bit resolution and clipping," "linear pcm vs common pcm,"
"theoretical acoustic experiment" just to mention a couple
of your gems) indicate you have some serious bandwidth
problems.

To save precious bandwidth, don't bother replying. It'd
be content-free anyway.

[Richard Crowley]

"Radium" wrote ...
> Do one thing. Try listening to a non-WMA compression file -- such as
> MP3 -- whose format is in stereo -- or better yet, 7.1 surround --
> with
> a sample-rate of 44.1 khz, and a bit-rate of 20 kbps. Tell me how you
> like it -- if you can still keep your sanity.

Listening to anyting at 20Kbps is insane by definition.
Why would I do such a silly thing in the first place?

I don't really have any interest in playing the audio
compression version of "How Low Can You Go?"
I'll leave you kids to your games.

[Mr.T]

"Walt" > wrote in message
...
> >>Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
> >>Adobe Audition to convert the stereo file to mono. But its more
> >>complicated than just stereo-to-mono conversion. First, I make two
> >>copies the stereo file. With one copy, I invert the phase of the left
> >>channel and then combine it with the right channel -- this causes
> >>whatever was identical the left and right channel to be cancelled.
> >>Before I combine these two channels, I decrease the loudness of each
> >>channel by 50%.
> >>
> >>With the other copy of the stereo file, I don't invert either the left
> >>or right channel. I simply decrease the loudness of both channels by
> >>77.5% and then I combine the left and right together to make a mono
> >>file.
> >>
> >>I then combine the audio of both copies together in a new WAV file.
> >>Voila! I get some nice monoaural audio with at least 44.1 khz and at
> >>least 16-bit.
> >
> > Anybody?
>
> Interesting approach. It appears he's making a mono sum by
>
> .5(R-L) + .775(L+R) = 1.25(R) + .25(L)
>
> I guess he hates the leftt channel. More cello, more bass, less violin,
> less percussion. Hey, whatever floats your boat.

The maths is not quite right. Since he is first doing a difference of L and
R, that can be .5(R-L) *or* .5(L-R)
(same thing if you don't ignore the fact that either can be negative
relative to the other depending on the original phase)
Therefore the left channel will NOT automatically be lower than the right.
However the result may or may not be desirable depending on exactly where
each instrument is panned in the mix.

Personally I would only do that if the original mix did not sum nicely to
mono in the first place.

MrT.

[Mr.T]

"Laurence Payne" <lpayne1NOSPAM@dslDOTpipexDOTcom> wrote in message
...
> >1. I want things to sound the same in all speakers
>
> Why? Stereo is good. Why discard it?

I thought he was pretty clear. Low bit rate MONO files are FAR superior to
listen to than low bit rate stereo files, since you don't waste bits
encoding the channel differences which are a much lesser auditory
requirement than low distortion and decent frequency response.
For most people anyway, there will always be some that prefer very low
quality stereo to higher quality mono I guess.

And the only reason for such low bit rates was also stated, *streaming
audio*.

MrT.

[Walt]

Mr.T wrote:
> "Walt" > wrote
>
>>>>Whenever I buy a CD, I put my favorite songs in my PC. I use Wavelab or
>>>>Adobe Audition to convert the stereo file to mono. But its more
>>>>complicated than just stereo-to-mono conversion. First, I make two
>>>>copies the stereo file. With one copy, I invert the phase of the left
>>>>channel and then combine it with the right channel -- this causes
>>>>whatever was identical the left and right channel to be cancelled.
>>>>Before I combine these two channels, I decrease the loudness of each
>>>>channel by 50%.
>>>>
>>>>With the other copy of the stereo file, I don't invert either the left
>>>>or right channel. I simply decrease the loudness of both channels by
>>>>77.5% and then I combine the left and right together to make a mono
>>>>file.
>>>>
>>>>I then combine the audio of both copies together in a new WAV file.
>>>>Voila! I get some nice monoaural audio with at least 44.1 khz and at
>>>>least 16-bit.
>>>
>>>Anybody?
>>
>>Interesting approach. It appears he's making a mono sum by
>>
>>.5(R-L) + .775(L+R) = 1.25(R) + .25(L)
>>
>>I guess he hates the leftt channel. More cello, more bass, less violin,
>> less percussion. Hey, whatever floats your boat.
>
>
> The maths is not quite right.

Yep, it should be:

..5(R-L) + .775(L+R) = 1.75(R) + .275(L)

> Since he is first doing a difference of L and R, that can be .5(R-L) *or* .5(L-R)

Well, he says he "invert[s] the phase left channel", so I'll take him at
his word. That's the only place he say he (intentionally) inverts polarity.

What's unclear is what he means by "decrease the loudness of both
channels by 77.5%" before summing to mono - I wrote that as .775(L+R)
when it may well be .225(L+R)

That would give:

..5(R-L) + .225(L+R) = .725(R) - .275(L)

Why anybody would want to do this is beyond me.



//Walt

[Richard Crowley]

"Mr.T" <MrT@home> wrote in message
u...
>
> "Laurence Payne" <lpayne1NOSPAM@dslDOTpipexDOTcom> wrote in message
> ...
>> >1. I want things to sound the same in all speakers
>>
>> Why? Stereo is good. Why discard it?
>
> I thought he was pretty clear. Low bit rate MONO files are FAR superior to
> listen to than low bit rate stereo files, since you don't waste bits
> encoding the channel differences which are a much lesser auditory
> requirement than low distortion and decent frequency response.
> For most people anyway, there will always be some that prefer very low
> quality stereo to higher quality mono I guess.
>
> And the only reason for such low bit rates was also stated, *streaming
> audio*.

He does have a point. This is approaching the definition of
"telephonic".

"Radium" > wrote in message
oups.com...
> Hi group:
>
> Lets say there is a 44.1 khz, monoaural, 32 kbps WMA file and a 44.1
> khz, monoaural, 32 kbps MP3 file; which one would rather listen to?
>
> I'd prefer the WMA. I readily notice the difference in WMA artifacts
> and MP3 artifacts; the differences are very difficult for me to
> describe, but they are signficant. Both resemble "digital tones" of old
> video games and binary signals but they are noticeably different from
> each other.
>
> BTW, what is responsible for those differences in audio artifacts
> resulting from a low-bit-rate WMAs vs. low-bit-rate MP3s?

I use WMA--not because it's clearly better, but because I got started with
WMA and have never seen any reason to change. Presently, I have 2600 hours
of music encoded in WMA. If I were to start today, I would probably use
AAC.

Norm Strong

[Mr.T]

"Walt" > wrote in message
...
>
> .5(R-L) + .775(L+R) = 1.75(R) + .275(L)
>
> > Since he is first doing a difference of L and R, that can be .5(R-L)
*or* .5(L-R)

> Well, he says he "invert[s] the phase left channel", so I'll take him at
> his word. That's the only place he say he (intentionally) inverts
polarity.

You are disregarding the original phase relationships. Consider the formula
for negative R, or negative L, and everything in between.

> What's unclear is what he means by "decrease the loudness of both
> channels by 77.5%" before summing to mono - I wrote that as .775(L+R)
> when it may well be .225(L+R)
>
> That would give:
>
> .5(R-L) + .225(L+R) = .725(R) - .275(L)
>
> Why anybody would want to do this is beyond me.

Agreed.

MrT.