[Radium]

On Feb 23, 5:45 am, "Rick Massey" wrote in
http://groups.google.com/group/sci.e...50bb50e?hl=en&
:

Not exactly "Wrong", but vastly oversimplified. Look up Sideband frequencies
in your resources that you use to learn about this stuff. It's an aural
illusion you're hearing more than anything else.

A solid grounding in FM theory will help a lot too -- there are lots of good
tutorials on the web, and you very well may find that some of the
limitations of your soundcard are shown out in this as you start dealing
with higher numbers of operators in more professional environments.

Doing as my research as I can this week, I still haven't found and
answer to how this "illusion" is generated by cancelling the identical
stereo signals. I've looked in detail about sidebands frequencies and
FM theory but still don't have a clue as to how the out-of-phase
signals are generated in CMS.

[[email protected]]

On Mar 3, 8:59 pm, "Radium" wrote:
On Feb 23, 5:45 am, "Rick Massey" wrote inhttp://groups.google.com/group/sci.electronics.basics/msg/ecfff083e50...
:

Not exactly "Wrong", but vastly oversimplified. Look up Sideband frequencies
in your resources that you use to learn about this stuff. It's an aural
illusion you're hearing more than anything else.

A solid grounding in FM theory will help a lot too -- there are lots of good
tutorials on the web, and you very well may find that some of the
limitations of your soundcard are shown out in this as you start dealing
with higher numbers of operators in more professional environments.

Doing as my research as I can this week, I still haven't found and
answer to how this "illusion" is generated by cancelling the identical
stereo signals. I've looked in detail about sidebands frequencies and
FM theory but still don't have a clue as to how the out-of-phase
signals are generated in CMS.

Dang I'm feeding the troll. OK Radium, take an average stereo
recording (with center vocalist) in CoolEdit, invert the phase of
either channel and then add them together. Presto, the vocalist is
mostly gone. Fiddling with levels and timing offsets can achieve a
nerly perfect null of the center field with all that remains being the
_difference_ between the channels.

As far as "identical stereo signals", there is no such thing. If the
channels are identical, it's called mono (comparatively boring).

What is CMS? Is it anything like PMS?

GG

[Laurence Payne]

On 3 Mar 2007 21:57:17 -0800, wrote:

take an average stereo
recording (with center vocalist) in CoolEdit, invert the phase of
either channel and then add them together. Presto, the vocalist is
mostly gone. Fiddling with levels and timing offsets can achieve a
nerly perfect null of the center field with all that remains being the
_difference_ between the channels.


....an average stereo recording...?

You speak from experience?

Ok - take half a dozen CDs at random from your collection. Take a
random track from each. There should be some "average" recordings in
there yes?

Now use the phase reversal trick, plus whatever further fiddling you
like, to remove lead vocals. Post some short examples of the "nearly
perfect null" you achieve.

[Rick Massey]

What is CMS? Is it anything like PMS?

Creative Music Synth. It's a program for editing sounds on a Creative Labs
sound card. Radium is obsessed with this little program the company created
a while ago, and for some unfathomable reason, has set this four operator
consumer level synth as his or her gold standard for synthesis.

For someone who works with professional level products, it is probably only
slightly less frustrating than PMS.

[Radium]

On Mar 4, 7:23 am, "Rick Massey" wrote:

Creative Music Synth. It's a program for editing sounds on a Creative Labs
sound card.

Actually, Creative Music Synth is not a program.

Creative Music Synth is the digital hardware chip-based FM synth
present on the SoundBlaster 16 ISA card.

Usually Creative Music Synth is set to an I/O address of 220, though
in some PCs I've seen it at 240.

Radium is obsessed with this little program the company created
a while ago, and for some unfathomable reason, has set this four operator
consumer level synth as his or her gold standard for synthesis.

Well, as I've in http://groups.google.com/group/sci.e...2cea68f?hl=en&

Despite doing my own research, I still don't have the foggiest idea
how the out-of-phase signals are generated in Creative Music Synth.
Sorry if my persistence on this topic is annoying but I am insanely
interested in it.

Creative Music Synth does not -- to any extent -- playback samples
from memory. It freshly generates its tones. It is physically built to
generates its tones from scratch without stealing any energy from the
PC's main CPU.

[[email protected]]

On Mar 4, 4:06 am, Laurence Payne lpayne1NOSPAM@dslDOTpipexDOTcom
wrote:
On 3 Mar 2007 21:57:17 -0800, wrote:

take an average stereo
recording (with center vocalist) in CoolEdit, invert the phase of
either channel and then add them together. Presto, the vocalist is
mostly gone. Fiddling with levels and timing offsets can achieve a
nerly perfect null of the center field with all that remains being the
_difference_ between the channels.

...an average stereo recording...?

You speak from experience?

Ok - take half a dozen CDs at random from your collection. Take a
random track from each. There should be some "average" recordings in
there yes?

Now use the phase reversal trick, plus whatever further fiddling you
like, to remove lead vocals. Post some short examples of the "nearly
perfect null" you achieve.

OK make it oldies. I just tried it on 'Saturday In The Park" from
Chicago V. Nearly total null of vocalist. Carole King "I Feel The
Earth Move" nearly perfect (estimate -40dB) null of Carole. "Goodby
To Love" by the Carpenters leaves the reverb portion of Karen
Carpenter. "New San Antonio Rose" on the ride with Bob by Asleep at
the Wheel also took out Dwight Yoakam. These samples were taken from
the CD .

GG

[Stuart]

"Laurence Payne" lpayne1NOSPAM@dslDOTpipexDOTcom wrote in message
...
On 3 Mar 2007 21:57:17 -0800, wrote:

take an average stereo
recording (with center vocalist) in CoolEdit, invert the phase of
either channel and then add them together. Presto, the vocalist is
mostly gone. Fiddling with levels and timing offsets can achieve a
nerly perfect null of the center field with all that remains being the
_difference_ between the channels.


...an average stereo recording...?

You speak from experience?

Ok - take half a dozen CDs at random from your collection. Take a
random track from each. There should be some "average" recordings in
there yes?

Now use the phase reversal trick, plus whatever further fiddling you
like, to remove lead vocals. Post some short examples of the "nearly
perfect null" you achieve.

Another trick some music producers do to prevent vocal removal is to
introduce a +/- 45 degree phase shift shuffle circuit. Couple with a
shuffled and crossed reverb return prevents simple L-R circuits from
achieving a cheap backing track. If you want to try it use Adobe Audition
load in iZotope Ozone Plug-in.

[Stuart]

"Stuart" wrote in message
...

"Laurence Payne" lpayne1NOSPAM@dslDOTpipexDOTcom wrote in message
...
On 3 Mar 2007 21:57:17 -0800, wrote:

take an average stereo
recording (with center vocalist) in CoolEdit, invert the phase of
either channel and then add them together. Presto, the vocalist is
mostly gone. Fiddling with levels and timing offsets can achieve a
nerly perfect null of the center field with all that remains being the
_difference_ between the channels.


...an average stereo recording...?

You speak from experience?

Ok - take half a dozen CDs at random from your collection. Take a
random track from each. There should be some "average" recordings in
there yes?

Now use the phase reversal trick, plus whatever further fiddling you
like, to remove lead vocals. Post some short examples of the "nearly
perfect null" you achieve.

Another trick some music producers do to prevent vocal removal is to
introduce a +/- 45 degree phase shift shuffle circuit. Couple with a
shuffled and crossed reverb return prevents simple L-R circuits from
achieving a cheap backing track. If you want to try it use Adobe Audition
load in iZotope Ozone Plug-in.
Actually it might be easier to set up with Sonitus:FX plug-in as it has a
preset fx for doing this

[Rick Massey]

Sorry if my persistence on this topic is annoying but I am insanely
interested in it.

That might be the single most telling statement you've made about it. It's
an obsession. Your resolve on this topic definitely eludes sanity.



Creative Music Synth does not -- to any extent -- playback samples
from memory. It freshly generates its tones. It is physically built to
generates its tones from scratch without stealing any energy from the
PC's main CPU.


Which makes it exactly like a lot of other synthesizers out there. And
sample playback on a soundcard is a truly different animal from sample
playback in a professional synthesizer, which is one of the many places your
tirade against them breaks down.

[Radium]

On Mar 4, 4:06 pm, "Rick Massey" wrote:

Sorry if my persistence on this topic is annoying but I am insanely
interested in it.

That might be the single most telling statement you've made about it. It's
an obsession. Your resolve on this topic definitely eludes sanity.

I am just really interested in those out-of-phase FM signals in CMS.
How are they generated, if not intentionally by the designer?

Creative Music Synth does not -- to any extent -- playback samples
from memory. It freshly generates its tones. It is physically built to
generates its tones from scratch without stealing any energy from the
PC's main CPU.

Which makes it exactly like a lot of other synthesizers out there. And
sample playback on a soundcard is a truly different animal from sample
playback in a professional synthesizer, which is one of the many places your
tirade against them breaks down.

If a soundcard-based sample playback synth/soundfont/synth-emulation/
softsynth has has a sample-rate, bit-resolutions, # of operators, # of
voices, # of channels** that are at least a 100 times whats necessary
for the average human auditory system to consider "music", AND [even
more important!!!!!!!!!] the recordings are that of actual, high-
quality, new, premium, clean, authentic, tradional instruments [such
as piano, violin, flute, guitar]. Sigh...... Then yes, I am willing to
give it a try and might actually enjoy it -- though, obviously not
*nearly* as much as Creative Music Synth.

But, unfortuntely, most soundcard-based sample playback synths/
soundfonts/synth-emulations/softsynths don't have sufficient sample-
rates, bit-resultions, # of operators, # of voices, # of channels**
near whats necessary for the average human auditory system to consider
"music". *More importantly*, the recordings are made from cheap,
dusty, broken-down, poorly-conditioned instruments -- which is why so
many sample playback synths/soundfonts/synth-emulations/softsynths
stinks like the human kakaa they are. Ever listend to M$ GS
"wavetable" synth? I feel like making and spreading a virus to infect
that softsynth piece of s---.

**Yamaha's OPL3 has 18 channels

[notejam]

This is slightly off topic, but the talk has made me wonder about
keeping the vocals. Is that possible?
I would like to keep the vocal information and remove band backing
from analog and or digital stereo recordings.

The reason I ask is band in a box now has a neat feature that can
create a backing track to any vocal.. Sometimes it works pretty good,
so it would be real nice if some one can create a vocal keeper program
rather than a vocal remover, as a tool to use along with this feature
in band in a box.

[Mr.T]

"notejam" wrote in message
ps.com...
This is slightly off topic, but the talk has made me wonder about
keeping the vocals. Is that possible?
I would like to keep the vocal information and remove band backing
from analog and or digital stereo recordings.

If you can successfully remove the vocal track from the original recording,
the you should be able to invert the phase of that, and add it back to the
original recording, giving you increased vocal instead.

MrT.

[Stuart]

"Mr.T" MrT@home wrote in message
...

"notejam" wrote in message
ps.com...
This is slightly off topic, but the talk has made me wonder about
keeping the vocals. Is that possible?
I would like to keep the vocal information and remove band backing
from analog and or digital stereo recordings.

If you can successfully remove the vocal track from the original
recording,
the you should be able to invert the phase of that, and add it back to the
original recording, giving you increased vocal instead.

MrT.



Sorry it doesn't work like that - consider the algebra of stereo

Left Total = Left channel info plus half centre channel (vocal) L+1/2C
Right Total = Right channel info plus half centre channel (vocal) R + 1/2C

Add them together and you get L+C+R = Mono

Subtract either one from the other and you get (L+1/2C) - (R+1/2C) = L-R
that is mostly the backing and it is a mono signal. R-L is exactly the same
signal merely flipped in phase for one channel and sounds exactly the same.

Now if you add or subtract this signal the L-R or a R-L from the L+C+R you
can see you will never get a pure C. That's because it never existed as a
discrete channel. Many have tried.... added to the fact that it is rare to
get an absolute balance between the channels and the deliberate decision to
fold in the reverb in a phase randomised manner...

The best you can do with getting a clean vocal is to treat the signal as
mono and then use a high pass filter to remove everything below say 120Hz
and do the same above say 7 or 8Khz, then put a mid range bump around 3Khz
and another one around 300Hz. By itself it will sound nasty but put an
accompaniment behind it and you might get away with it. Has been done
putting a new symphony orchestra behind some old acoustic records...BUT! a
lot of work.

[Mr.T]

"Stuart" wrote in message
...
This is slightly off topic, but the talk has made me wonder about
keeping the vocals. Is that possible?
I would like to keep the vocal information and remove band backing
from analog and or digital stereo recordings.

If you can successfully remove the vocal track from the original
recording,
the you should be able to invert the phase of that, and add it back to
the
original recording, giving you increased vocal instead.

Sorry it doesn't work like that - consider the algebra of stereo

Left Total = Left channel info plus half centre channel (vocal) L+1/2C
Right Total = Right channel info plus half centre channel (vocal) R +
1/2C

Assuming everything is panned hard left, hard right and centre.
But for the point of the execrcise we'll ignore that :-)

Add them together and you get L+C+R = Mono

Subtract either one from the other and you get (L+1/2C) - (R+1/2C) = L-R
that is mostly the backing and it is a mono signal. R-L is exactly the
same
signal merely flipped in phase for one channel and sounds exactly the
same.

Now if you add or subtract this signal the L-R or a R-L from the L+C+R you
can see you will never get a pure C. That's because it never existed as a
discrete channel. Many have tried.... added to the fact that it is rare
to
get an absolute balance between the channels and the deliberate decision
to
fold in the reverb in a phase randomised manner...

Agreed, maybe that's why I simply said "increased vocal".
If you reduce any part of L or R, whilst C remains untouched, then C *must*
increase in comparison.

MrT.

[Stuart]

"Mr.T" MrT@home wrote in message
...

"Stuart" wrote in message
...
This is slightly off topic, but the talk has made me wonder about
keeping the vocals. Is that possible?
I would like to keep the vocal information and remove band backing
from analog and or digital stereo recordings.

If you can successfully remove the vocal track from the original
recording,
the you should be able to invert the phase of that, and add it back to
the
original recording, giving you increased vocal instead.

Sorry it doesn't work like that - consider the algebra of stereo

Left Total = Left channel info plus half centre channel (vocal) L+1/2C
Right Total = Right channel info plus half centre channel (vocal) R +
1/2C

Assuming everything is panned hard left, hard right and centre.
But for the point of the execrcise we'll ignore that :-)

Add them together and you get L+C+R = Mono

Subtract either one from the other and you get (L+1/2C) - (R+1/2C) = L-R
that is mostly the backing and it is a mono signal. R-L is exactly the
same
signal merely flipped in phase for one channel and sounds exactly the
same.

Now if you add or subtract this signal the L-R or a R-L from the L+C+R
you
can see you will never get a pure C. That's because it never existed as a
discrete channel. Many have tried.... added to the fact that it is rare
to
get an absolute balance between the channels and the deliberate decision
to
fold in the reverb in a phase randomised manner...

Agreed, maybe that's why I simply said "increased vocal".
If you reduce any part of L or R, whilst C remains untouched, then C
*must*
increase in comparison.

MrT.



As you reduce the L-R level you move the signal closer to mono - if you
increase the level of the L-R component to "increase the stereo effect" it
reduces the vocal or rather any information that is common to both channels.
The benchmark for the balance between centre channel apparent loudness
compared to the backing is to listen to the total mix in mono. In practical
studio mixing nothing much is hard left or hard right but in a locus defined
around the centre line, especially if standard MS stereo microphones are
used. XY microphones are even more diffuse and the ORTF parameters ( 2
cardioids spaced at an angle of 110 degrees gives a superb spaciousness for
large scale orchestral works or ambience effects but dreadful for soloists.

In electronic music and midi based works then of course everything has to be
panned into the desired place. In commercial recordings the maximum centre
channel apparent loudness is set by monitoring in mono, once you switch to
stereo the instruments left or right will start to effect the balance also
the haas effect or precedence comes into play. So answering the op question
" I would like to keep the vocal information and remove band backing from
analog and or digital stereo recordings" is no it can't be done with normal
2 channel stereo recordings.

[Laurence Payne]

On 4 Mar 2007 20:27:52 -0800, "notejam" wrote:

This is slightly off topic, but the talk has made me wonder about
keeping the vocals. Is that possible?
I would like to keep the vocal information and remove band backing
from analog and or digital stereo recordings.

The reason I ask is band in a box now has a neat feature that can
create a backing track to any vocal.. Sometimes it works pretty good,
so it would be real nice if some one can create a vocal keeper program
rather than a vocal remover, as a tool to use along with this feature
in band in a box.


This isn't even theoretically possible by phase-reversal.

Adobe Audition has a plug-in that uses various techniques to attempt
what you want. It makes the best job of it that I've encountered yet.
But, I fear, nowhere near good enough to extract a vocal clean enough
for your purposes.

[Rick Massey]

which is why so
many sample playback synths/soundfonts/synth-emulations/softsynths
stinks like the human kakaa they are. Ever listend to M$ GS
"wavetable" synth? I feel like making and spreading a virus to infect
that softsynth piece of s---.


It's a soundcard. No one sane expects it to sound great. It is what it is,
something for sound on a computer aimed at the consumer market. It's
background for games and other multi-media applications. It's not designed
for any serious work in real music synthesis. For that, get a real
synthesizer or use csound.

[Radium[_3_]]

On Mar 5, 5:32 am, "Rick Massey" wrote:

It's a soundcard. No one sane expects it to sound great. It is what it is,
something for sound on a computer aimed at the consumer market. It's
background for games and other multi-media applications. It's not designed
for any serious work in real music synthesis. For that, get a real
synthesizer or use csound.

Anyways, lets get back on topic.

Doing as my research as I can this week, I still haven't found and
answer to how this "illusion" is generated by cancelling the identical
stereo signals. I've looked in detail about sidebands frequencies and
FM theory but still don't have a clue as to how the out-of-phase
signals are generated in Creative Music Synth.

Please assist.

[[email protected]]

On Mar 5, 1:22 pm, "Radium" wrote:
On Mar 5, 5:32 am, "Rick Massey" wrote:

It's a soundcard. No one sane expects it to sound great. It is what it is,
something for sound on a computer aimed at the consumer market. It's
background for games and other multi-media applications. It's not designed
for any serious work in real music synthesis. For that, get a real
synthesizer or use csound.

Anyways, lets get back on topic.

Doing as my research as I can this week, I still haven't found and
answer to how this "illusion" is generated by cancelling the identical
stereo signals. I've looked in detail about sidebands frequencies and
FM theory but still don't have a clue as to how the out-of-phase
signals are generated in Creative Music Synth.

Please assist.

FM is crap and nobody cares except you. Get Gigastudio and get real.
Let the flames begin.

GG

[Rick Massey]

"Radium" wrote in message
ps.com...
On Mar 5, 5:32 am, "Rick Massey" wrote:

It's a soundcard. No one sane expects it to sound great. It is what it
is,
something for sound on a computer aimed at the consumer market. It's
background for games and other multi-media applications. It's not
designed
for any serious work in real music synthesis. For that, get a real
synthesizer or use csound.

Anyways, lets get back on topic.

Doing as my research as I can this week, I still haven't found and
answer to how this "illusion" is generated by cancelling the identical
stereo signals. I've looked in detail about sidebands frequencies and
FM theory but still don't have a clue as to how the out-of-phase
signals are generated in Creative Music Synth.

Please assist.

You need a good understanding of acoustics and how wave interactions work
before any of this is going to make one lick of sense to you. And in getting
that grounding in the real theory of why this stuff works, you'll probably
answer your own question.

go find a good acoustics text book. A good freeware or shareware spectrum
analyzer will also help you also, as will an oscilloscope program. With
these tools, you'll probably find the answer to your question.