[New_idea]

Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter.
The PWM signal can be feed direct to a class D power amplifier.
This can save a lot of "signal conversion'"

[Don Pearce[_3_]]

On Thu, 8 Oct 2009 17:40:55 +0800, "New_idea"
wrote:

Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter.
The PWM signal can be feed direct to a class D power amplifier.
This can save a lot of "signal conversion'"



One word. Jitter.

d

[Tony[_11_]]

On Thu, 08 Oct 2009 10:16:35 GMT, (Don Pearce) wrote:

On Thu, 8 Oct 2009 17:40:55 +0800, "New_idea"
wrote:

Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter.
The PWM signal can be feed direct to a class D power amplifier.
This can save a lot of "signal conversion'"



One word. Jitter.

d

And even worse - a practical class D amp doesn't produce a straight PWM signal - it has to
account for supply variations. And there has to be a way to introduce a volume and maybe
tone controls - all impossible if the signal starts out as PWM

Tony

[GregS[_3_]]

In article , (Don Pearce) wrote:
On Thu, 8 Oct 2009 17:40:55 +0800, "New_idea"
wrote:

Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter.
The PWM signal can be feed direct to a class D power amplifier.
This can save a lot of "signal conversion'"



One word. Jitter.


More words, noise, pops, cracks, dropouts.


I sugested going out of the converter unfiltered, but basically you
need a digital to PWM converter to get the digital benefits.

greg

[William Sommerwerck]

There are practical issues involved. One is sampling rate. Another is the
way "Class D" is implemented in a given amplifier.

There's also the issue of whether (here we go!) the PWM should be analog or
digital. As another poster pointed out, jitter would be a problem, much more
for analog than digital -- not to mention the difficulty of exactly
controlling the pulse width in a polycarbonate pit.

It isn't an inherently bad idea. It's technically feasible, just not
commercially practical at this time.

[Don Pearce[_3_]]

On Thu, 8 Oct 2009 04:38:40 -0700, "William Sommerwerck"
wrote:

There are practical issues involved. One is sampling rate. Another is the
way "Class D" is implemented in a given amplifier.

There's also the issue of whether (here we go!) the PWM should be analog or
digital. As another poster pointed out, jitter would be a problem, much more
for analog than digital -- not to mention the difficulty of exactly
controlling the pulse width in a polycarbonate pit.

It isn't an inherently bad idea. It's technically feasible, just not
commercially practical at this time.


When you say digital, do you mean store the data as a number, as is
done now, and feed it to a DAC, the output of which is a pulse width
rather than a voltage? If so, that could be done with the CDs we have
right now.

d

[William Sommerwerck]

When you say digital, do you mean store the data as a number,
as is done now, and feed it to a DAC, the output of which is a
pulse width rather than a voltage? If so, that could be done with
the CDs we have right now.

Here we go...

When I say "digital", I mean "quantized" -- there is only a finite number of
possible pulse widths.

This is doable right now, and in fact was done 30 years ago, with the
original LaserDisks, which used a form of analog pulse-width modulation.

[Don Pearce[_3_]]

On Thu, 8 Oct 2009 05:11:34 -0700, "William Sommerwerck"
wrote:

When you say digital, do you mean store the data as a number,
as is done now, and feed it to a DAC, the output of which is a
pulse width rather than a voltage? If so, that could be done with
the CDs we have right now.

Here we go...

When I say "digital", I mean "quantized" -- there is only a finite number of
possible pulse widths.

This is doable right now, and in fact was done 30 years ago, with the
original LaserDisks, which used a form of analog pulse-width modulation.


OK, I get it. But why would the possible number of pulse widths be
finite (apart from quantum effects of course)? You can theoretically
make a pit any arbitrary length. Of course you would instantly be at
the mercy of scratches and dirt. Error correction would be a thing of
the past.

d

[William Sommerwerck]

When you say digital, do you mean store the data as a number,
as is done now, and feed it to a DAC, the output of which is a
pulse width rather than a voltage? If so, that could be done with
the CDs we have right now.

When I say "digital", I mean "quantized" -- there is only a finite
number of possible pulse widths.

This is doable right now, and in fact was done 30 years ago, with the
original LaserDisks, which used a form of analog pulse-width modulation.


OK, I get it. But why would the possible number of pulse widths be
finite (apart from quantum effects of course)? You can theoretically
make a pit any arbitrary length. Of course you would instantly be at
the mercy of scratches and dirt. Error correction would be a thing of
the past.

I feel like Peewee Herman going "Aaaaaahhh!"

Yes, the pit _can_ be of any arbitrary length. But if we quantize the
length, then we have a digital system. If the width falls within a
particular range, then we know -- with almost perfect certainty -- what
_number_ it represents.

As you point out, this is not possible in an analog system.

[Richard Crowley]

William Sommerwerck wrote:
There are practical issues involved. One is sampling rate. Another is
the way "Class D" is implemented in a given amplifier.

There's also the issue of whether (here we go!) the PWM should be
analog or digital. As another poster pointed out, jitter would be a
problem, much more for analog than digital -- not to mention the
difficulty of exactly controlling the pulse width in a polycarbonate
pit.

It isn't an inherently bad idea. It's technically feasible, just not
commercially practical at this time.

Not clear that it is inherently a good idea, either. One reason it isn't
commercially practical is that it gives up all the benefits of digital
encoding and transmission (like error detection and correction,
etc.). You end up with a much LESS robust system that doesn't
have any upside potential. An amusing thought problem, but no
apparent practial benefits.

As Mr. Dorsey has observed here before, there isn't really any
true analog recording. Mechanical recording is quantized by the
wax or vinyl molecules, etc. and magnetic recording is quantized
by the size of the magnetic monopoles, head gap, tape speed, etc.

[William Sommerwerck]

As Mr. Dorsey has observed here before, there isn't really any
true analog recording. Mechanical recording is quantized by the
wax or vinyl molecules, etc. and magnetic recording is quantized
by the size of the magnetic monopoles, head gap, tape speed, etc.

Yes, but... That's really stretching it.

[Richard Crowley]

"William Sommerwerck" wrote...
As Mr. Dorsey has observed here before, there isn't really any
true analog recording. Mechanical recording is quantized by the
wax or vinyl molecules, etc. and magnetic recording is quantized
by the size of the magnetic monopoles, head gap, tape speed, etc.

Yes, but... That's really stretching it.

I can hear the effect in both mechanical and magnetic recording.

[William Sommerwerck]

As Mr. Dorsey has observed here before, there isn't really any
true analog recording. Mechanical recording is quantized by the
wax or vinyl molecules, etc. and magnetic recording is quantized
by the size of the magnetic monopoles, head gap, tape speed, etc.

Yes, but... That's really stretching it.

I can hear the effect in both mechanical and magnetic recording.

What, to your ears, is the subjective effect?

[Arny Krueger]

"New_idea" wrote in message

Just wondering...

If the music producer can record music as PWM (pulse
width modulation) signal direct onto CD.

What's wrong with PCM?

Than a "class-D" CD player doesn't need a Digital-Analog
converter.

Why worry about solved problems?

DACs are now cheap and good.

The PWM signal can be feed direct to a class D
power amplifier.

Providing it used pure PWM, which is not a given.

This can save a lot of "signal conversion'"

Not a serious problem right now.

Here's a question - how will you implement tone and volume controls?

[Anahata]

On Thu, 08 Oct 2009 08:43:18 -0400, Arny Krueger wrote:

Here's a question - how will you implement tone and volume controls?

That's easy - just change the supply voltage to the power output stage.
(how to do that efficiently is left as an exercise for the reader....)

--
Anahata
==//== 01638 720444
http://www.treewind.co.uk ==//== http://www.myspace.com/maryanahata

[Anahata]

On Thu, 08 Oct 2009 08:44:42 -0500, anahata wrote:


That's easy - just change the supply voltage to the power output stage.
(how to do that efficiently is left as an exercise for the reader....)

I meant to say "volume's easy...." of course.

--
Anahata

[Arny Krueger]

"anahata" wrote in message
o.uk

On Thu, 08 Oct 2009 08:43:18 -0400, Arny Krueger wrote:

Here's a question - how will you implement tone and
volume controls?

That's easy - just change the supply voltage to the power
output stage. (how to do that efficiently is left as an
exercise for the reader....)

Even for just volume controls, that methodology gets dicy for high levels of
attenuation.

[Scott Dorsey]

New_idea wrote:
Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter.
The PWM signal can be feed direct to a class D power amplifier.
This can save a lot of "signal conversion'"

This is how SACD works. Sony marketed it aggressively for a while, but
seems to have given up on it. It has good points and bad ones, and
was extensively discussed here five or six years ago when it was a hot
topic.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Anahata]

On Thu, 08 Oct 2009 17:40:55 +0800, New_idea wrote:

Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter. The
PWM signal can be feed direct to a class D power amplifier. This can
save a lot of "signal conversion'"

Leaving aside all the arguments elsewhere in the thread about digital vs.
analog, information theory etc... There is another big problem with this
scheme that has been overlooked. (uless the "jitter" comment was intended
to cover the following)

Such a system would have to have a drive system that guaranteed perfect
noiseless constant speed of rotation of the disc, or at least constant
PWM pulse rate. A servo drive that kept the pulses phase locked to a
local oscillator might be possible to make, but any noise, rumble or
cogging in the drive system would impair the sound directly.

A digital CD player has a big FIFO buffer between the data coming off the
CD itself and the regularly clocked data stream fed into processing and D-
A conversion. The motor only has to maintain a correct average speed
within limits set by the buffer size and it therefor far easier to make.

A CD made this way would also be totally vulnerable to dirt and scratches
as no error correction would be possible.

--
Anahata
==//== +44 (0)1638 720444
http://www.treewind.co.uk

[William Sommerwerck]

Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter. The
PWM signal can be feed direct to a class D power amplifier. This can
save a lot of "signal conversion'"

Leaving aside all the arguments elsewhere in the thread about digital vs.
analog, information theory etc... There is another big problem with this
scheme that has been overlooked. (uless the "jitter" comment was intended
to cover the following)

Such a system would have to have a drive system that guaranteed perfect
noiseless constant speed of rotation of the disc, or at least constant
PWM pulse rate. A servo drive that kept the pulses phase locked to a
local oscillator might be possible to make, but any noise, rumble or
cogging in the drive system would impair the sound directly.

A digital CD player has a big FIFO buffer between the data coming off the
CD itself and the regularly clocked data stream fed into processing and D-
A conversion. The motor only has to maintain a correct average speed
within limits set by the buffer size and it therefor far easier to make.

A CD made this way would also be totally vulnerable to dirt and scratches
as no error correction would be possible.

All this assumes analog PWM. But if you quantized the pulse width -- that
is, digitized it to a finite number of states (widths) -- these problems
would be reduced.

Still, from a practical point of view, PCM is easier to implement in a
robust way than PWM (analog or digital).

[Don Pearce[_3_]]

On Mon, 12 Oct 2009 05:05:19 -0700, "William Sommerwerck"
wrote:

A CD made this way would also be totally vulnerable to dirt and scratches
as no error correction would be possible.

All this assumes analog PWM. But if you quantized the pulse width -- that
is, digitized it to a finite number of states (widths) -- these problems
would be reduced.

The only circumstance under which this would be true is if the scratch
or dirt was right at the end of the pit, and too small to change the
length into the next quantization level. The chances of this happening
must be vanishingly close to zero. Any other dirt or scratch would
have an identical effect in either a continuous or quantized system.

Even in the first circumstance I described, in an analogue system such
a length change would in any case be inaudible.

d

[William Sommerwerck]

"Don Pearce" wrote in message
...
On Mon, 12 Oct 2009 05:05:19 -0700, "William Sommerwerck"
wrote:

A CD made this way would also be totally vulnerable to dirt and
scratches
as no error correction would be possible.

All this assumes analog PWM. But if you quantized the pulse width -- that
is, digitized it to a finite number of states (widths) -- these problems
would be reduced.

The only circumstance under which this would be true is if the scratch
or dirt was right at the end of the pit, and too small to change the
length into the next quantization level. The chances of this happening
must be vanishingly close to zero. Any other dirt or scratch would
have an identical effect in either a continuous or quantized system.

Even in the first circumstance I described, in an analogue system such
a length change would in any case be inaudible.

Any quantized system reduces the effect of noise because the receiver can
make a good (and often correct) guess at what the value "should" be. This
isn't possible in an analog system.

[Don Pearce[_3_]]

On Mon, 12 Oct 2009 05:19:43 -0700, "William Sommerwerck"
wrote:

"Don Pearce" wrote in message
...
On Mon, 12 Oct 2009 05:05:19 -0700, "William Sommerwerck"
wrote:

A CD made this way would also be totally vulnerable to dirt and
scratches
as no error correction would be possible.

All this assumes analog PWM. But if you quantized the pulse width -- that
is, digitized it to a finite number of states (widths) -- these problems
would be reduced.

The only circumstance under which this would be true is if the scratch
or dirt was right at the end of the pit, and too small to change the
length into the next quantization level. The chances of this happening
must be vanishingly close to zero. Any other dirt or scratch would
have an identical effect in either a continuous or quantized system.

Even in the first circumstance I described, in an analogue system such
a length change would in any case be inaudible.

Any quantized system reduces the effect of noise because the receiver can
make a good (and often correct) guess at what the value "should" be. This
isn't possible in an analog system.


No it can't. All it can do is read what it gets and output that. If a
scratch reduces the pit length by 20%, then the output voltage drops
by 20% whether the system is analogue or quantized.

This is not a digital system, remember, just a quantized one, so there
is no way to apply any error correction or DSP to make things better.

d

[Anahata]

On Mon, 12 Oct 2009 05:19:43 -0700, William Sommerwerck wrote:

Any quantized system reduces the effect of noise

Don't forget that quantization *creates* noise....

because the receiver
can make a good (and often correct) guess at what the value "should" be.
This isn't possible in an analog system.

For that to work, the quantization steps would have to be bigger than the
noise introduced by the scratches. Therefore if the scratches were
audible, then the quantization noise itself would also be audible.

So you might be able to retrieve the correct quantized value, but that
quantized value was already a very coarse approximation of the signal in
the first place.

--
Anahata
==//== 01638 720444
http://www.treewind.co.uk ==//== http://www.myspace.com/maryanahata

[Ralph Barone]

In article ,
"William Sommerwerck" wrote:

Just wondering...
If the music producer can record music as PWM (pulse width modulation)
signal direct
onto CD.
Than a "class-D" CD player doesn't need a Digital-Analog converter. The
PWM signal can be feed direct to a class D power amplifier. This can
save a lot of "signal conversion'"

Leaving aside all the arguments elsewhere in the thread about digital vs.
analog, information theory etc... There is another big problem with this
scheme that has been overlooked. (uless the "jitter" comment was intended
to cover the following)

Such a system would have to have a drive system that guaranteed perfect
noiseless constant speed of rotation of the disc, or at least constant
PWM pulse rate. A servo drive that kept the pulses phase locked to a
local oscillator might be possible to make, but any noise, rumble or
cogging in the drive system would impair the sound directly.

A digital CD player has a big FIFO buffer between the data coming off the
CD itself and the regularly clocked data stream fed into processing and D-
A conversion. The motor only has to maintain a correct average speed
within limits set by the buffer size and it therefor far easier to make.

A CD made this way would also be totally vulnerable to dirt and scratches
as no error correction would be possible.

All this assumes analog PWM. But if you quantized the pulse width -- that
is, digitized it to a finite number of states (widths) -- these problems
would be reduced.

Still, from a practical point of view, PCM is easier to implement in a
robust way than PWM (analog or digital).

But, but, but... In a PWM system, the width of the pulse is
proportional to the amplitude of the original analog signal. So, if we
built a system which did parallel PCM and PWM conversion using the same
sample clock, then measured the width of the PWM output and encoded that
result using the same resolution as our PCM signal, wouldn't you end up
with the same encoding either way? Therefore the two are exactly
equivalent. In fact, you could build a DAC which ran off a 2.89 GHz
clock and merely outputted pulses whose width corresponded directly to
the value encoded in the PCM signal. Except for practical issues due to
the clock rates involved, that DAC should work just fine using a PCM
data input. In order to reduce the clock frequency, we use a lower
clock rate and noise shaping algorithms and call it a delta-sigma DAC.

Therefore, I contend that CDs are already Class-D (in a weird
topologically equivalent sort of way). For my next trick, I will prove
that black is white (and then get myself killed at the next crosswalk).

[William Sommerwerck]

Still, from a practical point of view, PCM is easier to implement
in a robust way than PWM (analog or digital).

But, but, but... In a PWM system, the width of the pulse is
proportional to the amplitude of the original analog signal. So,
if we built a system which did parallel PCM and PWM conversion
using the same sample clock, then measured the width of the
PWM output and encoded that result using the same resolution
as our PCM signal, wouldn't you end up with the same encoding
either way?

Yes, of course.

Therefore the two are exactly equivalent.

Thank you. That's what I've been trying to explain to these people, but they
won't listen.

More to the point, the encoding is equivalent at the point of quantization.
A quantized level is no different from the number that represents it.


Therefore, I contend that CDs are already Class-D (in a weird
topologically equivalent sort of way).

I object to the use of "Class" for anything other than it's original
purpose. But you're right, in terms of understanding the principles
involved.

The PCM on the disk, and a series of PWM pulses quantized to the values of
the PCM are fungible. They represent exactly the same data. What is so hard
to understand about this?

More to the point, why is the PCM digital, and the quantized pulses not?

[Anahata]

On Tue, 13 Oct 2009 05:47:26 -0700, William Sommerwerck wrote:

The PCM on the disk, and a series of PWM pulses quantized to the values
of the PCM are fungible. They represent exactly the same data. What is
so hard to understand about this?

That's only true in an imaginary world where there is no noise and the
surface of the disk is perfectly free of distortion, dirt and other
surface defects, and you have a very sensor and quantizer to measure the
pit length. You have to be able to measure each pulse length (a few
microns on the disk) with an accuracy of better than 1 part in 65536,
assuming you want to match the 16 bit resolution of conventional CD.

In practice those defects will push the measured pulse width of your PWM
version many quantized steps away from the correct value, so you'll end
up choosing some exact quantized value, but the wrong one.

With PCM you only have to distingush between a pit and the absence of a
pit, so far higher noise and dirt levels can be tolerated. You have to do
it 16 times instead of once per sample, but the chances are still much
higher of getting all of those 16 bits right.

--
Anahata
==//== 01638 720444
http://www.treewind.co.uk ==//== http://www.myspace.com/maryanahata