Bromo writes:

The recording does not have to have a brick wall filter - but a slow rolloff
filter - which will have much less phase effect. From what I remember about
my filter design classes, is that in the vast majority (all of them? I do
not recall) the phase starts rolling off sometime before the amplitude in a
filter response.

This is only true for analog filters - I'd be surprised to see any
modern digital audio component doing this.

In recording at 44.1kHz, with a "brick wall" style of filter, the phase
begins shifting around 1-2kHz.

But very few cd players are made this way these days. Even my
first CD player (A Philips - their first model) did this, and had
perfect phase response.

---Ketil

Bromo wrote:


When you get to soft passages, the quantization noise in digital is much
closer to the signal peaks - and at the high end of the spectrum (10-20kHz)
the effects of smearing are greater which is usually found to be sopranos,
snare drums, and violins can be very grating.

Aren't you ignoring dither, which is a key part of digital audio?

What smearing effects are you talking about?

On 6/13/04 8:46 PM, in article rz6zc.18160$Hg2.12062@attbi_s04, "chung"
wrote:

Bromo wrote:

On 6/12/04 12:48 PM, in article CtGyc.74669$3x.33764@attbi_s54, "chung"
wrote:

Most mid range
priced ($300) CD players made from c.1996 on later tend to have better
DACs
and decent analog stages.

Please provide examples justifying your claim. Show some test results of
inferior performance from CD players earlier than 1996, compared to
newer ones.

Why on earth should I do that, chung?


I was hoping that you could back up that claim.


Actually, you can go through the back issues of UHF magazine and compare the
200Hz square wave reproduction and the 1kHz sine wave at 60dB down. They
start c. 1990 and should prove the point handily.

Bromo wrote:
On 6/13/04 5:37 PM, in article eO3zc.17010$Hg2.13507@attbi_s04, "Steven
Sullivan" wrote:

Why would I want to do that? CD playback is 44.1/16 not 22/11.5.
Recording is more like 96/24 these days. The question is whether
*these* specs intrinsically and audibly 'mangles' frequencies from 10-20 kHz.

The recording does not have to have a brick wall filter - but a slow rolloff
filter - which will have much less phase effect. From what I remember about
my filter design classes, is that in the vast majority (all of them? I do
not recall) the phase starts rolling off sometime before the amplitude in a
filter response.

In recording at 44.1kHz, with a "brick wall" style of filter, the phase
begins shifting around 1-2kHz.

Well, let's test that hypothesis:

Let's build a 7th order Chebyshev filter, with a cut-off at 22 KHz. I
ran the simulation using SPICE, and here are the results:

Group delay at 10 Hz: 39.6957 microsec.
Group delay at 2 KHz: 39.4285 microsec.

Deviation from constant group delay at 2 KHz: -267.113 nanosec.

That corresponds to a phase error (from linear phase) of 0.2 degree.

You said that phase distortion is apparent at 1-2 KHz. Want to
substantiate that claim?


In 24/96, you can put the poles of the filter much, much higher than 20kHz
and make it a lower order filter to gently roll off which means the phase
shift happens at frequencies that are not audible.

Of course, your point is rather moot in the first place, since the vast
majority of CD players since the late '80's have used oversampling.

On 6/14/04 7:36 PM, in article , "chung"
wrote:

In 24/96, you can put the poles of the filter much, much higher than 20kHz
and make it a lower order filter to gently roll off which means the phase
shift happens at frequencies that are not audible.

Of course, your point is rather moot in the first place, since the vast
majority of CD players since the late '80's have used oversampling.

Sorry - I was referring to recording more than playback. I should have made
that more clear.

On 6/14/04 12:09 PM, in article K4kzc.22469$Hg2.15230@attbi_s04, "chung"
wrote:

Bromo wrote:


When you get to soft passages, the quantization noise in digital is much
closer to the signal peaks - and at the high end of the spectrum (10-20kHz)
the effects of smearing are greater which is usually found to be sopranos,
snare drums, and violins can be very grating.

Aren't you ignoring dither, which is a key part of digital audio?

What smearing effects are you talking about?


The sound, to me, of those seems to have unnatural tinny sounds - kind of
like having a nail driven into your head.

On 6/13/04 8:43 PM, in article , "chung"
wrote:

Frequency domain is described with *sinusoidal* basis waveforms.

True - and it is difficult to display non periodic transient information
with sine waves - you would end up with a frequency domain spectrum that
changes over time.

The original poster was wondering if sine waves were good representations of
music - and the answer is *yes* and *no*. This was before you felt the need
to jump on correct some of my errors and berate me.

On 6/13/04 8:43 PM, in article , "chung"
wrote:

You need to study your math books. Taylor series is a power series
expansion of a function around an operating point. Fourier analysis is
the transformation of a time-domian function to the frequency domain.
Frequency domain is described with *sinusoidal* basis waveforms. Of
course, in Fourier analysis, amplitudes of the sinusiodal functions can
be complex, and frequencies both negative and positive.

I am surprised that an RF engineer has trouble with these concepts

Oops. Strike Taylor and put in Fourier.

On 6/14/04 7:36 PM, in article , "chung"
wrote:

In 24/96, you can put the poles of the filter much, much higher than 20kHz
and make it a lower order filter to gently roll off which means the phase
shift happens at frequencies that are not audible.

Of course, your point is rather moot in the first place, since the vast
majority of CD players since the late '80's have used oversampling.

Actually, I was mostly talking about RECORDING signals digitally, not
decoding.

On 6/13/04 8:44 PM, in article , "chung"
wrote:
Note that the OP stated that DAC's mangled 5KHz signals. He did not say
that it was poor implementation of ADC's that mangled them.

And you seem to agree with him that the 5KHz signals are mangled.

Nope - just what that person would deem "mangled" might not be what you ro I
deem "mangled."

Mangled is a judgment, not a measurement.

I guess that could have been highly incompetent implementations that
screw up phase response at 1 or 2 KHz, but it is child's play nowadays
to not do that.

Actually, it is not - but the "work around" is to sample at 24 bit and 96 or
192kHz sample rate to allow the "brick wall" filter to be a slow roll off
filter.

Have you heard of oversampling, a technique that has been in use for 20
years?

Yes I have - and it is exceedingly hard to do on a recording :-P. It is
much easier on playback.


No, it has nothing to do with 24 bit or 96/192 KHz sampling rates, which
were obviously not common at all 20 years ago.

Of course not - it is removing a lot of the repeated information above
22.05kHz up to some frequency the designer deems appropriate. This allows
the use of cheap filters instead of expensive multi-pole filters.

Again, useful for a CD PLAYER - not so useful for a digital recorder. Best
to use a high sample rate for them.


Neither technique could be "child's play" - as it is more difficult and
technically challenging to get a decent digital recording than a analog
recording.

Wrong. You can get a decent digital recording much easier, and much
cheaper. Notice how many people have successfully recorded vinyl LP's to
CD's. Or notice how many have used the "lowly" minidisc recorders for
*decent* amatuer recordings in digital. You can easily get soundcards
that outperform analog recorders costing orders of magnitude more.

You are expressing an opinion. It all depends upon what you deem "decent"
and what I deem "decent" - the very best recordings I have heard have been
either with really high end digital recorders, or old style analog
machinery. To me this is "decent" - if you find the level of "decent" to be
Minidisc level - then so be it - you have achieved audio nirvana much easier
than many people.

Again - for me, at least, "decent recordings" are *not* child's play.

Also - which sound cards compared to which analog recorders?

On 6/14/04 7:36 PM, in article , "chung"
wrote:

In recording at 44.1kHz, with a "brick wall" style of filter, the phase
begins shifting around 1-2kHz.

Well, let's test that hypothesis:

Let's build a 7th order Chebyshev filter, with a cut-off at 22 KHz. I
ran the simulation using SPICE, and here are the results:

Group delay at 10 Hz: 39.6957 microsec.
Group delay at 2 KHz: 39.4285 microsec.

Deviation from constant group delay at 2 KHz: -267.113 nanosec.

That corresponds to a phase error (from linear phase) of 0.2 degree.

You said that phase distortion is apparent at 1-2 KHz. Want to
substantiate that claim?

I said the phase shift starts shifting at 1-2kHz. You showed a start of the
shift. Thanks.

Bromo wrote:

On 6/13/04 8:44 PM, in article , "chung"
wrote:
Note that the OP stated that DAC's mangled 5KHz signals. He did not say
that it was poor implementation of ADC's that mangled them.

And you seem to agree with him that the 5KHz signals are mangled.

Nope - just what that person would deem "mangled" might not be what you ro I
deem "mangled."

Mangled is a judgment, not a measurement.

So you think "mangled" cannot be supported by measurements.


I guess that could have been highly incompetent implementations that
screw up phase response at 1 or 2 KHz, but it is child's play nowadays
to not do that.

Actually, it is not - but the "work around" is to sample at 24 bit and 96 or
192kHz sample rate to allow the "brick wall" filter to be a slow roll off
filter.

Have you heard of oversampling, a technique that has been in use for 20
years?

Yes I have - and it is exceedingly hard to do on a recording :-P. It is
much easier on playback.

Exceedingly hard? Have you heard of oversampling ADC's? How about
sampling at 88.2K or 96K and downsample to 44.1K? All these have been in
use for a long time.



No, it has nothing to do with 24 bit or 96/192 KHz sampling rates, which
were obviously not common at all 20 years ago.

Of course not - it is removing a lot of the repeated information above
22.05kHz up to some frequency the designer deems appropriate. This allows
the use of cheap filters instead of expensive multi-pole filters.

Again, useful for a CD PLAYER - not so useful for a digital recorder. Best
to use a high sample rate for them.

So now you are changing your argument to the ADC's instead of CD
players. Have we agreed that the DAC in the CD player is not "mangling"
5 KHz waveforms?



Neither technique could be "child's play" - as it is more difficult and
technically challenging to get a decent digital recording than a analog
recording.

Wrong. You can get a decent digital recording much easier, and much
cheaper. Notice how many people have successfully recorded vinyl LP's to
CD's. Or notice how many have used the "lowly" minidisc recorders for
*decent* amatuer recordings in digital. You can easily get soundcards
that outperform analog recorders costing orders of magnitude more.

You are expressing an opinion. It all depends upon what you deem "decent"
and what I deem "decent" - the very best recordings I have heard have been
either with really high end digital recorders, or old style analog
machinery. To me this is "decent" - if you find the level of "decent" to be
Minidisc level - then so be it - you have achieved audio nirvana much easier
than many people.

How about +/- 0.2 dB from 20-20KHz, 100dB dynamic range? 44.1K/48K/96K
sampling? Unmeasureable wow and flutter?

Want to list your analog recorder measurements?


Again - for me, at least, "decent recordings" are *not* child's play.

Also - which sound cards compared to which analog recorders?

There is a host of prosumer sound cards now available for recording
applications. Check your local professional music stores. Or visit this
site:

http://www.pcavtech.com/soundcards/compare/index.htm

For example, look at the LynxTwo.

Now compare some of these top cards against your favorite analog recorders.

Bromo writes:

On 6/14/04 7:36 PM, in article , "chung"
wrote:

In 24/96, you can put the poles of the filter much, much higher than 20kHz
and make it a lower order filter to gently roll off which means the phase
shift happens at frequencies that are not audible.

Of course, your point is rather moot in the first place, since the vast
majority of CD players since the late '80's have used oversampling.

Sorry - I was referring to recording more than playback. I should have made
that more clear.

But then, doing "oversampling" at the recording end, and then
decimating the signal down to 44.1 is nothing new, either. What _is_
your point, exactly ?

---Ketil

Bromo wrote:

On 6/14/04 12:09 PM, in article K4kzc.22469$Hg2.15230@attbi_s04, "chung"
wrote:

Bromo wrote:


When you get to soft passages, the quantization noise in digital is much
closer to the signal peaks - and at the high end of the spectrum (10-20kHz)
the effects of smearing are greater which is usually found to be sopranos,
snare drums, and violins can be very grating.

Aren't you ignoring dither, which is a key part of digital audio?

What smearing effects are you talking about?


The sound, to me, of those seems to have unnatural tinny sounds - kind of
like having a nail driven into your head.


So you're saying all CD's have this effect on you? That's an amazing
claim. I'm starting to feel sorry for you .

Bromo wrote:

On 6/14/04 7:36 PM, in article , "chung"
wrote:

In recording at 44.1kHz, with a "brick wall" style of filter, the phase
begins shifting around 1-2kHz.

Well, let's test that hypothesis:

Let's build a 7th order Chebyshev filter, with a cut-off at 22 KHz. I
ran the simulation using SPICE, and here are the results:

Group delay at 10 Hz: 39.6957 microsec.
Group delay at 2 KHz: 39.4285 microsec.

Deviation from constant group delay at 2 KHz: -267.113 nanosec.

That corresponds to a phase error (from linear phase) of 0.2 degree.

You said that phase distortion is apparent at 1-2 KHz. Want to
substantiate that claim?

I said the phase shift starts shifting at 1-2kHz. You showed a start of the
shift. Thanks.

Actually, the phase shift starts at DC, for any filter, regardless of
how gentle the slope is, and regardless of whether you sample at 44.1K
or 192KHz. Note that you said "with the lower order filters the phase
shift happens at frequencies that are not audible".

At the risk of beating a dead horse, here is what you also said: "...the
phase distortion is apparent as low as 1-2kHz.".

Would you call a 0.2 degree phase error a distortion that is apparent? I
believe you even said you could see these errors on a scope, too. Can
you really see 0.2 degree phase difference?

On 6/15/04 11:13 AM, in article 1eEzc.54470$HG.37166@attbi_s53, "chung"
wrote:

Exceedingly hard? Have you heard of oversampling ADC's? How about
sampling at 88.2K or 96K and downsample to 44.1K? All these have been in
use for a long time.

Yup. They have. Recording in all of its aspects is much harder than
playback. Period.

On 6/15/04 6:47 PM, in article W%Kzc.55571$HG.45771@attbi_s53, "Ketil
Kirkerud Elgethun" wrote:

Bromo writes:

On 6/14/04 7:36 PM, in article , "chung"
wrote:

In 24/96, you can put the poles of the filter much, much higher than 20kHz
and make it a lower order filter to gently roll off which means the phase
shift happens at frequencies that are not audible.

Of course, your point is rather moot in the first place, since the vast
majority of CD players since the late '80's have used oversampling.

Sorry - I was referring to recording more than playback. I should have made
that more clear.

But then, doing "oversampling" at the recording end, and then
decimating the signal down to 44.1 is nothing new, either. What _is_
your point, exactly ?

That recording is more difficult than playback is all.

On 6/15/04 6:51 PM, in article , "chung"
wrote:

"with the lower order filters the phase
shift happens at frequencies that are not audible".

The reason for this - is that this was in the context of the higher sampling
rate - a lower order filter can be made that will have less phase shift.

Would you call a 0.2 degree phase error a distortion that is apparent? I
believe you even said you could see these errors on a scope, too. Can
you really see 0.2 degree phase difference?

I don't recall making that claim...

Bromo wrote:
On 6/13/04 8:46 PM, in article rz6zc.18160$Hg2.12062@attbi_s04, "chung"
wrote:

Bromo wrote:

On 6/12/04 12:48 PM, in article CtGyc.74669$3x.33764@attbi_s54, "chung"
wrote:

Most mid range
priced ($300) CD players made from c.1996 on later tend to have better
DACs
and decent analog stages.

Please provide examples justifying your claim. Show some test results of
inferior performance from CD players earlier than 1996, compared to
newer ones.

Why on earth should I do that, chung?


I was hoping that you could back up that claim.


Actually, you can go through the back issues of UHF magazine and compare the
200Hz square wave reproduction and the 1kHz sine wave at 60dB down. They
start c. 1990 and should prove the point handily.

Given that you appear to have access to that magazine and I don't, can
you share with us those measurement results that lead to your claim?

"chung" wrote in message
...
Bromo wrote:

On 6/14/04 12:09 PM, in article K4kzc.22469$Hg2.15230@attbi_s04, "chung"
wrote:

Bromo wrote:


When you get to soft passages, the quantization noise in digital is
much
closer to the signal peaks - and at the high end of the spectrum
(10-20kHz)
the effects of smearing are greater which is usually found to be
sopranos,
snare drums, and violins can be very grating.

Aren't you ignoring dither, which is a key part of digital audio?

What smearing effects are you talking about?


The sound, to me, of those seems to have unnatural tinny sounds - kind
of
like having a nail driven into your head.


So you're saying all CD's have this effect on you? That's an amazing
claim. I'm starting to feel sorry for you .

Actually, Chung, that is the way CD technology had many of us feeling, at
least until recently. And I can still hear it when switching from DVD-A or
SACD to CD, even with otherwise identical sounding mixes. Cymbals sound
tinny; snares too "sharp", violins a bit "screechy" compared both to live
sound and these other, superior forms of mastering and decoding.

And before you start, no I can't "prove" it. But I can tell you that it
kept me from buying a CD player for eight years, and when I did, it led me
to buying what is now considered a "classic", the Phillips 880, because of
its superior sound quality. Then to a Proceed external DAC with higher
resolution in combination with a DTI Pro Jitter Buster/Noise Shaper in order
to improve it even more, ever so slightly. But DVD-A and SACD are better
yet. And as you know from previous discussions, the Panasonic upsampling
also seems to get rid of the Nasties. I and Bromo are not alone; there is a
substantial segment of audiophilia out there who feel similarly.

On 6/15/04 6:50 PM, in article , "chung"
wrote:

The sound, to me, of those seems to have unnatural tinny sounds - kind of
like having a nail driven into your head.


So you're saying all CD's have this effect on you? That's an amazing
claim. I'm starting to feel sorry for you .

Heh. Either I have really good high frequency hearing, or really bad. When
playing with an tone generator, I can hear it as high as 18kHz. My wife can
hear out to 20-21kHz (or so she tells me when I play the tone and ask if she
can hear it).

On 6/15/04 11:13 AM, in article 1eEzc.54470$HG.37166@attbi_s53, "chung"
wrote:

Nope - just what that person would deem "mangled" might not be what you ro I
deem "mangled."

Mangled is a judgment, not a measurement.

So you think "mangled" cannot be supported by measurements.

I am unwilling to agree or disagree with his judgment - I do not see a plot
of the sine wave, nor do I have a clear definition of "mangled."

Without such an exact definition and the supporting data, all I can conclude
is that he does not like it for whatever reason, whatever level it might
have been.

For me, personally, I like subtractive synthesizers so "mangling" sinewaves
at all frequencies (deliberately) is hours of entertainment!

Bromo wrote:

On 6/15/04 6:51 PM, in article , "chung"
wrote:

"with the lower order filters the phase
shift happens at frequencies that are not audible".

The reason for this - is that this was in the context of the higher sampling
rate - a lower order filter can be made that will have less phase shift.

Would you call a 0.2 degree phase error a distortion that is apparent? I
believe you even said you could see these errors on a scope, too. Can
you really see 0.2 degree phase difference?

I don't recall making that claim...


Here's what you said on 6/12:

"If you were to look at a waveform on an oscilliscope out of a typical
CD player - you will notice errors on the sine wave - even in a player
that is not broken".

You also said that "...phase distortion is apparent as low as 1-2 KHz"
in a different post on 6/12.

On 6/16/04 6:24 PM, in article , "chung"
wrote:

Would you call a 0.2 degree phase error a distortion that is apparent? I
believe you even said you could see these errors on a scope, too. Can
you really see 0.2 degree phase difference?

I don't recall making that claim...


Here's what you said on 6/12:

"If you were to look at a waveform on an oscilliscope out of a typical
CD player - you will notice errors on the sine wave - even in a player
that is not broken".

Sure - made no claims that one could *hear* it - just *see* it.

You also said that "...phase distortion is apparent as low as 1-2 KHz"
in a different post on 6/12.

But made no claim if one could or could not hear it.

On 6/16/04 11:16 AM, in article 6vZzc.50093$0y.7184@attbi_s03, "chung"
wrote:

I was hoping that you could back up that claim.


Actually, you can go through the back issues of UHF magazine and compare the
200Hz square wave reproduction and the 1kHz sine wave at 60dB down. They
start c. 1990 and should prove the point handily.

Given that you appear to have access to that magazine and I don't, can
you share with us those measurement results that lead to your claim?

I would recommend you going to www.uhfmag.com and find for yourself.

Good mag, BTW.

Harry Lavo wrote:

"chung" wrote in message
...
Bromo wrote:

On 6/14/04 12:09 PM, in article K4kzc.22469$Hg2.15230@attbi_s04, "chung"
wrote:

Bromo wrote:


When you get to soft passages, the quantization noise in digital is
much
closer to the signal peaks - and at the high end of the spectrum
(10-20kHz)
the effects of smearing are greater which is usually found to be
sopranos,
snare drums, and violins can be very grating.

Aren't you ignoring dither, which is a key part of digital audio?

What smearing effects are you talking about?


The sound, to me, of those seems to have unnatural tinny sounds - kind
of
like having a nail driven into your head.


So you're saying all CD's have this effect on you? That's an amazing
claim. I'm starting to feel sorry for you .

Actually, Chung, that is the way CD technology had many of us feeling, at
least until recently. And I can still hear it when switching from DVD-A or
SACD to CD, even with otherwise identical sounding mixes.

How do you know that they are otherwise identical?

Cymbals sound
tinny; snares too "sharp", violins a bit "screechy" compared both to live
sound and these other, superior forms of mastering and decoding.

Hmmm, "forms of mastering". Does that mean the skills of the mastering
engineer? Or do you simply believe that redbook CD is incapable of
producing a sound that you like? If it's the latter, would a CD recorded
from a SACD player's output change your mind, if you can't tell them apart?

Not sure what you meant by forms of decoding. Redbook CD is linear PCM.


And before you start, no I can't "prove" it.
But I can tell you that it
kept me from buying a CD player for eight years, and when I did, it led me
to buying what is now considered a "classic", the Phillips 880, because of
its superior sound quality. Then to a Proceed external DAC with higher
resolution in combination with a DTI Pro Jitter Buster/Noise Shaper in order
to improve it even more, ever so slightly. But DVD-A and SACD are better
yet. And as you know from previous discussions, the Panasonic upsampling
also seems to get rid of the Nasties.

If it gets rid of the nasties, then the nasties are not in the CD
recording. That appears to be a different issue than what Mr. Bromo
seems to be saying: that all CD's with high frequency content hurt him
and that the problem was with the recording.

I and Bromo are not alone; there is a
substantial segment of audiophilia out there who feel similarly.

Well, I am not sure about that. Mr. Bromo is the first one who said that
listening to CDs' with high freqeuncy content is like having a nail
driven into his head. I certainly do not know of anyone else who feels
that way.

I just got some original 1969 german pressings of I and II and when
they arrive I will compare them (they were sold as mint)

I look forward to you report. Please let us know exactly what you are comparing
them to. There are several variations of the domestic issues of these LPs. The
best ones are the ones mastered by Bob Ludwig. They can be identified by the
initials RL at the end of the stamper numbers.


I just played the 1969 led zeppelin I and II german pressings and the
sound is more like I remembered from my old vinyl:
eq is different with more bass less high-mids and less top.

The classic reissues seem to have a wider stereo image, clearer reverb
tails (more present than 1969 pressings which are drier and darker)
sharper transients but maybe a little too much 3-5 KHZ emphasis? don't
know if this was on the master tapes or if it was eq'd in.
the classic records are more revealing overall.

The 1969 pressings seem to place more important parts foreward and
backround parts get lost a little.
My experience tells me that this could be a result of passing through
early solid state devices, that seem to have this effect on a full
mix, but it is only a guess.

the only advantage to the 1969 pressings is in the low mids, where the
sounds sem to be a bit fuller (the bass sounds a bit more round).
the Classics sound more like a bass sound that would come off a master
tape with no processing, but again a first guess.

Bromo wrote:

On 6/16/04 6:24 PM, in article , "chung"
wrote:

Would you call a 0.2 degree phase error a distortion that is apparent? I
believe you even said you could see these errors on a scope, too. Can
you really see 0.2 degree phase difference?

I don't recall making that claim...


Here's what you said on 6/12:

"If you were to look at a waveform on an oscilliscope out of a typical
CD player - you will notice errors on the sine wave - even in a player
that is not broken".

Sure - made no claims that one could *hear* it - just *see* it.

Yes, and that was what I said you claimed: that you can see it. Can you
see this on a scope?


You also said that "...phase distortion is apparent as low as 1-2 KHz"
in a different post on 6/12.

But made no claim if one could or could not hear it.


Yes, that was what I said you claimed: that you can see it. Of course if
you don't think you can hear it, one has to wonder why you even brought
up the pahse distortion at 1-2KHz in the first place...

Bromo wrote:

On 6/16/04 11:16 AM, in article 6vZzc.50093$0y.7184@attbi_s03, "chung"
wrote:

I was hoping that you could back up that claim.


Actually, you can go through the back issues of UHF magazine and compare the
200Hz square wave reproduction and the 1kHz sine wave at 60dB down. They
start c. 1990 and should prove the point handily.

Given that you appear to have access to that magazine and I don't, can
you share with us those measurement results that lead to your claim?

I would recommend you going to www.uhfmag.com and find for yourself.

Good mag, BTW.

They do not have any articles available with measurements on c. 1990 CD
players. I was hoping that since you seem to base your claim on some of
these articles, you can provide those results that support your claim.

Do you remember how are the square wave measurements different? And how
are the -60dB sinewaves different?

chung wrote:
Harry Lavo wrote:

"chung" wrote in message
...
Bromo wrote:

On 6/14/04 12:09 PM, in article K4kzc.22469$Hg2.15230@attbi_s04, "chung"
wrote:

Bromo wrote:


When you get to soft passages, the quantization noise in digital is
much
closer to the signal peaks - and at the high end of the spectrum
(10-20kHz)
the effects of smearing are greater which is usually found to be
sopranos,
snare drums, and violins can be very grating.

Aren't you ignoring dither, which is a key part of digital audio?

What smearing effects are you talking about?


The sound, to me, of those seems to have unnatural tinny sounds - kind
of
like having a nail driven into your head.


So you're saying all CD's have this effect on you? That's an amazing
claim. I'm starting to feel sorry for you .

Actually, Chung, that is the way CD technology had many of us feeling, at
least until recently. And I can still hear it when switching from DVD-A or
SACD to CD, even with otherwise identical sounding mixes.

How do you know that they are otherwise identical?

Heck, if he's not doing it blind, how does he know he's not simply
experiencing expectation bias? (Assuming too that his perception
of difference arose from lengthy, yet non-comparative,
listening that employed only the proper parts of his brain).


Cymbals sound
tinny; snares too "sharp", violins a bit "screechy" compared both to live
sound and these other, superior forms of mastering and decoding.

Hmmm, "forms of mastering". Does that mean the skills of the mastering
engineer? Or do you simply believe that redbook CD is incapable of
producing a sound that you like? If it's the latter, would a CD recorded
from a SACD player's output change your mind, if you can't tell them apart?

Even mastering engineers tend to say that *other* factors typical to mastering
-- such as application of EQ, NR, and choice of source tape --
are *far* more likely to generate audible difference, than digital
'encoding/decoding'.

Not sure what you meant by forms of decoding. Redbook CD is linear PCM.

And too, one would want to be sure that Harry's SACD, DVD-A , and CD
playback settings and processing paths treat the formats the same.
Which is damned hard to test at home. Even if I set all speakers to
large (eliminating any variation in crossover frequencies and slopes),
and listen only to the stereo tracks (eliminating differences in
time alignment implementation), and set the speaker levels the same --
how do I know that th player doesn't, say, output SACD's a little
louder than other formats? One coudl measure the outputs,
assuming one had a reference disc that has been verified to
contain the exact same test track, recorded at the same level
and EQ, in all formats


I and Bromo are not alone; there is a
substantial segment of audiophilia out there who feel similarly.

Well, I am not sure about that. Mr. Bromo is the first one who said that
listening to CDs' with high freqeuncy content is like having a nail
driven into his head. I certainly do not know of anyone else who feels
that way.

Besides, audiophilia long ago forfeited credibility as an arbiter
of what effects are real.


--

-S.
Why don't you just admit that you hate music and leave people alone. --
spiffy

On 6/17/04 10:06 AM, in article vzhAc.106959$3x.9172@attbi_s54, "chung"
wrote:

Cymbals sound
tinny; snares too "sharp", violins a bit "screechy" compared both to live
sound and these other, superior forms of mastering and decoding.

Hmmm, "forms of mastering". Does that mean the skills of the mastering
engineer? Or do you simply believe that redbook CD is incapable of
producing a sound that you like? If it's the latter, would a CD recorded
from a SACD player's output change your mind, if you can't tell them apart?

Not sure what you meant by forms of decoding. Redbook CD is linear PCM.

I know that there are a lot of bad recodings on CD - and with equalization
and the compression people do - a lot of pop recodings are processed and
recorded badly.

A good CD sounds fine usually - though for some reason - seems to not matter
the player - the Anonymous4 sound grating recorded and divine in person. I
think it is mostly the recording , though many people think they spend a lot
of time in the frequency ranges that are difficult to reproduce regardless
of the medium

From: (maxdm)
Date: 6/17/2004 9:54 AM Pacific Standard Time
Message-id: a1kAc.46870$2i5.17355@attbi_s52

I just got some original 1969 german pressings of I and II and when
they arrive I will compare them (they were sold as mint)

I look forward to you report. Please let us know exactly what you are
comparing
them to. There are several variations of the domestic issues of these LPs.
The
best ones are the ones mastered by Bob Ludwig. They can be identified by
the
initials RL at the end of the stamper numbers.


I just played the 1969 led zeppelin I and II german pressings and the
sound is more like I remembered from my old vinyl:
eq is different with more bass less high-mids and less top.

The classic reissues seem to have a wider stereo image, clearer reverb
tails (more present than 1969 pressings which are drier and darker)
sharper transients but maybe a little too much 3-5 KHZ emphasis? don't
know if this was on the master tapes or if it was eq'd in.
the classic records are more revealing overall.

The 1969 pressings seem to place more important parts foreward and
backround parts get lost a little.
My experience tells me that this could be a result of passing through
early solid state devices, that seem to have this effect on a full
mix, but it is only a guess.

the only advantage to the 1969 pressings is in the low mids, where the
sounds sem to be a bit fuller (the bass sounds a bit more round).
the Classics sound more like a bass sound that would come off a master
tape with no processing, but again a first guess.








The common complaint about the first two Led Zeppelin LPs on Classicsis that
they are too bright. It seems your observations go with that complaint. It
might be worth your while to seek out the Bob Ludwig domestic pressings. They
have all the qualies of the Classics by and large without the brightness.

Bromo wrote:

On 6/17/04 10:06 AM, in article vzhAc.106959$3x.9172@attbi_s54, "chung"
wrote:

Cymbals sound
tinny; snares too "sharp", violins a bit "screechy" compared both to live
sound and these other, superior forms of mastering and decoding.

Hmmm, "forms of mastering". Does that mean the skills of the mastering
engineer? Or do you simply believe that redbook CD is incapable of
producing a sound that you like? If it's the latter, would a CD recorded
from a SACD player's output change your mind, if you can't tell them apart?

Not sure what you meant by forms of decoding. Redbook CD is linear PCM.

I know that there are a lot of bad recodings on CD - and with equalization
and the compression people do - a lot of pop recodings are processed and
recorded badly.

OK, so are you now saying that it is the mastering of the CD's that give
you the headaches, not the technology?


A good CD sounds fine usually - though for some reason - seems to not matter
the player - the Anonymous4 sound grating recorded and divine in person. I
think it is mostly the recording , though many people think they spend a lot
of time in the frequency ranges that are difficult to reproduce regardless
of the medium

Why is it hard to reproduce those frequencies in CD, SACD or DVD-A media?

On 6/17/04 6:30 PM, in article , "chung"
wrote:

Bromo wrote:

On 6/16/04 11:16 AM, in article 6vZzc.50093$0y.7184@attbi_s03, "chung"
wrote:

I was hoping that you could back up that claim.


Actually, you can go through the back issues of UHF magazine and compare
the
200Hz square wave reproduction and the 1kHz sine wave at 60dB down. They
start c. 1990 and should prove the point handily.

Given that you appear to have access to that magazine and I don't, can
you share with us those measurement results that lead to your claim?

I would recommend you going to www.uhfmag.com and find for yourself.

Good mag, BTW.

They do not have any articles available with measurements on c. 1990 CD
players. I was hoping that since you seem to base your claim on some of
these articles, you can provide those results that support your claim.

Do you remember how are the square wave measurements different? And how
are the -60dB sinewaves different?

IN the latest issue the Creek CD50 mk2 player had no ringing in the square
wave, and the -60dB sine wave had no visible distortion. This was the best
measurement I recall seeing - since most of the square waves look like they
are missing some harmonic information - and ringing somewhat. I have no
idea how Creek is doing it - but there they are.

I recently set up by c. 1985 Magnavox CD player next to the NAD 541i. It
will be interesting to see how they;re different.

On 6/17/04 6:28 PM, in article , "chung"
wrote:

Bromo wrote:

On 6/16/04 6:24 PM, in article , "chung"
wrote:

Would you call a 0.2 degree phase error a distortion that is apparent? I
believe you even said you could see these errors on a scope, too. Can
you really see 0.2 degree phase difference?

I don't recall making that claim...


Here's what you said on 6/12:

"If you were to look at a waveform on an oscilliscope out of a typical
CD player - you will notice errors on the sine wave - even in a player
that is not broken".

Sure - made no claims that one could *hear* it - just *see* it.

Yes, and that was what I said you claimed: that you can see it. Can you
see this on a scope?

Yes, that would be how.

You also said that "...phase distortion is apparent as low as 1-2 KHz"
in a different post on 6/12.

But made no claim if one could or could not hear it.


Yes, that was what I said you claimed: that you can see it. Of course if
you don't think you can hear it, one has to wonder why you even brought
up the pahse distortion at 1-2KHz in the first place...

I don't know if it is phase distortion that causes the high end of the
sounds to sound raqther ... Well. Like a drill going through my head in some
recordings. The sounds are about 4kHz, though.

I got the 1-2kHz numbers from an article in Stereophile (IIRC) where they
did an interview with a recording engineer who had been around since before
the advent of digital and was really gung ho on the medium (and still is),
though critical of the difficulties in early recording and how it looks
now...

Harry Lavo wrote:

... [On CD] Cymbals sound
tinny; snares too "sharp", violins a bit "screechy" compared both to live
sound and these other, superior forms of mastering and decoding.

I have to agree with this, and those are precisely the
instruments that I hear as affected. To me the effect is much
less on recent CDs and some may not have it.

I make no claims as to where this problem comes from. I have
suspected peaky microphones, poor electronics in the recording
chain, overload (or poor performance) of the ADC used to make
the recording, and other problems. Whatever it is, it seems
to be accentuated by the frequency balance of CDs, which tend
to be mastered hotter in the high frequencies than LPs were.

The few DVD-A that I've heard did not exhibit this problem.


Mike Prager
North Carolina, USA

On 6/18/04 10:40 AM, in article _9DAc.133990$Ly.5003@attbi_s01, "chung"
wrote:

I know that there are a lot of bad recodings on CD - and with equalization
and the compression people do - a lot of pop recodings are processed and
recorded badly.

OK, so are you now saying that it is the mastering of the CD's that give
you the headaches, not the technology?

The technology and the medium are hardly perfect. Evan on a well recorded
CD - it sometimes is hard.

A good CD sounds fine usually - though for some reason - seems to not matter
the player - the Anonymous4 sound grating recorded and divine in person. I
think it is mostly the recording , though many people think they spend a lot
of time in the frequency ranges that are difficult to reproduce regardless
of the medium

Why is it hard to reproduce those frequencies in CD, SACD or DVD-A media?

I have no idea. I am speculating.

Bromo wrote:

On 6/17/04 6:28 PM, in article , "chung"
wrote:

Bromo wrote:

On 6/16/04 6:24 PM, in article , "chung"
wrote:

Would you call a 0.2 degree phase error a distortion that is apparent? I
believe you even said you could see these errors on a scope, too. Can
you really see 0.2 degree phase difference?

I don't recall making that claim...


Here's what you said on 6/12:

"If you were to look at a waveform on an oscilliscope out of a typical
CD player - you will notice errors on the sine wave - even in a player
that is not broken".

Sure - made no claims that one could *hear* it - just *see* it.

Yes, and that was what I said you claimed: that you can see it. Can you
see this on a scope?

Yes, that would be how.

Are you saying that you can see 0.2 degree phase distortion on a scope?

What does the "error" on a sinewave coming out of a CD player look like?
Measrements show that harmonics are down 90 dB on a full-scale sine
wave. And a properly dithered recording would have very little
distortion even when the amplitude of the sinewave is small.

Have you seen a sinewave out of an LP? Ever compared that against a
sinewave from a CD?


You also said that "...phase distortion is apparent as low as 1-2 KHz"
in a different post on 6/12.

But made no claim if one could or could not hear it.


Yes, that was what I said you claimed: that you can see it. Of course if
you don't think you can hear it, one has to wonder why you even brought
up the pahse distortion at 1-2KHz in the first place...

I don't know if it is phase distortion that causes the high end of the
sounds to sound raqther ... Well. Like a drill going through my head in some
recordings. The sounds are about 4kHz, though.

I got the 1-2kHz numbers from an article in Stereophile (IIRC) where they
did an interview with a recording engineer who had been around since before
the advent of digital and was really gung ho on the medium (and still is),
though critical of the difficulties in early recording and how it looks
now...

So, as an RF engineer you just believe everything those guys say? Why
not believe in me instead?

Bromo wrote:

On 6/17/04 6:30 PM, in article , "chung"
wrote:

Bromo wrote:

On 6/16/04 11:16 AM, in article 6vZzc.50093$0y.7184@attbi_s03, "chung"
wrote:

I was hoping that you could back up that claim.


Actually, you can go through the back issues of UHF magazine and compare
the
200Hz square wave reproduction and the 1kHz sine wave at 60dB down. They
start c. 1990 and should prove the point handily.

Given that you appear to have access to that magazine and I don't, can
you share with us those measurement results that lead to your claim?

I would recommend you going to www.uhfmag.com and find for yourself.

Good mag, BTW.

They do not have any articles available with measurements on c. 1990 CD
players. I was hoping that since you seem to base your claim on some of
these articles, you can provide those results that support your claim.

Do you remember how are the square wave measurements different? And how
are the -60dB sinewaves different?

IN the latest issue the Creek CD50 mk2 player had no ringing in the square
wave, and the -60dB sine wave had no visible distortion. This was the best
measurement I recall seeing - since most of the square waves look like they
are missing some harmonic information - and ringing somewhat. I have no
idea how Creek is doing it - but there they are.

First of all, a Creek CD50 mk2 at $1,500 is not your average CD player.

Second, I don't see any distortion for a -60dB sinewave from a c.1990 CD
player, assuming that the recording has dither applied.

Third, you have to be careful when you said there is no ringing. Are you
comparing two CD players playing the same square wave? A 200 Hz sq. wave
will look more "ideal" than a 1KHz sq. wave, from any CD player. The
ringing *has* to be there if the CD player is working correctly, because
there is no information above 20 KHz. In other words, some harmonics
*have* to be missing. You cannot say that one CD player has a better DAC
or output stage just by comparing square waves. You have to know what
the perfectly reproduced square wave looks like.

Seems to me that your claim is not supported yet, given the info you
provided.


I recently set up by c. 1985 Magnavox CD player next to the NAD 541i. It
will be interesting to see how they;re different.

Bromo wrote:
Do you remember how are the square wave measurements different? And
how are the -60dB sinewaves different?

IN the latest issue the Creek CD50 mk2 player had no ringing in the
square wave, and the -60dB sine wave had no visible distortion. This
was the best measurement I recall seeing - since most of the square
waves look like they are missing some harmonic information - and
ringing somewhat. I have no idea how Creek is doing it - but there
they are.

I recently set up by c. 1985 Magnavox CD player next to the NAD 541i.
It will be interesting to see how they;re different.

*All* square waves will be missing harmonic information because the signal
is band limited, it is just a question of how close you zoom into the edges
and how low you choose the frequency of the square wave. The output filter
will make the ringing and it is pretty easy to produce very little overshoot
with proper filter topologies. With the oversampling technique this should
not be too difficult. But then too, as long as the ringing is above the
hearing range it will not matter to the sound.
Square waves and dirac pulses are also not part of a music material, even
our ears cannot perceive them fully. They are great for testing and theory
tho. But a difference will not automatically be audible.

--
ciao Ban
Bordighera, Italy