glen herrmannsfeldt wrote:
> Jerry Avins wrote:

   ...

>> My confusion. I took your meaning of compression to be data 
>> compression, not dynamic range compression. Now that I understand what 
>> you meant, I don't understand how that's relevant to this discussion.
> 
> I am not sure either what the OP meant.  I just wanted to indicate
> that the two meanings are not unrelated.

How are they related?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

glen herrmannsfeldt wrote:
> Jerry Avins wrote:
> 
>> glen herrmannsfeldt wrote:
> 
>>> Without oversampling, as would have been the case in the early years,
>>> you have from 20kHz to 22.05kHz, which is not easy, but not impossible
>>> for an analog (RLC) filter.  Is there any evidence that the CD designers
>>> considered digital filtering as part of the standard?
> 
>> The CDs are just the same now as they originally were. The 
>> oversampling you refer to is not an increased sample rate of the data 
>> on the CD, but an implementation detail of the D/A hardware. But for 
>> its advantage as an obfuscating marketing ploy, a different term would 
>> have been used to describe the D/A converter in your player.
> 
> I usually think of it as a way to make good filters cheaper.
> With the ever increasing density of digital ICs, and analog circuitry
> not changing nearly as fast, it makes sense.  I am wondering if this
> occurred to those designing the original CD standard.

Whatever its purpose, it's not the _signal_ that's oversampled. Naive 
explanations of sampling reproduce the output signal by connecting the 
"dots", so having more of them make a smoother curve, and "oversampling" 
seems to imply that the sample instances _of_the_signal_ are closer 
together. T'ain't so. The oversampling here is a way to make a better 
reconstruction filter. It's a different technique that should have a 
different name.

> My belief is that they didn't, partly because of the BCD track
> numbers.  They weren't considering how fast digital electronics
> would change.

The filter problem wasn't quite as bad ad we make it seem. The 
contemplated upper frequency was about 18 KHz. The BCD track numbers 
must have been dictated by some management ignoramus. Converting hex to 
BCD in hardware takes only a handful of gates, and that's the hard way. 
I suspect it was the same sort of decision as the one by IBM that a PC 
would never need more than 640K of RAM.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Jerry Avins wrote:

> glen herrmannsfeldt wrote:
> 
>> Jerry Avins wrote:

(snip)
>>> Straight PCM augmented with error-correction redundancy. Where's the 
>>> compression?

>> As far as I know, rock music is often compressed.  That is,
>> in the audio sense not the data processing sense.  It helps
>> it sound loud, even at lower amplifier power levels.

>> With a low dynamic range, it might just as well be
>> compressed at the digital signal level.

> My confusion. I took your meaning of compression to be data compression, 
> not dynamic range compression. Now that I understand what you meant, I 
> don't understand how that's relevant to this discussion.

I am not sure either what the OP meant.  I just wanted to indicate
that the two meanings are not unrelated.

-- glen

Jerry Avins wrote:

> glen herrmannsfeldt wrote:

>> Without oversampling, as would have been the case in the early years,
>> you have from 20kHz to 22.05kHz, which is not easy, but not impossible
>> for an analog (RLC) filter.  Is there any evidence that the CD designers
>> considered digital filtering as part of the standard?

> The CDs are just the same now as they originally were. The oversampling 
> you refer to is not an increased sample rate of the data on the CD, but 
> an implementation detail of the D/A hardware. But for its advantage as 
> an obfuscating marketing ploy, a different term would have been used to 
> describe the D/A converter in your player.

I usually think of it as a way to make good filters cheaper.
With the ever increasing density of digital ICs, and analog circuitry
not changing nearly as fast, it makes sense.  I am wondering if this
occurred to those designing the original CD standard.

My belief is that they didn't, partly because of the BCD track
numbers.  They weren't considering how fast digital electronics
would change.

-- glen

glen herrmannsfeldt wrote:
> Jerry Avins wrote:
> 
>> Ron N. wrote:
> 
> (snip)
> 
>>>                                These days, it might be
>>> better to think of the data on an audio CD as a fixed rate
>>> compression format, which is usually seriously processed
>>> in the digital domain (maybe not as much as mp3) for
>>> input and output.
> 
>> Straight PCM augmented with error-correction redundancy. Where's the 
>> compression?
> 
> As far as I know, rock music is often compressed.  That is,
> in the audio sense not the data processing sense.  It helps
> it sound loud, even at lower amplifier power levels.
> 
> With a low dynamic range, it might just as well be
> compressed at the digital signal level.

My confusion. I took your meaning of compression to be data compression, 
not dynamic range compression. Now that I understand what you meant, I 
don't understand how that's relevant to this discussion.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

glen herrmannsfeldt wrote:

   ...

> Without oversampling, as would have been the case in the early years,
> you have from 20kHz to 22.05kHz, which is not easy, but not impossible
> for an analog (RLC) filter.  Is there any evidence that the CD designers
> considered digital filtering as part of the standard?

The CDs are just the same now as they originally were. The oversampling 
you refer to is not an increased sample rate of the data on the CD, but 
an implementation detail of the D/A hardware. But for its advantage as 
an obfuscating marketing ploy, a different term would have been used to 
describe the D/A converter in your player.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

glen herrmannsfeldt wrote:

> Without oversampling, as would have been the case in the early years,
> you have from 20kHz to 22.05kHz, which is not easy, but not impossible
> for an analog (RLC) filter.

Typically, the cutoff is set to 80...85% of Fs/2, which is about 18kHz 
for 44.1kHz system. It is indeed possible to build the analog elliptic 
filter of the 9th order which provides for more then 90dB of alias 
attenuation. However the only reason to attenuate the aliases is because 
of the nonlinear distortion in the analog power amplifiers. There is 
actually no need to have that much of attenuation. Somewhat 20dB at 
44.1kHz is sufficient. A simple 3rd order analog filter will do it.


   Is there any evidence that the CD designers
> considered digital filtering as part of the standard?

The early CD audio DACs were conventional R-2R structures without any 
oversampling.

Vladimir Vassilevsky

DSP and Mixed Signal Design Consultant

http://www.abvolt.com

Jerry Avins wrote:

> Ron N. wrote:

(snip)

>>                                These days, it might be
>> better to think of the data on an audio CD as a fixed rate
>> compression format, which is usually seriously processed
>> in the digital domain (maybe not as much as mp3) for
>> input and output.

> Straight PCM augmented with error-correction redundancy. 
> Where's the compression?

As far as I know, rock music is often compressed.  That is,
in the audio sense not the data processing sense.  It helps
it sound loud, even at lower amplifier power levels.

With a low dynamic range, it might just as well be
compressed at the digital signal level.

-- glen

Vladimir Vassilevsky wrote:
(someone wrote)

>> ONLY 8X oversampling??  all of the cheap sigma-delta chips have higher
>> oversampling than that.  you MUST have higher oversampling than 8X for
>> decent 1-bit D/A and any cheap CD player is using off-the-shelf 1-bit
>> D/A converters, no?  (they are oversampling by 64x or 128x, i think.)

I believe, though, that oversampling, in the CD player sense, came 
later.  My first CD player was from 1985, and I don't remember any
discussion about oversampling at the time.

(snip)

>> oh, yeah.  but you gotta oversample more than 8x to do decent digital
>> brickwall filtering and expect a passive RC filter to take care of the
>> analog filtering (and eliminate the images at 128x and above).

> This is actually an interesting question. The passive 1-st order RC 
> should have a cutoff at about 30kHz. The 128x oversampling will result 
> in the alias attenuation of ~40dB. In addition to that, there will be a 
> lot of noise shaping residual in the near ultrasound, that's why the 
> simple RC is not adequate.

Without oversampling, as would have been the case in the early years,
you have from 20kHz to 22.05kHz, which is not easy, but not impossible
for an analog (RLC) filter.  Is there any evidence that the CD designers
considered digital filtering as part of the standard?

-- glen

Ron N. wrote:

> The CD does not contain the data from the sampler,

Of course it does. The original recording rate is irrelevant. Its being 
a mix of several microphones originally recorded on different tracks is 
irrelevant. What's on the CD, provided that an ordinary CD player can 
reproduce it, is two channels of straight 16-bit PCM at a 44.1 KHz 
sample rate each, error-correcting codes, and format information.

> it contains some processed data from potentially a very
> different sample rate, which might be played back
> at yet another sample rate.

The playback rate is fixed at 44.1 KHz.

>                                These days, it might be
> better to think of the data on an audio CD as a fixed rate
> compression format, which is usually seriously processed
> in the digital domain (maybe not as much as mp3) for
> input and output.

Straight PCM augmented with error-correction redundancy. Where's the 
compression?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������