Thread: Should SA-CD.net take Blu-Ray Audio on board?

Posts: 352
Page: prev 1 ... 31 32 33 34 35 36

Post by tailspn August 27, 2014 (351 of 352)
off the grid said:

Hi Fitz,

Yup you are correct a classic case of being over worked and seeing what I hoped to see, not what was present.

Don't let the flatlanders dissuade you Neal, you're on the right track. The analogy of each bit time being a 1 or 0 is a marker for whether the signal is increasing, or decreasing is an ancient oversimplification, and incorrect. It's much more sophisticated than that, but still understandable.

While there are very few true 1-bit two level (DSD) A/D converters in practice today, and certainly not simple first order, there's an excellent explanation of the workings of a PDM first order modulator here:

http://www.embedded.com/design/debug-and-optimization/4406844/The-basics-of-sigma-delta-analog-to-digital-converters-

Try working through the 16 steps in Table 1 of the modulator as it seeks equilibrium, and you'll understand much more of how the process works than anyone advising you here. The actual first order modulator is within the dashed lines, and it's resolving a DC level of 3/8 the maximum positive level.

The key words of the modulators action are:

"Note that a repetitive pattern develops every sixteen samples, and that the average of the signal W over samples 1 to 16 is 3/8, thus showing that the feedback loop forces the average of the feedback signal W to be equals to the input X."

The output at D is analog derivative expressed in a sequence of 1's and 0's of the input signal (a 3/8 max DC level in this case) that can be recovered with a simple analog integrator (average of the density of the 1's and 0's bits).

Yes, of course there's a digital carrier, but the bit density envelope within that carrier is an analog signal.

Post by Fitzcaraldo215 August 28, 2014 (352 of 352)
tailspn said:

...

http://www.embedded.com/design/debug-and-optimization/4406844/The-basics-of-sigma-delta-analog-to-digital-converters-

...

The output at D is analog derivative expressed in a sequence of 1's and 0's of the input signal (a 3/8 max DC level in this case) that can be recovered with a simple analog integrator (average of the density of the 1's and 0's bits).

Yes, of course there's a digital carrier, but the bit density envelope within that carrier is an analog signal.

Tom - I am not here to fight with you, a guy I have tremendous respect and admiration for. I am just trying to make sure that the record left for posterity in this forum accurately reflects the undistorted, accurate truth. Everything I have seen from all sources indicates a universally accepted truth to the subject of this friendly, I hope, debate, so it is not just about opinions. There are accepted facts here beyond any opinion. To deny them is foolish.

As I look at the article you cited that pertains to a conceptual PDM analog to digital converter and further at the article cited by it, I see digital all over the place. First, the article is entitled The Basics of Sigma Delta Analog to DIGITAL Converters. You specifically cite point D in Figure 2, the modulator. At that point, the circuit flows to both a 1-bit DAC, that's a Digital-to-Analog Converter for us plain folk, or to a Digital Filter for final output. Ergo, the result of the modulator at point D must be digital. How you can reach a different conclusion that this is still about nothing but analog is beyond me. It is your citation, after all.

Yes, there may be analog steps in the process leading up to final output. There would have to be since the original input signal being converted by this conceptual ADC is analog. But, DSD/PDM are the digital RESULT of the a-to-d process, not the process itself, circuit diagram or the interim steps. DSD/PDM are a digital representation, as the PDM Wiki article definition states, not a specific process. The circuit block diagram only confirms this. Yet, you consistently seem befuddled by this, interpreting it only as you wish to interpret it.

That the digital output signal can, or more correctly could, be converted to analog via a simple analog integrator is, again, confusing the process with the definition of DSD/PDM. I have no doubt, as you say, that real world circuit diagrams and processes are more complex than this. I also said several posts ago that all electronic signals, even digital ones, are, if you look deeply enough, actually analog. Digital electronic transmission is a special case of analog transmission, where digital encodes its digital information on electronic flows, which are analog, as a series of discrete pulses corresponding to the values of either zero or one. It is so highly special a case that we refer to and distinguish it conceptually as the Digital Domain vs. the Analog Domain.

Yes, engineers can work in a circuit with a digital input, ignore its digital nature and treat it directly as an encoded analog electronic signal for subsequent analog processing steps. In the case of a DAC, that would lead up through the circuit stages to eventual recovery of a replica of the originally sampled analog signal. But, that ability to shift from the Digital to the Analog Domain applies to any and every digital signal input including PCM. At the micro circuit level, there is also much analog content involved in the a-d and d-a conversion of PCM. But, again, the in-circuit process, even if it shifts domains between digital and analog, does not change the definition of the thing itself. DSD/PDM and PCM, too, are all in the Digital Domain and they are digital representations of a discretely sampled analog signal. But, for you, it is "guilt by association". If the process has analog in it, the definition of the whole modulation scheme must be analog. That is false reasoning. You would have to say all digital is analog, including PCM, and that the whole Digital Domain does not exist. But, that gets us nowhere. Try telling that to Intel, IBM, Apple, etc., etc., etc.

That DSD/PDM is even conceptually easy to convert from its digital representation to analog is still confusing the process with the definition. But, conceptually easy does not even square with the real world. You have repeatedly cited Ted Smith's new and apparently very good PS Audio DAC. But, even a casual look at his block diagrams or circuit photos speaks to great complexity and much digital content in the circuit makeup. The notion that it is easy collapses quite quickly. And if it were so easy, why did it take so long to achieve, now that we are, what, about 15 years or so into DSD? You have inadvertently destroyed your own argument by citing his work.

So, Tom, I do not think this article provides any proof of your assertion that DSD is analog. Actually, it provides further proof to the contrary. I do not think you will find that proof you are seeking anywhere.

Wishing you the best,

Carl

Page: prev 1 ... 31 32 33 34 35 36

Closed