dsd DSD Offshoot Discussion From MQA Topic

[Miska]

22 minutes ago, crenca said:

Ok, thanks - I might be getting it (but probably not 😋)!  Yes by "processing" I meant modulation.  Now, what is distinguishing DSD from 1 bit PCM is DSD is modulation (both modulated at  "regular intervals" as the article says - the clock), but where as DSD is modulated into a analog waveform, PCM is demodulated into an analog waveform.  Is this correct to say?

 

No, both are demodulated into analog waveform.

 

Modulated form is the digital representation.

[crenca]

4 minutes ago, Miska said:

 

No, both are demodulated into analog waveform.

 

Modulated form is the digital representation.

 

Ok thanks.  So what kind of modulator is used to create DSD?

[Jud]

People know this can and does often start as analog, right?   

 

My impression is that what starts as analog is usually digitized first as SDM (just as I believe most DACs use SDM as the last step before analog). Most "PCM" these days is converted from SDM.  

 

Back in the day, digitization direct to "PCM" was the usual thing, with a few scattered Pacific Microsonics ADCs still working that way.  (I think I recall @manisandher having sold his to a Paris studio.)

[crenca]

12 minutes ago, Jud said:

People know this can and does often start as analog, right?   

 

My impression is that what starts as analog is usually digitized first as SDM (just as I believe most DACs use SDM as the last step before analog). Most "PCM" these days is converted from SDM.  

 

Back in the day, digitization direct to "PCM" was the usual thing, with a few scattered Pacific Microsonics ADCs still working that way.  (I think I recall @manisandher having sold his to a Paris studio.)

 

Miska seems to be saying that the modulator defines rather it really is PCM (of all bit depth) or PSM (DSD).  I think he simply means PWM vs. PCM...so the modulator spits out either or, but then I wonder if he is saying more that...trying to follow.

[Miska]

28 minutes ago, crenca said:

Ok thanks.  So what kind of modulator is used to create DSD?

 

Sigma-Delta (or Delta-Sigma) modulator. It can be two- (DSD) or more levels, like in all modern audio ADC and DAC devices.

 

In ADC, SD-modulator is hybrid analog-digital construction. If the ADC has PCM output, it is then converted to typical binary encoded integer LPCM. ADC can also have DSD or multi-level SDM output. You can see the increasing noise floor in modern ADC's PCM output that in most cases looks just like DSD128 converted to PCM.

 

In DAC, SD-modulator is purely digital used to convert PCM inputs into conversion stage's native format.

 

I myself try to avoid talking much about "DSD" because it is marketing term for end-to-end SDM delivery system. I rather talk about SDM or PWM or something that has technical meaning.

[manisandher]

21 minutes ago, Jud said:

(I think I recall @manisandher having sold his to a Paris studio.)

 

Yep. It seemed a total waste of its ability using it just for my modest needle drops. The Paris mastering studio was already very well equipped with the 'latest and best' of everything, but the guys there told me that the PM2 was now their favourite ADC... mainly for its bottom end. Due to being 'pure' PCM? Who knows?

 

Mani.

[Jud]

On 3/5/2019 at 2:15 PM, Jud said:

 

If you could simply give a good, correct real world example of PDM and PWM, it would be helpful to me, at least.

 

On 3/5/2019 at 2:16 PM, mansr said:

There's no point discussing anything while Miska is here. Sorry.

 

I asked for some examples to aid in understanding.  Your reaction seemed rather petulant to me.

 

44 minutes ago, mansr said:

I guess I should congratulate Miska on utterly confusing everybody.

 

And now you've stopped by seeming rather petulant again.  If people ask for help understanding and you don't provide it, I'm not sure you can blame someone else for any "confusion."

 

(Using quotes because what Miska is saying seems to me to align with the sources I've read up 'til now, which you are saying are all incorrect or unreliable. That's fine, but hard for me to accept without some type of alternative explanation. And I'm not motivated to buy an $84 book that I almost certainly don't have the math background to understand anyway, so without something considerably less expensive and more convenient, I'm stuck where I am.)

[Jud]

15 minutes ago, Miska said:

modern ADC's PCM output that in most cases looks just like DSD128 converted to PCM.

 

Starts as analog, converted initially to PDM/PWM in most cases, then PCM? At least that's what I remember Tom Caulfield the recording engineer saying.

[Miska]

44 minutes ago, Jud said:

Starts as analog, converted initially to PDM/PWM in most cases, then PCM? At least that's what I remember Tom Caulfield the recording engineer saying.

 

Yes, that's the case...

 

[fas42]

7 hours ago, Jud said:

 

Yes, that's what I meant. A lot of people are fans of filtering that doesn't cut much, including various NOS DACs, the three-letter word that spawned a thread that spawned this one, and the PONO and Ayre filters.  Does this presumably "warmer," more "organic" sound (at least for some people) provide an appeal that the producer and artist intend as part of the work?

 

"Warmer," more "organic" is actually the true nature of recordings, of all types - one has to deliberately distort at the recording stage to get a cold, "non-organic" feel to a sound.

 

The typical use of heavy digital processing, and inadequate filtering circuitry in the playback area injects artifacts in the sound, which interfere with the 'naturalness' of what one hears. One workaround is the use of NOS DACs, which may enable one to get closer to the true nature of the recording - at the expense of introducing other issues.

[Jud]

2 hours ago, Miska said:

 

Yes, that's the case...

 

 

Thanks!  Decimation and high pass for PCM, that makes sense.

[Don Hills]

10 hours ago, Miska said:

... PCM is not PWM, DSD/SDM is. ...

 

You have said several times that DSD is PWM. Are you sure you meant to say that?

(Assuming that "DSD is PWM" translates to "the result of encoding an analogue signal is a train of pulses of varying width corresponding to the value of the analogue input.")

[Miska]

3 hours ago, Don Hills said:

You have said several times that DSD is PWM. Are you sure you meant to say that?

 

SDM is PWM/PDM, yes...

 

Quote

(Assuming that "DSD is PWM" translates to "the result of encoding an analogue signal is a train of pulses of varying width corresponding to the value of the analogue input.")

 

It doesn't have to be "analogue". Again referring to (and Fig 4):

https://en.wikipedia.org/wiki/Pulse-width_modulation#Delta-sigma

 

Also:

https://en.wikipedia.org/wiki/Delta-sigma_modulation

 

P.S. I'm pretty sure "DSD" originally meant "Direct Sigma-Delta" or "Delta-Sigma Direct", before marketing department at Sony and Philips decided that it is probably not very consumer friendly so they came up with "Direct Stream Digital".

[fas42]

 

[Don Hills]

20 hours ago, Miska said:

 

SDM is PWM/PDM, yes...

...

Yes, but which is it closest to? PWM and PDM are two quite different modulation schemes.

[Miska]

2 hours ago, Don Hills said:

Yes, but which is it closest to? PWM and PDM are two quite different modulation schemes.

 

No they are not, they are essentially the same, only difference is interpretation of the data... You can think PDM as RTZ version of the same.

[Don Hills]

1 hour ago, Miska said:

 

No they are not, they are essentially the same, only difference is interpretation of the data... You can think PDM as RTZ version of the same.

 

Ah yes, I see what you mean. I come from a telecommunications background and see things somewhat differently... so PWM can be thought of as a limit case of PDM.

[Jud]

12 hours ago, Miska said:

 

No they are not, they are essentially the same, only difference is interpretation of the data... You can think PDM as RTZ version of the same.

 

For anyone unfamiliar (as I was before looking it up a moment ago), RTZ= Return To Zero after each cycle/pulse.