DSD comparison to PCM.

[Orb]

Nope, SC is used to reduce jitter sensitivity, but I'm with Ragnar on this - never heard a CMOS opamp I've liked the sound of.

I think BB/TI moved away from multibit because they really believe S-D is superior for audio purposes (after all it boasts better numbers across the board), and its also a lot cheaper to make ( = presumably better margins).

Are you talking specifically about SC filter or the implementation of the opamp analogue stage by the audio manufacturer?
Only asking as in some EE degree lectures they teach SC providing such benefit as I mentioned and not just relating to jitter - not arguing about opamp and context is specifically only the hybrid multi-bit SD rather than 1-bit that have a lot more challenges.

Appreciate your being very general but this covers a heck of a lot of audio gear that you say has poor sound, such as all top end Esoteric, amongst others who use "multi-bit SD" DAC with SC filter.
My point being IMO it is not necessarily the multi-bit SD DAC chip with SC filter/architecture at fault in this context of slew rate and linearity (of course meaning error correction in the digital domain).
Cheers
Orb

 

[Ragnar]

Are you talking specifically about SC filter or the implementation of the opamp analogue stage by the audio manufacturer?
Only asking as in some EE degree lectures they teach SC providing such benefit as I mentioned - not arguing about opamp and context is specifically only the hybrid multi-bit SD rather than 1-bit that have a lot more challenges.
Orb

SC filters should relieve the opamp's slew requirement somewhat by providing an RC filter ahead of its input; the downside is that the kind of capacitors found inside a CMOS chip are not especially linear with voltage.

I think BB/TI moved away from multibit because they really believe S-D is superior for audio purposes (after all it boasts better numbers across the board), and its also a lot cheaper to make ( = presumably better margins).

Their shareholders believe many more dacs can be sold at <<$10. The co-linear dacs require much more care in processing as well as considerably more wafer area and then trimming. Since most manufacturers are up against stiff competition, and specs sell, cost and measured performance is a dominant factor.

You can be sure the semiconductor industry moves ahead with hardly any regard for our passion for sound, driven by much larger markets than those served by audiophile companies which build only hundreds of a product a year. Companies that once sold "audio" ladder dacs, Intersil, Harris, ADI went out of that business when SD dacs took hold. SD dacs now fit the bill for medical and instrumentation use. The major market today for ladder dacs is video and that requires only 8-10 bit performance which doesn't need wafer trimming. The same can be said for discreet semiconductors; where are the great JFETs of yor? Even bipolars, used in almost all low noise mic preamps are going entirely to smt which need smaller, noisier die sizes. Even the fabs used for the older parts are going obsolete, replaced by processes that give smaller feature size, better for digital but equating to more noise for analog.

 

[opus111]

Are you talking specifically about SC filter or the implementation of the opamp analogue stage by the audio manufacturer?

I guess if a SC stage can be effected in a suitable technology it could sound decent, but I've only heard of it being done in CMOS with CMOS analog stages (opamps).

Only asking as in some EE degree lectures they teach SC providing such benefit as I mentioned and not just relating to jitter - not arguing about opamp and context is specifically only the hybrid multi-bit SD rather than 1-bit that have a lot more challenges.

I must have missed that part - what benefits do SC filters bring to the party, apart from jitter immunity? Sure they're able to be varied with the clock rate, so that's a plus for the manufacturer in being able to implement multiple SRs nice and simply. SC filters exist as stand alone chips (Linear Tech make one or two nice ones) - I can't countenance using one of those for audio as the noise performance doesn't pass muster, but they're pretty good for less demanding applications like instrumentation.

Appreciate your being very general but this covers a heck of a lot of audio gear that you say has poor sound, such as all top end Esoteric, amongst others who use "multi-bit SD" DAC with SC filter.

As I've not heard most of the 'high end' gear that uses the CMOS opamps, its only my conjecture that they'll sound poor. Perhaps they have some secret sauce to fix up the problems? But I do chuckle when I see examples with multiple paralleled CMOS opamp stages, as if their problems get fixed up by just stacking them

My point being IMO it is not necessarily the multi-bit SD DAC chip with SC filter/architecture at fault in this context of slew rate and linearity (of course meaning error correction in the digital domain).

As far as I can see it is S-D DACs with CMOS opamps which suck for SQ. But my listening to those is limited to CS4398 which Thorsten seems to think has quite a decent analog stage - I gave up trying to tweak it it sounded so rubbish against my cheapo multibit. S-D converters with external opamps suck less and the opamp chosen makes a difference to their sound. Faster slew rate opamps sound better, in general.

 

[Orb]

Opus will send a PM with the link for one of the lectures as it is not intended for general viewing (I think).
Cheers
Orb

 

[Lee]

I've not read the entire thread but I can offer up one observation based on recording classical quartets and trios. We have split mic cable into both DSD and PCM boxes. The DSD more closely captures the live event when played back after sound checks.

 

[Orb]

You can be sure the semiconductor industry moves ahead with hardly any regard for our passion for sound, driven by much larger markets than those served by audiophile companies which build only hundreds of a product a year. Companies that once sold "audio" ladder dacs, Intersil, Harris, ADI went out of that business when SD dacs took hold. SD dacs now fit the bill for medical and instrumentation use. The major market today for ladder dacs is video and that requires only 8-10 bit performance which doesn't need wafer trimming. The same can be said for discreet semiconductors; where are the great JFETs of yor? Even bipolars, used in almost all low noise mic preamps are going entirely to smt which need smaller, noisier die sizes. Even the fabs used for the older parts are going obsolete, replaced by processes that give smaller feature size, better for digital but equating to more noise for analog.

That is a good point, as audio market is incredibly small and niche in the scheme compared to global industry requirements for DACs.
Raises some food for thought and where we will be in 5-10 years regarding digital playback of music.
Cheers
Orb

 

[jap]

I've not read the entire thread but I can offer up one observation based on recording classical quartets and trios. We have split mic cable into both DSD and PCM boxes. The DSD more closely captures the live event when played back after sound checks.

What ADC/DAC boxes did you use?

 

[mep]

It appears on this thread that anytime anyone says DSD sounds better than PCM, it's always the fault of the person's PCM DAC-it's never the *right* one.

 

[MarinJim]

I've not read the entire thread but I can offer up one observation based on recording classical quartets and trios. We have split mic cable into both DSD and PCM boxes. The DSD more closely captures the live event when played back after sound checks.

Some would disagree. That what makes the world go around.

 

[Lee]

What ADC/DAC boxes did you use?

Korg MR1000 for DSD and Sound Devices 722 and 744 for PCM. AKG 414 BULS mics.

 

[Lee]

It appears on this thread that anytime anyone says DSD sounds better than PCM, it's always the fault of the person's PCM DAC-it's never the *right* one.

I have noticed this as well.

I learned recording from Chesky Records in a fully PCM environment (Katz and Kipnis were among the first hirez PCM pioneers). For my recordings (soon to be on HDTracks), I used 24/176 PCM. But when we started hearing Todd Garfinkle's recordings we secured a Korg and did some tests. DSD clearly sounded closer to the mic feeds on our classical work.

In any event, you can get great results from either format (and conversions from one to the other are pretty easy).

I just prefer DSD.

What I do know from experience is that 16/44 is entirely inadequate for proper capture of an audio event.

P.S. We use a Benchmark AD1 for PCM too so the comparison is not limited to the ADCs in the 722. The AD1 is superb; the 722 less so.

 

[MarinJim]

I prefer PCM because it sounds right in my system. But really, it is not PCM or DSD, it is the implementation of the format, IMHO.

 

[Lee]

I prefer PCM because it sounds right in my system. But really, it is not PCM or DSD, it is the implementation of the format, IMHO.

I think it is both the format and the implementation. Prolly more differences come from the latter though.

 

[opus111]

I've not read the entire thread but I can offer up one observation based on recording classical quartets and trios. We have split mic cable into both DSD and PCM boxes.

Which particular PCM boxes? (Ah I now see you've answered in a later post, apologies. I'll do a tech check on the 722 and 744 to see what they're using.)

Chesky don't use a fully PCM ADC as far as I'm aware, its Bob Adams' S-D design.

 

[Bruce B]

Which particular PCM boxes? (Ah I now see you've answered in a later post, apologies. I'll do a tech check on the 722 and 744 to see what they're using.)

Chesky don't use a fully PCM ADC as far as I'm aware, its Bob Adams' S-D design.

David is using a MSB Tech Platinum Studio ADC now

 

[opus111]

As its from MSB, is that a multibit one?

 

[Lee]

Which particular PCM boxes? (Ah I now see you've answered in a later post, apologies. I'll do a tech check on the 722 and 744 to see what they're using.)

Chesky don't use a fully PCM ADC as far as I'm aware, its Bob Adams' S-D design.

You misunderstand...my new record label uses the gear I mentioned, not Chesky Records. I learned the recording business while working with David Chesky and team in the 90s. I wish we had $30K+ for such a reference DAC like they use.

 

[opus111]

You misunderstand...my new record label uses the gear I mentioned, not Chesky Records.

Yeah I got that part - but you said that Chesky was 'a fully PCM environment' and I'm saying that they're not because the quantizer in the ADC they were using is low bit (I recall 15 or 16 levels, so 4 bits). Hence they're not a pure PCM enviroment because the ADC is S-D.

 

[Lee]

Yeah I got that part - but you said that Chesky was 'a fully PCM environment' and I'm saying that they're not because the quantizer in the ADC they were using is low bit (I recall 15 or 16 levels, so 4 bits). Hence they're not a pure PCM enviroment because the ADC is S-D.

I was referring to Chesky being PCM format only back in the 90s when I was involved - all pre-DSD.

 

[opus111]

There's much more to being PCM format than merely not being DSD. Most people listen to PCM nowadays on non-PCM DACs and which has been recorded with non-PCM ADCs. For example, your Sound Devices 722 (having checked it out) doesn't use a pure PCM ADC so when you say that its not sounding as good as DSD, you're not actually listening to a PCM system.