ANOTHER Example of Why I HATE DSD and Why Customers Who Bought Sony's Boloney Are So Annoying

[Hiro]

DSD is not a domain ...

 

In other words, DSD is a marketing name, SDM is the domain.

[SoNic67]

Actually, SDM is not a domain either, is a method.

Correct is pulse-density or PCM 1 bit.

Pulse-density modulation - Wikipedia, the free encyclopedia

[Miska]

Do you think the volume control in dcs DAC is also done in SDM domain?

 

I don't know, maybe for DSD sources. For PCM sources it could be done in PCM or inside the modulator.

 

ESS and Cirrus Logic DAC chips have digital volume for DSD. I think AK4399 also has, but AKM doesn't tell anything how they do it.

[Miska]

Actually, SDM is not a domain either, is a method.

Correct is pulse-density or PCM 1 bit.

 

Where do you get that PCM there? It's not PCM in any way. Multi-bit SDM is not PCM either.

 

- If you feed 8-bit SDM to an 8-bit PCM DAC output is garbage.

- If you feed 8-bit PCM to an 8-bit SDM DAC output is garbage.

 

Check also the linked document:

1-bit A/D and D/A Converters

[tailspn]

In the professional studio world there're basically two competing methods for level changing DSD content. One is the PCM route performed by Pyramix which uses the Philips developed DXD 352.8KHz PCM format. Here the DSD stream is converted to 352.8KHz 32-bit PCM, and all operations preformed in that form. It's then converted back to DSD for release. Other like DAW's work the same way, except not at 352.8KHz. Steinberg, Pro tools etc.

 

The second method employed by Sonoma and SADiE use the Sony developed e-chip, which fundamentally converts the 1-bit, two level (DSD) stream into 8-bit, two level words at the same sample rate, and operates in that format. It's then converted back to 1-bit, two level (DSD) for release, but since it's at the same sample rate, no decimation filtering is involved.

[SoNic67]

At the end of the day, DSD is PCM... The modality to obtain them is different, but resulting signal is 1 bit PCM. You can have PCM that is RZ encoded, true?

[wisnon]

I dont buy that sonic. PCM has to be decoded. DSD carries music directly and will play (badly) if connected to a transducer and well if LPF activated.

 

PCM is CODED. DSD is NOT, it is pulse width modulation (same as switching tech).

[Miska]

At the end of the day, DSD is PCM... The modality to obtain them is different, but resulting signal is 1 bit PCM. You can have PCM that is RZ encoded, true?

 

No it is not, in common multi-bit SDM formats bits have different meaning than bits have in PCM. Like I've said before, if for you any series of data (including this text encoded in UTF-8) means PCM I cannot help it.

 

You can transform any wordlength PCM into NRZ coding, but it won't play back on a DSD DAC.

 

In PCM you can have power of two number of levels; 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048...

In SDM you can have any number of levels; 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37...

 

P.S. Actually UTF-8 is closer to SDM than PCM...

[Adyc]

In the professional studio world there're basically two competing methods for level changing DSD content. One is the PCM route performed by Pyramix which uses the Philips developed DXD 352.8KHz PCM format. Here the DSD stream is converted to 352.8KHz 32-bit PCM, and all operations preformed in that form. It's then converted back to DSD for release. Other like DAW's work the same way, except not at 352.8KHz. Steinberg, Pro tools etc.

 

The second method employed by Sonoma and SADiE use the Sony developed e-chip, which fundamentally converts the 1-bit, two level (DSD) stream into 8-bit, two level words at the same sample rate, and operates in that format. It's then converted back to 1-bit, two level (DSD) for release, but since it's at the same sample rate, no decimation filtering is involved.

 

Thanks. So the second method is considered purer? I.e., it will not introduce noises like the first method?

[Superdad]

The second method employed by Sonoma and SADiE use the Sony developed e-chip, which fundamentally converts the 1-bit, two level (DSD) stream into 8-bit, two level words at the same sample rate, and operates in that format. It's then converted back to 1-bit, two level (DSD) for release, but since it's at the same sample rate, no decimation filtering is involved.

 

Ah, if only the Sony e-chips were still being made--and made available to independent h/w developers. Then we could have near-flawless (would it be?) level attenuation in the digital domain. Can you please tell me how wide a range of attenuation is offered with this method without loss of resolution? Can it cover a 60dB range?

 

Is anyone aware of any developers replicating (with an FPGA I would expect) the Sony method for DSD stream level control? Does Sony have patents on the whole concept or just their implementation for studio workstations as detailed here? http://www.jamminpower.com/PDF/DSD%20Editing%20System.pdf

 

In the same vein, Miska, can you tell us more about the method, range, and resolution of the s/w level control you offer for SDM in HQPlayer? I am really interested.

[Miska]

In the same vein, Miska, can you tell us more about the method, range, and resolution of the s/w level control you offer for SDM in HQPlayer? I am really interested.

 

I'm much more happy to talk about "what" than "how"...

 

Range is practically unlimited, so the noise floor is limiting factor, not resolution.

 

And I'm all the time improving/adding algorithms when I figure out new ways of doing things, so things also improve over time, release by release.

[tailspn]

Ah, if only the Sony e-chips were still being made--

 

Great find on that paper! A varient of the E-Chip is in every Sony Viao laptop, supporting the DSD Disk creation feature. DSD Disks play on Sony PS3 players, and the Sony 5400. Andreas Koch, who contributed to their development says the algorithms are certainly mountable in a FPGA, and improvable.

 

The range of level adjustment on the Sonoma is +6db to -60dB. I use it all the time to re-balance multi-channel mixes.

[tailspn]

Thanks. So the second method is considered purer? I.e., it will not introduce noises like the first method?

 

Well, it does have the advantage of not requiring decimation filtering for the down conversion from the 2.84MHz sampling rate of 64fs DSD to whatever sampling rate of the chosen PCM format. The effect of that filtering is the phase non linearity that introduces pre and post ringing, as well as non linear transfer characteristics. It's not night and day, especially if the PCM rate is as high as practical, but it is noticeable IMO.

[alainbil]

Just to be accurate...

The Naim DAC (as opposed to the DAC-V1), the NDS streamer and CDS and 555 players (Naim's top end products) all use the TI 1704 r2r DAC. The other devices use various TI delta-sigma DACs.

 

Eloise

 

Dont both CD5XS and CDX2 CD players use the Burr-Brown 1704 also ?

[Miska]

Dont both CD5XS and CDX2 CD players use the Burr-Brown 1704 also ?

 

Yes, at least CD5XS seems to be. I've been mostly checking DAC-V1 (PCM1791A), ND5XS (PCM1791A), NDX (PCM1791A), UnitiLite (PCM1793), NaimUniti2 (PCM1793) and UnitiQute2 (PCM1793?).

[Charles Hansen]

Why not design an all in one Ayre with UPnP support of stereo and multichannel PCM/DSD and give it a superb analog control.

 

That would be good, but would also be what is known as a "sub-optimal" solution.

 

If you are trying to simplify your system and just want to have one volume control, it is bad engineering to put it in the source component. Then every source component would need its own volume control. Also, the analog volume control would be placed inside the box with all of the digital components, with the high-speed clocks and lots of RFI in the same box.

 

A far better solution is to move the volume control into the same box as the power amp. Then the volume of ANY source can be controlled by just one volume control, and that volume control is inside a box that is completely free of any digital (RFI) signals. Go check out our AX-5 integrated amplifier. it is the most pure signal path ever made. There isn't even a volume control. It is just a power amp with multiple inputs. And the gain of the power amp is adjustable t o match any required playback level. Devilishly ingenious!

[SoNic67]

In PCM you can have power of two number of levels; 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048...

In SDM you can have any number of levels; 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37...

P.S. Actually UTF-8 is closer to SDM than PCM...

Well, 1 bit PCM has only 2 levels. Using a RZ encoding, you can add pulses one next to another and voala... you have the DSD. With all that 'nice' DC component.

Dont both CD5XS and CDX2 CD players use the Burr-Brown 1704 also ?

For $2000 they are way "out" there. There are other players with PCM1704 that can be snatched for $300...

[Charles Hansen]

That's exactly what I suggested earlier. But you would have to use a different DAC, something that is actually capable of playing back DSD-256 recordings made with the QA-9. Merging HORUS, and exaSound e20 come to mind...

 

No, that would be wrong. The QA-9 can create DSD-256 recordings, but then to have a fair comparison with PCM, you would hav to use something like AT LEAST DXD (352/32) or higher (768/24).

 

 

To keep things fair it is best to keep the overall bit rate close to the same. I think that (not perfect, but fairly close) would be 176/24 vers DSD-64.

[Charles Hansen]

This is wrong, you don't have to convert to PCM. You can keep it in SDM domain all the way. If you don't know how to do it, doesn't mean it doesn't exist.

 

If you double the SDM rate while doing it, the output can easily have 10 - 20 dB lower noise floor than the source material.

 

Same amount of noise is added in digital in both PCM and SDM cases, if you use same input and output resolutions.

 

Miska,

 

The question is not what is the world's best system the MISKA THE GREAT can dream up.

 

The question is of the two widely available digital high-resolution formats 192/24 and DSD (64), which sounds better. That's all.

[Hiro]

No, that would be wrong. The QA-9 can create DSD-256 recordings, but then to have a fair comparison with PCM, you would hav to use something like AT LEAST DXD (352/32) or higher (768/24).

 

Didn't you claim that nothing greater than 4x PCM was needed?

[tailspn]

The question is of the two widely available digital high-resolution formats 192/24 and DSD (64), which sounds better. That's all.

 

That's easy, you (me) can easily hear the difference between a 64fs natively recorded DSD file, and that same file then converted to 352.8/32 (DXD) PCM. Since all the available 192/24 content (with the exception of the no longer manufactured 20 bit(on a good day maybe) Pacific Microsonics Model One and Two) use A/D converters front ended with a Delta-Sigma Modulator converter (including the ARDA 1201)), and go through the same detectable decimation filtering and format conversion process as the Pyramix DSD > DXD, 192/24 PCM is no contest.

[Charles Hansen]

That's easy, you (me) can easily hear the difference between a 64fs natively recorded DSD file, and that same file then converted to 352.8/32 (DXD) PCM. Since all the available 192/24 content (with the exception of the no longer manufactured 20 bit(on a good day maybe) Pacific Microsonics Model One and Two use A/D converters front ended with a Delta-Sigma converter (including the ARDA 1201)) go through the same detectable decimation filtering and format conversion process at the Pyramix DSD > DXD, 192/24 PCM is no contest.

 

Hello Spinner,

 

You are going too fast for my drug-addled brain. Please help me one step at a time:

 

1) You say that DSD-64 file converted to DXD causes degradation, therefore DXD is less good than DSD, correct?

 

I say that I am not talking about ANY conversion being acceptable. Every laptop notebook has a 192/24 A/D converter in it, yet i don't expect that it will produce perfectly transparent conversion.

 

I am speaking specifically of the Ayre QA-9 as the machine used for the 192/24 conversion, and that also the digital filter must be set to "Listen" position.

 

2) I am not quite following you here. Again, you pick a specific machine (Pyramix) and say that it introduced audible artifact, therefore all Delta-Sigma ADC outputs converted to PCM must also introduce audible artifacts?

 

If that is what you are saying, I disagree strongly.

 

I'm sure I am misinterpreting your post, so please feel free to clarify.

 

Thanks.

[Charles Hansen]

Didn't you claim that nothing greater than 4x PCM was needed?

 

Yes, to compete against DSD. In fact it would probably compete well against DSD-128. But Misa is constantly talking about DSD-1024. I don't know how such a comparison would end. It may be we have reached the limit of today's electronics. Or it may be that we have reached the limit of the human ear/brain's resolution -- I really don't know. So it could be that PROPERLY IMPLEMENTED quad rate PCM would be sufficient.

 

But to change the rules immediately before the race begins is not fair at all.

 

For over a decade we have had TWO STANDARD HIGH-RES AUDIO FORMATS:

 

192/24

DSD

 

One group thinks one sound better.

 

Another group thinks the other sounds better.

 

I think that which sounds better depends on the implementation.

 

BUT I think that the BEST implementation of 192/24 sounds better than the BEST implementation of DSD.

 

I have developed hardware to test this, and it is the most fair test that I can think of. I have been developing testing procedures for over 40 years and am confident in the ability of my testing procedure. As soon as get time to do this, I will do it and report back.

 

Previously we could not do it because there was no DoP standard for DSD A/D conversion. Now we have preliminary spec that will allow us this possibility. Before this our SACD output was SDIF-2 or SDIF-3 (selectable), but the only equipment we had that would accept this was a borrowed Meitner DAC. It would be silly to try to compare while using two completely different DACs!!!!

 

Now that we have prototype DoP for A/D we can record source into both PCM and DSD using EVERY SINGLE ITEM IN THE ENTIRE SIGNAL PATH EXACTLY THE SAME (EXCEPT FOR THE FIRMWARE OF THE XILINX FPGA). THERE iS NO POSSIBLE WAY TO MAKE THE TEST MORE FAIR THAN I CAN THINK OF. AS I HAVE NOTED IN A PREVIOUS POST, THE DAC CHIP IS AN ES9016, SO THERE COULD BE A SLIGHT ADVANTAGE TO THE DSD SIGNALS.

 

My prediction is that the two signal chains will be EXTREMELY CLOSE sonically, but we will see. The DSD will have much more out-of-band noise, but our circuitry has no problems in this area, so I don't expect much difference here. NORMAL PCM (EVEN QUAD RATE) uses brickwall filters that damage the sound, but our system does not. This adds no expense -- in fact it make the system LESS expensive! But since there will be no aliasing into the audio band with signals of less that 170 kHz, I don't think this will be a problem. No musical instruments or microphones have output at these frequencies so it is a non-issue.

 

BUT if the DSD somehow magically sounds better, I will let you all know that I was wrong and eventually change all of my products to be DSD based. If they sound the same or if PCM sounds better, then we can stop all of this DSD nonsense, (WHICH, BY THE WAY IS THE ATTITUDE THAT 99.999% OF ALL RECORDING STUDIOS HAVE!), and just enjoy well recorded PCM -- SIMPLY ELIMINATE THE BRICKWALL FILTERS.)

[SoNic67]

I really doubt that any human can hear differences between DXD352.8/32 and DSD128. I am the first one to fight against people that say "that difference is cannot be heard" when is about 16 versus 20 bit or 44.1K versus 96K. But the DXD352/32 and DSD128 are not humanly detectable.

 

They are not even capable technically to be fully resolved. The best DAC's today have only a 20 bit resolution. Best ADC - 19 bit. Even in DSD, you cannot have better analog resolution, using the same stages. If any, you will have worse THD in bypass mode, because you don't use the multilevel D-S. Using higher DSD modes will NOT bypass the limitations imposed by the analog stages inside DAC's (or ADC's). is just wasted bandwidth.

 

Another reason why DSD sucks is lack of mixing ability. You need to rely on ONE microphone to capture the signal. that gives you at th ebest 100dB range - that's like 17 bit actual resolution. Even DSD cannot bring out what is buried under mic's noise and THD.

In PCM recordings, you have multiple mics, each with 100dB resolution. Added together, the final dynamic range will increase over that 100dB, to the maximum allowed by the actual bit-depth...

[Hiro]

But to change the rules immediately before the race begins is not fair at all.

 

For over a decade we have had TWO STANDARD HIGH-RES AUDIO FORMATS:

 

192/24

DSD

 

But we today also have DSD-128 and DSD-256, so it wouldn't be prudent to not use the higher DSD rates in the test. The fact that both recordings (DSD-256 and 4x PCM) would be made on the same ADC (QA-9) would make for a fair comparison.

 

You could post both recordings online so that everybody could listen to them via their DSD256/4xPCM-ready DACs (e.g. exaSound e20, Merging Horus).