PCM and DSD issue

 

This PCM and DSD issue has too much in it, and no product currently available is ‘perfect’ as it was always a premise that there is no ‘infinite’ thing around here, and again, the DAC makes PCM thinner, while DSD has other problems that prevent it from achieving its potential. PCM is still better if upsampled, and on PCs, for low budget listeners, PCM is pretty good – DSD makes those details emerge, but who needs those details? We get it from upsampled PCM too – but in a more favourable way. All that remains is SBMD.

pcm converted from dsd compared with plain pcm

So is it theoretically possible to convert the PCM output back into a DSD stream so that it matches the original DSD stream? In other words is a DSD->PCM->DSD conversion theoretically lossless?

But right now, I'm only concerned with the theory at the moment.  There are many people who claim that DSD is superior to PCM (they say it sounds better).  But on the vast majority of converter chips that give you a PCM signal, the conversion is a 1-bit delta-sigma stream that is converted to PCM.  DSD is just that 1-bit stream before the conversion.

In discussions I've had on this topic, I say that there isn't any difference between DSD and PCM if the conversion to PCM (in other words the delta-sigma stream conversion to pcm) is done with care.  What I don't know is if theoretically and mathematically, is the conversion actually lossless.  So is it theoretically possible (not just good enough for audio) to go from DSD to PCM and then back to DSD without losing any information?  If so, it seems that the DSD/PCM discussions are moot, and it all boils down to how good the actual implementation is.

And is there a better forum to post this in?  I just posted here due to the topic of DSD->PCM conversion.

DSD and PCM exist in various formats so you'd have to be a bit more specific. "DSD" is usually 64Fs DSD but there are also 128Fs converters. When DSD is converted to PCM for editing/mixing purposes 24bit 352.8 kHz seems to be a popular choice (Pyramix workstation e.g.). Still, some recording engineers claim (no proof) that there is an audible difference (example). Another option is 8 bit 2.8 MHz PCM (Sonoma and Sadie workstation). My guess is that when the conversion involves sample rate and/or wordlength reduction, it can't be considered lossless. Most DSD users I know don't think the quality "loss" is something to worry about.

I gladly leave the final answer to the mathematical experts on this forum.

I'm mildly curious how this method compares and contrasts to Sony's own SBM Direct conversion system for DSD, bit and sample rate levels aside--in terms of perceived sound quality. In my experience SBM Direct-encoded CDs are among the least fatiguing and irritating redbook cds that I have heard, especially with players that are prone to jitter. But if this system preserves more of the audio information of the original DSD recording then that benefit may be offset, assuming an appropriate D-to-A converter is used for playback of the converted recording.

with Sony DSD Direct software you can convert stereo PCM to DSF, but there's no program (I know of) to convert back to PCM.

Correct. The info.txt file says it all.

Input: raw mono DSD where the bytes' bits are processed from LSB to MSB. The file size must be a multiple of 4.

Output: raw 24bit PCM at 88.2 kHz in little endian format.

The decimation is done in two stages:

1. 8:1 (from 44100*64 to 44100*8 Hz)

2. 4:1 (from 44100*8 to 44100*2 Hz)

If you leave out the 2nd stage you basically have what's known as DXD (PCM at 352kHz sampling frequency) which seems to be popular as an intermediate/processing format.

If you test it and the result sounds too noisy chances are that the bit order of your file and the expected bit order doesn't match. If you know a little about hacking code it should be easy to change it. I already have a table in there that for reversing the bits in a byte, for example.

The implementation of the first stage is a little tricky. Instead of 8 MACs (multiply-and-accumulate) for 8 bits in a byte I just use one table lookup. The bitreversal is needed anyway to exploit the FIR filter's impulse response's symmetry and reduce the size of the tables. If someone wants to code something similar in C I can recommend this approach.

I am not an expert but it would be cool to compare to other DSD to PCM converters out there (if possible). There's a DSD plugin for WMP which concerts stereo DSDIFF/DSF to 16-bit 44.1kHz.

As I understand it there's many options you can have when converting DSD to PCM and several different approaches, some good, some not as good.

What would be a higher quality approach than 1. converting to 32 bit samples 2. low passing 3. downsampling?

Of course, the choice of a specific low pass filter implies the usual trade-offs (delay, ringing, aliasing, ...). But your actual choice for DSD2PCM probably wouldn't look different than the one you would pick for a regular PCM2PCM decimation, for example 96 kHz -> 48 kHz, to the same target sample rate. Or not?

The lowpass filter's roll off starts at something around 25 kHz (IIRC). That's IMHO high enough because there's already an awful lot of noise just above it. (The DSD propaganda tells you that DSD has a "great dynamic range". That only applies to frequencies below 20kHz which is not surprizing).

Interesting that if you perform the conversion badly (so that all the ultrasonic noise aliases back down into the audible band!) you hear noise that's anything-but-constant!

I know that exactly what you get is an artefact of whatever method you use to "wrongly" downsample - but even so it's interesting that this ultrasonic noise seems to vary so much - both in response to the music, and in some less predictable way.

If this was dither, it would be deemed to be very broken.

I know it's inaudible in normal playback, but if ultrasonic noise has any effect on replay equipment (assuming much of it reaches it - depends on the player) then having it vary is a bad thing - it's another potential cause of distortion which a supposedly "hi res" format can do without.

The noise is indeed a problem with the DSD format and, format-wise, something only improved by DSD128 (pushes it up twice as high) and furthermore with DXD (even less noise).

When converting to PCM, my opinion would be that one should have the option to be able to tweak such filter settings as they see fit. If ones objective would be to perform the most accurate "capture" (for lack of a better word) of what's from the original DSD signal (within the frequency range of your PCM output), you should be able to capture the whole thing, noise and all, if you wanted. Even if just for analysis and debugging.

The problem with the "most accurate capture" is that DSD players do 'know' that terrible amounts of noise are there and their analog low-pass probably reaches far enough down to compensate somewhat. PCM doesn't have this problem, a good 96 kHz DAC will roughly output at least 40 kHz of perfect analog bandwidth. This would be a capture of the full digital DSD signal, but probably not an accurate capture of the equivalent DSD low passed 'experience'.

Yes, solve the problem by wasting space for insanely high bitrates, to rescue this purely marketing motivated technology, instead of using proper PCM. DSD has not a single advantage over PCM other than DA converter costs. Good DSD DAC can be built pretty cheaply, kind of strange when you're targeting the higher end. But excellent PCM DACs are also available at commodity prices nowadays. Just accept it, the format is dead.

48kHz files do not preserve 0-22kHz band perfectly,  the brick wall effect means that the high pass filters have to cut into much lower frequencies for seamless playback.

Wow

Friday, September 23, 2016

More on DSD vs PCM

Here is the most helpful and informative page I have ever read about DSD, by Charles Hansen of Ayre Acoustics.  DSD ("Direct Stream Digital") is simply a meaningless trademark term which Sony has in this case defined as 1-bit delta sigma modulation at 2.8224Mhz and 7th order noise shaping.  They own the trademark so they can say it means anything they like.  Any time you deviate from 1-bit, as is essential for any kind of mixing or mastering or even level setting, you are forced out of the delta sigma domain into the PCM domain.  The Sonoma so-called DSD workstation is really a PCM workstation that happens to operate at 2.8224Mhz with 8 bit data.  DSD and PCM are interpreted by the same delta sigma DACs just with different digital filter algorithms.  The difference in filters explains everything people hear--it has to because there are no other differences.  Any superiority comes from the loss of the need for brick wall filters in high speed systems.  Now that we have 4x PCM, we don't need brick wall filters in PCM any more either, so we can achieve the same benefits with PCM which is far easier to work with, but few have ventured into this new landscape yet (except Ayre of course, who has a QA-9 A/D converter with no brick wall filter, instead it uses a "moving average" filter which has not time smear or ringing).  The only "pure" DSD recordings are all analog then converted to DSD, or live performances.  And there are just a very small number of those.  [And btw, Charles Hansen is the greatest!!!  After reading this hugely informative yet no-nonsense post, I'm a fan.]

Here's a long discussion at Steve Hoffman, which features fans on both sides, and reasonable civility.

Lotsa people think DSD and HighRes PCM are pretty equivalent.  I think that's a reasonable view, though one I don't exactly agree with (I still favor PCM), with the equivalence being DSD is about the same as 88.2/20 or 96/20.

Of course, DSD fanboys have always claimed that DSD is some special magic, that NO PCM could equal.  (Don't tell them about the feedback that delta sigma systems rely on.  That might collapse the magic.)

Many if not most think 44.kHz/16 bit "perfect sound forever" is still perfectly fine, and I know a number of people who think CD quality PCM is superior to DSD, especially in the highs, with a common thread that the highs in DSD sound fake, for which there is a tiny bit of technical justification (more noise and noise shaping is going on, meanwhile there is less brick-wall filtering, so you could also take this the other way).

Famously much SACD and DSD content is made from PCM sources, often defeating one of the long running claims (which probably most serious audio engineers would regard as hype) that DSD bypasses several phases of processing used in PCM.

Although unleashing DSD onto the world, Sony supported it poorly according to many industry insiders.  AFAIK, and in contrast many earlier format releases, Sony did not sell any DSD mastering equipment.  Instead, they gave it away to specific "partners."  If you were not one of the handful of chosen, you were out of luck, you would have to do your mastering in high rez PCM and convert to PCM.  Even the equipment Sony gave away may not have been fully featured, apparently Sony designed a fully featured DSD mixing system with a European partner, then never actually bothered to make it.  It's not impossible to make such a thing, and I believe that there are now, 18 years after launch, fully DSD mixing and EQ system(s) available from companies other than Sony.

Speaking of how DSD allegedly bypasses the decimation and integration phases (the hype which some believe as the magic of DSD that makes it inherently better), there are a bunch of problems with the argument (in addition to the one that PCM processing is nearly always used anyway).  Even if you had pure DSD mastering and playback (almost never the case) the claim would be inaccurate because:

1) First, it assumes that 1-bit DAC's are being used at the DSD sampling rate.  This is almost never the case anymore.  Almost all DAC's used for DSD now are Delta Sigma DAC's.  It's still considered DSD if you use Multi-bit Delta Sigma DAC's at the DSD frequency or higher, which requires a lot of complex mathematics to do  optimally.

The last 1-bit DAC's were used in devices like my 2001 DVP-9000ES.  Those were Sony DACs which actually operated at 70Mhz if I understand correctly, which would be something like 24x DSD.  Sony was doing some kind of way up sampling to increase dynamic range.  So it was never as simple as the cute block diagram Sony used to make DSD look simpler.

(Interestingly enough, it does not appear that the spec sheets for the Sony converter chips used in DVP-9000ES and back to CDP-707ES, have ever been made public.  But Sony did advertise these as 70 Mhz 1-bit converters.  I wonder if Sony made these at the long closed Sony Semiconductor factory in San Antonio, Texas.  Sony subsequently found it cheaper to buy off the shelf multibit sigma delta converters from the likes of Burr Brown.  Cynics

2) Second it assumes that 1-bit Sigma Delta ADC's are used.  I haven't found much discussion about this, but I believe that in the early days of digital audio, sigma delta ADC's were considered too noisy.  Noiseshaping is required when you use a sigma delta ADC.   Also, very high oversampling.  I believe some if not all of the earliest ADC's were actually SAR (successive approximation) which is one of most widely used approaches for analog to digital conversion.   Even now when Sigma Delta ADC's are used, they are used with multi bit converters and high oversampling.

Even if Delta Sigma ADC's are used, there's a lot more going on than you might think.  Quoting from the above linked article:

    These are usually very-high-order sigma-delta modulators (for example, 4th-order or higher), incorporating a multibit ADC and multibit feedback DAC. 

Sigma Delta systems are inherently approximate (aka noisy) systems which almost always require feedback to operate correctly.  This is something NEVER mentioned.  This is one reason why I've personally moved back to PCM as much as possible.  PCM does not require feedback to work correctly.  The dirty word "feedback" would destroy the claimed "magic" of simplicity.

Of course it is also because of feedback that delta sigma systems can get near perfect linearity without requiring extensive trimming the way PCM systems do.  You do the fine tuning before, when you can only guess, or you do the fine tuning after the fact, which can always be perfect.

Now this also is probably a non-issue.  While the feedback used in Delta Sigma would smear the highs, Delta Sigma ADC's and DAC's generally operate at such high frequencies that high frequency information might even be better preserved as compared with slower PCM systems.  It's actually quite hard to know without extensive analysis and/or testing which system preserves the high frequency integrity better.

However, one can also just look at the measured performance.  DSD does quite well compared to 16 bit systems in the midrange, but has much more noise in the upper octave 10-20kHz.  That greater high frequency noise means that by definition high frequency information is NOT being preserved as well.  OTOH, there is ultimately response to an even higher frequency, and there may be less phase shift in the upper audible octave.  So it looks like a toss up.

Listening Tests

The best published investigation of audible differences between PCM and DSD was done in Germany using some of the very best megabuck PCM and DSD equipment.  (IIRC the PCM was either 88.2kHz/20 or 96kHz/20, so as to have comparable bandwidth and bit depth.)  Monitoring was done with Stax headphones (you can't get more transparent than that).  And the result was: there is no audible difference!  Not only was the null hypothesis not rejected but most identification was no better than random for nearly all people.

I believe this is basically correct.  DSD is simply an inefficient high resolution system which takes more bits to achieve 88.2kHz/20bit fidelity than PCM does, and PCM is more easily worked with in many ways, including incrementally increasing fidelity with just a few more bits.  The very idea of 2xDSD and 8xDSD is monstrous--a monstrous waste of bits.

I've argued that DSD operates a bit as if it has an infinitely varying digital filter.  Varying the digital filters in 44.1/16 can make a slight audible difference (or larger if you throw out the book with NOS, which is not high fidelity IMO).  Once you get to modern apodizing reconstruction filters using ordinary PCM, it's not clear from published research that better can be achieved or is necessary, but an end-to-end apodizing system like MQA promises to be would be a step better.  That is, a step better than DSD-in-principle.

To DSD or not to DSD

If Only Sony had marketed DSD on a fairly straightforward technical basis, I might have signed on in 1999 and never looked back.  Forget the simplicity crapola, the real technical advantage of DSD compared to plain vanilla PCM is the superb impulse response.

[Update: after the nth revision of this post, I discovered that Charles Hansen had already debunked the above graph in great detail.  It's a pack of lies from beginning to end!  It's no wonder that Sony didn't plaster this on everything, more people might have called them out.  This is not to say that you couldn't come up with a relatively more honest graph to make the point that DSD has better impulse response that the usual 44.1kHz plus brick wall filtering usually used, but in that case there would be competing hirez PCM systems that could do as well.  The way the graph is shown no real systems can produce those results at all.  BTW, I'm now a little bit concerned that the MQA impulse response graph shown in TAS is also inaccurate, though in showing more rather than less time smearing with standard PCM.]

Now, PCM defenders will argue, and they have got to be at least mostly correct, that this difference, which is caused by high frequency phase response in the anti-aliasing and reconstruction filters, is not audible.  But it sure looks like it would be important.

I had hunted all around for a clear picture like the above, and found it posted by Hiro, a senior member of ComputerAudiophile, at this 64 page mega argument about DSD, which looks to be one of the better discussions on the topic.

Hiro starts by calling most of John Siau's arguments as wrong or misleading, then makes a pretty interesting (and misleading) argument himself.  He claims that Archimago measured the same noise from DSD64 as from 192/24 on the Teac interface.  So I had to go and read Archimago's blog.

Hiro is plain WRONG!

In this page of measurements, Archimago shows the astoundingly high noise level of plain DSD in a scope trace of a 1kHz sine wave.  Then, later on, in the 6th graphic on the page he shows the noise level of the Teac under several conditions, DSD with FIR1-4, and 24/192 with sharp filter.  These could hardly be more different above 20k.  The DSD with all 4 possible FIR filters rises from -145dB to about -75dB around 90kHz, reaching -100dB at 40kHz.  Meanwhile, 24/192 rises from -145dB at 20khz to -115dB at 90kHz, reaching only -140dB at 40kHz.  At 40kHz, the point where Hiro claimed that the noise levels were still the same, there is actually a 40dB difference.

Now, in the previous page of measurements, Archimago shows that 24/192 is noisier than 24/96 in the Teac and this is typical of all DAC's (and part of the reason why Dan Lavry and some others recommend 24/96 instead of 24/192 for the highest fidelity).  Perhaps it is not surprising that Hiro takes the worst case for PCM noise, 24/192 as his basis of comparison with DSD.  But even there he is wrong, as I just reported.

Even if you take the worst case for PCM noise, 24/192, and then combine it with No Oversampling (NOS) which as I always argue isn't really high fidelity or standard PCM, do you get the noise to rise a bit closer to DSD.  But the DSD noise is still higher.  Archimago doesn't show the NOS and DSD noise spectra on the same graph, or even the same page, but I can compare them and they are still quite different.  At 40kHz, 24/192 with NOS reaches -113dB.  Meanwhile, DSD64 has reached -100dB, which is 13dB worse.

Now Hiro was wrong in what he said.  But perhaps he merely misspoke.  Perhaps he meant to say that DSD128 is comparable to 24/192 in noise level.  And for that comparison, Archimago does show both on one graph.  At 45kHz, actually pretty much everywhere, the noise level for 24/192 is lower, but just barely until 50kHz and above.  At 40kHz it breaks down like this:

24/192/sharp   -139dB

DSD/FIR2      -137dB

DSD/FIR1      -131dB

I still wouldn't say "they are the same below 45kHz" but close.  But I'm not even sure why we are doing THIS comparison.  Well Hiro also mentions that DSD64 can very simply be up sampled to DSD128.  Now here we have an interesting case, however.  Upsampling will push the digital noise upwards.  But it seems to me very much unlike "noise shaping" in one critical way.  As a purely 1-way process, up sampling cannot possibly restore lost information.  The information loss from the original DSD64 encoding cannot be undone.  So while the noise will be reduced, the lost information cannot be restored, and I'd predict a kind of dark sound, the same thing you get in clunkier fashion with noise gating.

Meanwhile, I would have been (and was) rightfully turned off by a large number of things about DSD right from the start:

1) DSD recorders have been almost unobtainable (there were no consumer DSD recorders until 2007 or so, right now one is available for $999).

2) SACD discs are impossible for most people to make, they require a manufactured watermark (some old machines will accept a fake DVD/SACD, and the newest ones will read DSD files).

3) DSD does not lend itself to simple DSP for crossover and room correction functions--so conversion to and from PCM is required anyway, so the best approach is high rez PCM end-to-end.

I'm less bothered by (3) than I was years ago, for an interesting reason.  The reason is that conversion to and from PCM is extremely transparent.  It's so transparent that I find I often prefer taking the analog outputs of digital devices and resampling to digital at 24/96 than just letting the 44.1/16 pass through all the way.  So, if I'm fine with resampling analog to PCM, why not DSD to PCM, or even DSD to Analog to PCM?  I see now I can fit DSD into my system as a perfectly fine music delivery system, though not as a final digital conversion approach.

Of course, as many have pointed out, SACD was an attempt to impose DRM on an industry.  If Sony could have led everyone to abandon PCM, we'd be locked into their new system with a DRM system that has not even yet been broken.  Of course, we know in retrospect this was never going to happen.

But from the beginning, there was no consumer recording of the new formats, and, very curiously, the first generation SACD machines had problems dealing with user-recorded media that had already become well established by that time.  As if to send a message to the industry.  Well it was too late.

Now I just said that PCM conversion is very transparent, as was demonstrated by the Meyer/Moran experiments in 2006, up to 10 levels of PCM conversion/deconversion was still found to be audibly transparent.  Very little noise is added, however there is a increasing amount of high frequency phase shift.  This doesn't look good on photos but has never been proven to be audible.

Meanwhile, DSD is not amenable to multiple generations because of high frequency noise that keeps on growing until you get overloading in the highs.

However, DSD128 is looking like it might have reasonably low noise levels in the near supersonic, and still of course give you the natural (noise shaping aka feedback driven) impulse response.  DSD64 is so noisy you can easily see the HF noise on high bandwidth oscilloscope traces of sine waves, as Archimago shows.  DSD128 looks just like analog on the scope.

I'm not sure we've seen the end of this, since now filter designers are showing how perfect impulse response can be obtained with PCM and slightly higher sampling and end-to-end mathematical apodizing.  This retains the advantages of PCM in relative compactness and mathematical tractibility--it can easily be worked with in DSP.

DSD counfounds mathematics not because of sigma delta itself--that's the trivial part that had me fooled for the longest time.  Equally fundamental to DSD is Noise Shaping, based on continuous high level feedback.  This means, in effect, each pulse is NOT equal.  Each pulse is in the context of everything before and after it, which actually determines what it means.  This context dependence makes the mathematics infinite.  You can't just "add things up" to make a mixer, etc.

Meanwhile, PCM is reborn every few years with some interesting innovations, though I consider apodizing important but little else.  IMO, by the time we get to CD players like the Pioneer PD-75 around 1991 we're in the modern era of high sound quality, thanks to high linearity, low jitter and high stability, and flat but closer to linear phase digital filters: plain old 44.1/16 bits done fairly well is incredibly good!  For the longest time, the best published research in JAES was that it was perfectly transparent.  It may never have been perfectly transparent, but it's obviously quite close.  It has only barely been established in the AES literature that it isn't perfectly transparent, that significantly improved apodizing can be slightly audibly better and demonstrated in DBT (published by Meridian).  This has not been scientifically established for DSD, in fact the reverse has been demonstrated in the most recent and well done experiment--it is indistinguishable from comparable PCM.  Most talk to the contrary has not been well founded.

DSD stays alive simply by slowly making what used to be impossible less so.  And I'm happy to play with it as I can without huge expense.  I will never have full DSD end-to-end because that would require me to give up DSP.  But I can accept DSD inputs, converted through analog resampling to 96/24.

Which in a way, is not surprising.  Recall that DSD was originally invented, in the first place, not as a mixing or mastering format, but as an archival format.  Now I'm not sure it's especially good at that either, because of noise, but for an archival format there isn't much concern regarding mixing and mastering, and even distribution and playback.  Also, DSD56 was invented specifically for the mastering of 44.1 and 48kHz, the two popular rates of the time, but not for the high rez PCM formats of today. 

Now certainly someone as astute as Ted Smith would understand the practical and mathematical difficulties of DSD.  Nevertheless, he built a DSD DAC.  Maybe he has some answers to the other problems too.  I find his progression from first time electronic builder to advanced DAC builder unbelievable.  In this story line, it all happens in a few months in 2010, while he's apparently also listening to Johnny Cash.

Archimago does usefully propose combining DSD128 with lossless compression.  If it can indeed be compressed to the same size as 192/24, perhaps it's not that bad.  But we have no reason to believe this complexity is needed.  As far as we know now, 24/96 PCM is as high definition as is needed, and it is far easier to work with than DSD128.

Here's a comment by Charles Hanson saying flat out that DSD is unnecessary (because he has incorporated the goodness of DSD into his quad rate PCM with the QA-9).  He says his quad rate PCM sounds better than DSD at any rate.  I believe him.  Hanson gives the background story of DSD vs DVD-Audio in a subsequent comment.

Posted by Audio Investigator at 4:26 PM

http://audioinvestigations.blogspot.com/2016/09/more-on-dsd-vs-pcm.html

this is great stuff – too much, so make sure to read more!

Interview with Ted Smith from PS audio

 

4.1 Before doing the D-A conversion on DirectStream DAC, no matter PCM or DSD signal, they are both being converted to the DSD first. Are you suggesting that DSD is superior to PCM?

 

Before discussing it, I must give some guide of the basic DSD concept so that normal people can understand it better.

 

In terms of the format, DSD is equally good as PCM. They both can record a huge amount of music information. The difference is just the format. The reason I insist on DSD development is not because PCM this format is not good enough, but rather the way of PCM D-A conversion is not ideal enough. Simply put: multi-bit PCM dac’s structure is too complex and it can only rely on digital filer to “predict” the original analogue waveform, which can somehow mask the original “appearance” of the digital signal. In other words, PCM dac cannot exploit the advantage of PCM format’s information to the full.

 

But for the DSD DAC, it is such an ideal D-A conversion approach for me to be much simpler, more linear and closer to the analogue. In theory, DSD dac only needs one resistor and one capacitor to form a simple low-pass filter circuit, which converts DSD signal to analogue signal. This kind of conversion not only has much lower error, but also doesn’t need strict component matching like PCM dac. However, please note that these advantages I mentioned above can be only achieved by 1 bit DSD D-A conversion. On that premise, I personally think 1 bit DSD dac is much better than multi-bit PCM dac even for processing the PCM signal. If PCM signal is converted to the DSD first, 1 bit DSD dac will be capable of exploiting all hidden information in that PCM-converted DSD signal.

 

In conclusion, no matter PCM format or DSD format, 1 bit DSD dac is the most ideal approach for D-A conversion.

Edited April 15, 2017 by louisxiawei

Each "tap" is nothing but a pair of filter coefficients.  Number of taps = number of coefficient pairs.  The greater the number of taps, the "longer" the filter is.

 

So "tap-length" is a misnomer for filter length--which is just number of taps.  Hope that is more clear.  

Almost,

an FIR filter is a sequence of delay lines, the connection between each delay element is "tapped off", run through a multiplier (the "tap coefficient") and and the result of these multiplies are all added together. The value before the first delay element is also tapped.

The filter length is the number of delay elements, the number of taps is the filter length plus 1.

 

In digital FIRs the delay elements are flip flops, arranged as registers, for audio data something like 32 bits longs. Each delay element is one of these registers. The output of a register feeds into the next, all clocked by the same clock. The hard part is the multiplier. Each tap takes its value and multiplies its coefficient. (usually considered to be between zero and one), this multiply takes significant hardware to implement.

 

Thus for a 100 length filter you have 100 32 bit registers, 101 taps, each of which is a 32x32 bit multiplier, and one 101 input adder, it adds all the 101 numbers from the multipliers together. The output of the adder is the "output" of the filter for that particular time slice.

 

The value of those 101 coefficients determines what the filter does.

 

Even a small number of taps can do significant filtering. The more taps you have the more control you have over exactly what the filter does. A large number of taps does not necessarily produce a "better" filter, it gives the designer more control to implement exactly the filter they want.   

Yes, the number of taps is not useful as measure of filter quality, it is pretty much arbitrary number.

 

It is not a problem run a million tap filters on computer. For example room correction filters are typically 65000 to 256000 taps and processing that takes negligible amount of CPU time at PCM rates, including 768k... When you have something like 16 million taps at 24 MHz sampling rate, then we start talking about noticeable amount of CPU time.

Interview with Ted Smith from PS audio

4.3 But many people think DSD dac has high-frequency problems, what's your opinion?

 

The high-frequency problem does exist, therefore DSD dac needs Noise shaping to solve it. By taking advantage of the Noise Shaping, the very high-frequency noise can be eliminated and keep the intact information of the music, which our ears can hear.

 

To solve this problem completely, DirectStream DAC does a 10X upsampling for all DSD signal beofore D-A conversion so as to push the noise to the very high-frequency area that human ears cannot detect, then covert back to 2x DSD. I think this can completely solve the so-called “inborn flaw” of the DSD dac.

Interview with Ted Smith from PS audio

4.6 Don't you think the conversion between DSD and PCM will cause any distortion?

 

The conversion between DSD and PCM is used by FPGA I developed for nearly 10 years, which can make sure that the distortion from the conversion to the lowest. Many would think it might be better with less conversion, but the merit of this conversion is greater than its disadvantage. As I mentioned, in terms of format, DSD is equally good as PCM, the key is to use 1 bit DSD DAC to do the D-A conversion.

 

 

I personally am a little disappointed regarding Ted's 4.6's answer. It means nothing to me. I hope someone can give us more detailed answers about the potential lossy signal/distortion from the DSD-PCM conversion.  

Interview with Andreas Koch from playback

2.1 Your playback DAC convert files no matter PCM or DSD into DSD first and then do the D-A conversion afterwards, do you think it is a better way to do this even for the PCM signal?

 

The multi-bit PCM DAC has its non-linear distortion problem because the proportion of each bit’s output is different. To decrease this non-linear distortion, the dac has to reply on high-precision components and circuit, and consequently, leading to high manufacturing cost.

 

Moreover, most of multi-bit PCM dacs use Brickwall filters that can produce pre-ringing. Generally speaking, this pre-ringing is the sound that most people loathe and being described as “ annoying digital sounding” quite often. By comparison, DSD dac is much more linear and do not need digital filter, which can solve many PCM DACs’ flaw.

Interview with Andreas K

2.2 Many people think only 1 bit DSD D-A conversion is the genuine and correct way to work, what do you think?

 

Yes, that’s true. The premise of DSD dacs being better than PCM ones is that DSD dacs must be a true 1-bit DSD modulation. However, the problem is that most of Delta sigma dac chips on the market nowadays will make the DSD signal get through low-pass-filter circuit first to be converted to lower-sampling multi-bit signal before doing the D-A conversion. Under that circumstance, DSD signal in fact has been converted to PCM one already, which can also cause many PCM-related problems. As far as I know, most of DAC chips on the market are in the category of multi-bit Delta sigma dac modulation, which all convert DSD signal to 5-6 bit PCM signal.

 

Only by building separate circuit for DSD and PCM, we then can achieve 1 bit native DSD D-A conversion, and all Playback Designs dacs are made in this fashion.

Disagree.

 

It is the flaw of 1-bit SDM engine that caused DAC chip maker to produce 5-6bit SDM engine. DSD is not converted to PCM in these engines as there is no decimation fllter applied in the process.

 

Mr Andreas has been saying the SAME thing again and again in various interviews/workshops, It is a bit disappointing that a man with his level of expertise, so many of tasteless marketing talks and so little of technical tidbits have been revealed to the community.  

I too was shocked to read that mis-information from Mr. Koch.  At first I assumed that the translation was in error, but to see you report that he has made such a false claim elsewhere leaves me scratching my head.

 

Multi-level PDM (as used in S-D DAC chips and in shift-register discrete designs such as Miska's DSC-1 and others) is in no way shape or form similar to binary-twos-complement multibit PCM!  Not in rate, not in encoding, not in resolution, and not in where it puts noise.

 

Remember folks, DSD is just the marketing name for 1-bit (well really two-level) variant of PDM (pulse data modulation), and if done right, there are lots of good reasons for a designer to harmlessly move a 1-bit DSD/PDM stream into a multi-level PDM format.

 

[That said, the very different, and quite critical steps of taking Redbook and interpolating that first to high rate PCM and then through SDM modulators to produce a PDM stream--that's where the art comes in, ala HQ Player or some of the bettter DACs with the horsepower and carefully refined code/filters for the task.

By the way, if we had a home computer interface and cable protocol to send multi-level PDM (call it DSD-wide if you wish), then DACs could be built for that and Jussi would likely be happy to oblige multi-level output from his fine SDM engine.]

PBD doesn't, AFAIK, use DAC chips. It is all built around FPGA...

There is some strange misconception that multi-bit SDM DACs would be related to PCM and operate in similar fashion. It would be plain stupid to do it like that. PCM is wasting a lot of bits. If we think a sinewave that is -12 dB level, it means that top two bits are never used to represent it, that already loses a huge range of values, only quarter is used. In addition, many more of the bits are only used around peak regions of the waveform. Only good thing is that this reduces also biggest problem of R2R ladders because there the biggest errors come from MSBs. But there is very little resolution left at the lowest levels / around zero-crossing.

 

For example TI/BB chips use 5-level SDM which would be about 2.3 bits in PCM terms. Meaning that anything below -6 dBFS level would be 1-bit anyway and the MSB would practically always unused. Even for "6-bit" in PCM style, signal would become 1-bit below -36 dB. However, SDM always uses all bits/levels/elements, regardless if signal level. It is most precise at the lowest levels / around zero-crossing.

 

My DSC1 DSD-DAC is "5-bit" if you think in terms of two's complement -> log2(32) = 5, or more precisely log2(33) = 5.0444 bits. Although that kind of bit figure is completely useless in this context. The meaningful part is that it can produce 33 different output current levels which are turned into voltage levels by I/V stage.

The problem I have is that whatever people say negatively about DSD, it doesn't bear with what I hear in my setup.  

If you take the trouble to make the investment in the computing power needed to process DSD (and a DSD DAC), it is a very enjoyable sound.

https://www.computeraudiophile.com/forums/topic/31445-debate-of-dac-design-regarding-dsd-vs-pcm-among-5-vips/?page=2

Rob Watts on Chord Mojo tech

We talk to Rob Watts about the digital engineering behind Chord Electronics remarkable Mojo portable DAC. Rob explains just what ‘taps’ are and why they matter, and there’s a lot more to it than digital audio technology.

Jason Kennedy: You like to talk about taps in DAC chips, the more the RW: merrier seems to be the gist, but what are they and why does no one else mention them?

Rob Watts: OK, the interpolation filter’s job in a DAC is to re- create the missing parts of the original analogue signal - the signal in between one sample and the next. This is done with an FIR filter. In a simple way, a FIR filter consists of a data memory (this stores previous data samples) and a coefficient memory (this is a fixed memory with all of the coefficients that the filter algorithm has created). To create an intermediate data, you simply multiply and add all of the stored data samples with a particular coefficient, and once you have added all of the values you end up with the intermediate value you need. Now in the early days, you used a delay line to store the previous data samples, and you tapped into this delay line in order to access the stored data. Hence the word taps.

So why is it only me that goes on about taps and stuff? The problem is about understanding - when I was at university, I studied electronics. But I was passionate about audio, and was interested in the physiology of hearing. I thought if I understood that, I could make better audio electronics. One of the things I was very interested in was how the brain processed the output from the ears. Now we take our hearing for granted, but the brain does some amazing things to give us auditory perception - separating individual sounds into separate entities with placement data (where are sounds located) is an amazing feat, requiring considerable processing. And we know very little as to how the brain does this. Anyway, what I did learn was that transients were a very important perceptual cue, and that the timing of transients was crucial. From transients the ear brain locates sounds in space, it is also used to compute pitch (particularly for bass) and it’s used for getting the timbre of an instrument. I spent a lot of time researching this in the psychology library, which was close to my hall of residence. Anyway, one of the courses in electronics was Whittaker Shannon sampling theory. And this is the basis of digital audio. From this it is a fact that if you had an infinite tap length FIR filter you would perfectly reconstruct the original bandwidth limited analogue signal in the ADC. It would make no difference if it was sampled at 22uS or 22pS you would have the same digital signal. But it was very clear to me that having a limited tap length would create timing errors. And I know from my studies and from my own listening tests that that would be a major subjective problem.

rob-watts.jpg

Rob Watts at the Chord Mojo launch

Now, unfortunately, nobody else has recognised this problem for two reasons. One is electronic engineers do not study hearing, and the second problem is that they are stuck on the idea that filters are a frequency domain problem and not a time domain problem. So if you design a filter where your only concern is frequency, then a 100 taps or so is enough. But if you think from the timing perspective, it categorically is not enough. What I have done is to make no assumptions about whether something makes a difference to the sound unless I actually do a listening test. And listening to increasing tap lengths always improves the sound quality. With Dave, I am at 164,000 taps, and I know that that is not the end of it and that further improvements are possible.

Is an FPGA a DAC chip that the designer can customise rather than a fixed set of parameters as found in regular converter chips?

No an FPGA is not a DAC chip, it’s a sea of gates that you can connect together to make any digital device you like. You could make a PC processor out of an FPGA, or a device that controls a rover on Mars, or the digital parts of a DAC. I also create IP and designs for audio to make silicon chips. And my designs could be used to make a dedicated DAC chip, or it could be used to program an FPGA. The benefits you have using an FPGA are considerable, as you can have thousands of times more processing power than is found inside high end audio silicon chips. Indeed, Mojo has 500 times more processing power than conventional high performance DACs.

Is there a separate analogue output stage in Mojo or is that derived directly from the FPGA?

Absolutely. An FPGA is entirely digital and cannot have analogue outputs. This is done with discrete components (flip flops, resistors, capacitors and separate reference circuitry). Doing the analogue discretely has big benefits. In designing silicon DACs there are enormous problems - the clock has to be distributed, and this increases jitter, the substrate injects noise and distortion into the analogue parts, the reference circuitry can't have low enough impedance and is noisy, resistors are non-linear, capacitors are non-linear too. None of these problems apply with discrete DACs.

So does this mean anybody can design their own DAC's using FPGAs? No I am afraid not. Creating the internal modules, getting them right, getting the DAC technology right, has taken me 30 years to do. This is not easy to do.

Chord-mojo_and-Hugo.jpg

Hugo and Mojo, guess which has the most advanced tech inside.

I am getting rather different sound from coax and USB inputs with the latter sounding better, what’s the reason for this?

It’s complicated and depends upon a number of factors - principally the amount of RF noise injected into the Mojo, and the amount of correlated noise that gets in. It will depend upon the source device as to which sounds best. My preference is optical, as this has the smoothest sound quality and best depth, as it does not suffer from both of the aforementioned problems.

Does Mojo upsample and if so what does it upsample up to?

Yes all my DAC's up-sample to 2048 times that is at least 16 times more than typical. What does this do? Well its not just about up- sampling but filtering out the RF noise that is present on a digital signal. Its essential to do this, as it gets you closer to the original analogue signal in the ADC (and this is the DAC's job to recover the analogue signal not the digital data). This extensive filtering reduces jitter sensitivity by a factor of 64, and allows the DAC to eliminate noise floor modulation. Now this is a very important problem, as it makes the DAC sound hard and less smooth and is a major problem with DACs - all other DACs have very large noise floor modulation, Mojo has zero measurable noise floor modulation (I have plots at home proving this). This is a major reason why Mojo sounds so smooth and natural.

How low a headphone sensitivity can Mojo sensibly drive?

I wanted Mojo to work with the most sensitive IEMs available, and to do that I had to improve noise. That's why Mojo has a 125 dB dynamic range. To do this was not easy, but it means that you can drive any headphone on the market as Mojo has extremely low noise, but is able to deliver over 5v RMS and 0.5A RMS of current.

You say you tuned the Mojo to have a smoother sound, was this in order to make it more forgiving of real world formats like MP3?

Yes I wanted it smoother and warmer, not so much with AAC or MP3, but more because it is likely to be partnered with harder sounding headphones.

http://www.the-ear.net/how-to/rob-watts-chord-mojo-tech

To DSD or not to DSD?

That is all ?

 

You try to convince happy DSD listeners to doubt because the theories, algorithms , upconvertion, et al?

 

Now, you are not asking questions for your knowledge. Your goal is to attack DSD.

 

Do you think the vast majority of music listeners are interested in the theory behind DSD or PCM? The proof of the pudding is in the eating, but you don't need the recipe to be convinced about his taste.

 

Please go to help a lying politician to convince the people that their lies are true, hiding behind an "innocent" question ...!

DSD format specs are public and conversion tools exist, both commercial and free. I had interest on DSD topic only because I found my DAC sounds better in DSD mode. I found enough information on this forum to understand the DSD topic on the level I need it. If you are interested to know more about an area you don't understand, start to self study. That's better recommendation than yours ...

But what about the problem Rob Watts brings forth concerning DSD's lack of perception of depth? Thus musicality.

Sorry to print out his entire thread on this subject but I would like to hear others opinions on it's merits. Dave is the new flagship dac soon to be released by Chord but designed by Rob Watts.

 

By Rob Watts

 

~"There are actually two independent issues going on with DSD that limits the musicality - and they are interlinked problems.

 

The first issue is down to the resolving power of DSD. Now a DSD works by using a noise shaper, and a noise shaper is a feedback system. Indeed, you can think of an analogue amplifier as a first order noise shaper - so you have a subtraction input stage that compares the input to the output, followed by a gain stage that integrates the error. With a delta sigma noise shaper its exactly the same, but where the output stage is truncated to reduce the noise shaper output resolution so it can drive the OP - in the case of DSD its one bit, +1 or -1 op stage. But you use multiple gain stages connected together so you have n integrators - typically 5 for DSD. Now the number of integrators, together with the time constants will determine how much error correction you have within the system - and the time constants are primarily set by the over-sample rate of the noise shaper. Double the oversampling frequency and with a 5th order ideal system (i.e. one that does not employ resonators or other tricks to improve HF noise) it converges on a 30 dB improvement in distortion and noise.

 

So where does lack of resolution leave us? Well any signal that is below the noise floor of the noise shaper is completely lost - this is completely unlike PCM where an infinitely small signal is still encoded within the noise when using correct dithering. With DSD any signal below the noise shaper noise floor is lost for good. Now these small signals are essential for the cues that the brain uses to get the perception of sound stage depth - and depth perception is a major problem with audio - conventional high end audio is incapable of reproducing a sense of space in the same way one can perceive natural sounds. Now whilst optimising Hugo's noise shaper I noticed two things - once the noise shaper performance hit 200 dB performance (that is THD and noise being -200 dB in the audio bandwidth as measured using digital domain simulation) then it no longer got smoother. So in terms of warmth and smoothness, 200 dB is good enough. But this categorically did not apply to the perception of depth, where making further improvements improved the perception of how deep instruments were (assuming they are actually recorded with depth like a organ in a cathedral or off stage effects in Mahler 2 for example. Given the size of the FPGA and the 4e pulse array 2048FS DAC, I got the best depth I could obtain.

 

But with Dave, no such restriction on FPGA size applied, and I had a 20e pulse array DAC which innately has more resolution and allows smaller time constants for the integrator (so better performance). So I optimised it again, and kept on increasing the performance of the noise shaper - and the perception of depth kept on improving. After 3 months of optimising and redesigning the noise shaper I got to 360 dB performance - an extraordinary level, completely way beyond the performance of ordinary noise shapers. But what was curious was how easy it was to hear a 330 dB noise shaper against a 360 dB one - but only in terms of depth perception. My intellectual puzzle is whether this level of small signal accuracy is really needed, or whether these numbers are acting as a proxy for something else going on, perhaps within the analogue parts of the DAC - I am not sure on this point, something I will be researching. But for sure I have got the optimal performance from the noise shaper employed in Dave, and every DAC I have ever listened too shows similar behaviour.

 

The point I am making over this is that DSD noise shapers for DSD 64 is only capable of 120 dB performance - and that is some 10 thousand times worse than Hugo - and a trillion times worse than Dave. And every time I hear DSD I always get the same problem o perception of depth - it sounds completely flat with no real sense of depth. Now regular 16 bit red book categorically does not suffer from this problem - an infinitely small signal will be perfectly encoded in a properly dithered system - it will just be buried within the noise.

 

Now the second issue is timing. Now I am not talking about timing in terms of femtosecond clocks and other such nonsense - it always amuses me to see NOS DAC companies talking about femtosecond accuracy clocks when their lack of proper filtering generates hundreds of uS of timing problems on transients due to sampling reconstruction errors. What I am talking about is how accurately transients are timed against the original analogue signal in that the timing of transients is non-linear. Sometimes the transient will be at one point in time, other times delayed or advanced depending upon where the transient occurs against the sample time. In the case of PCM we have the timing errors of transients due to the lack of tap length in the FIR reconstruction filter. The mathematics is very clear cut - we need extremely long tap lengths to almost perfectly reconstruct the original timing of transients - and from listening tests I can hear a correlation between tap length and sound quality. With Dave I can still hear 100,000 taps increasing to 164,000 taps albeit I can now start to hear the law of diminishing returns. But we know for sure that increasing the tap length will mean that it would make absolutely no difference if it was sampled at 22 uS or 22 fS (assuming its a perfectly bandwidth limited signal). So red book is again limited on timing by the DAC not inherently within the format.

 

Unfortunately, DSD also has its timing non-linearity issues but they are different to PCM. This problem has never been talked about before, but its something I have been aware of for a long time, and its one reason I uniquely run my noise shapers at 2048FS. When a large signal transient occurs - lets say from -1 to +1 then the time delay for the signal is small as the signal gets through the integrators and OP quantizer almost immediately. But for small signals, it can't get through the quantizer, and so it takes some time for a small negative signal changing to a positive signal to work its way through the integrators. You see these effects on simulation, where the difference of a small transient to a large transient is several uS for DSD64.

 

Now the timing non linearity of uS is very audible and it affects the ability of the brain to perceive the starting and stopping of instruments. Indeed, the major surprise of Hugo was how well one can perceive that starting and stopping of notes - it was much better than I expected, and at the time I was perplexed where this ability was coming from. With Dave I managed to dig down into the problem, and some of the things I had done (for other reasons) had also improved the timing non-linearity. It turns out that the brain is much more sensitive that the order of 4 uS of timing errors (this number comes from the inter-aural delay resolution, its the accuracy the brain works to in measuring time from sounds hitting one ear against the other), and much smaller levels degrade the ability for the brain to perceive the starting and stopping of notes.

 

But timing accuracy has another important effect too - not only is it crucial to being able to perceive the starting and stopping of notes, its also used to perceive the timbre of an instrument - that is the initial transient is used by the brain to determine the timbre of an instrument and if timing of transients is non-linear, then we get compression in the perception of timbre. One of the surprising things I heard with Hugo was how easy it was to hear the starting and stopping of instruments, and how easy it was to perceive individual instruments timbre and sensation of power. And this made a profound improvement with musicality - I was enjoying music to a level I had never had before.

 

But the problem we have with DSD is that the timing of transients is non-linear with respect to signal level - and unlike PCM you are completely stuck as the error is on the recording and its impossible to remove. So when I hear DSD, it sounds flat in depth, and it has relatively poor ability to perceive the starting and stopping of notes (using Hugo/Dave against PCM). Acoustic guitar sounds quite pleasant, but there is a lack of focus when the string is initially struck - it sounds all unnaturally soft with an inability to properly perceive the starting and stopping. Also the timbre of the instrument is compressed, and its down to the substantial timing non-linearity with signal level.

 

Having emphasised the problems with delta-sigma or noise shaping you may think its better to use R2R DAC's instead. But they too have considerable timing errors too; making the timing of signals code independent is impossible. Also they have considerable low level non linearity problems too as its impossible to match the resistor values - much worse than DSD even - so again we are stuck with poor depth, perception of timing and timbre. Not only that they suffer from substantial noise floor modulation, giving a forced hard aggressive edge to them. Some listeners prefer that, and I won't argue with somebody else's taste - whatever works for you. But its not real and it not the sound I hear with live un-amplified instruments.

 

So to conclude; yes I agree, DSD is fundamentally flawed, and unlike PCM where the DAC is the fundamental limit, its in the format itself. And it is mostly limited by the format. Additionally, its very easy to underestimate how sensitive the brain is to extremely small errors, and these errors can have a profound effect on musicality.

Back in 1999 we could read in a paper on High Efficiency Power Amplifiers that "Unfortunately 256fs=11MHz is too high for power switching" (http://icd.ewi.utwente.nl/publications/icpub1999/dis_rvanderzee.pdf) and yet in 2004 Sharp developed just that -- a power amplifier with 11.2896 MHz 1-bit switching (Sharp SM-SX300).

 

 

Below I quote some interesting info on the Sharp development:

 

Sharp Corporation has been the first in the world to succeed in developing an 11.2 MHz (256 fs) ultra high-speed 1-Bit digital switching amplifier technology that boosts a switching frequency nearly twice as high as previous techniques by further upgrading its proprietary 1-Bit digital amplifying technology, which is capable of faithfully amplifying and reproducing the original sound.

 

Capitalizing on this technological development, Sharp will continue to develop new-generation 1-Bit digital amplifiers that have a much higher audio resolution and can reproduce sound with a strong, clear rise, thereby pursuing the ideal of “reproducing original sound.”

 

In August 1999, Sharp was the first in the world to release the SM-SX100 1-Bit Digital Amplifier, which enabled high-speed 1-Bit digital switching with an approximately 2.8 MHz (2,822,400 times per second) frequency 64 times (64 fs) greater than compact discs. In addition, by switch-amplifying 1-Bit digital signals, the minimum unit of audio information, it reproduced the original sound recorded via microphone as is. The digital device also realized substantial reductions in energy and size.

 

In December 2001, Sharp released the SM-SX200 1-Bit Digital Amplifier, which enabled high-speed 1-Bit digital switching with an approximately 5.6 MHz (5,644,800 times per second) frequency that is 128 times (128 fs) greater than compact discs. The newly-developed 1-Bit amplifying technology has quickened sampling and 1-Bit digital switching by four times and two times as much as the SM-SX100 and SM-SX200, respectively, thereby boosting the minimum unit of switching pulse to approximately 11.2 MHz (256 fs: 11,289,600 times per second). With this ultra high-speed 1-Bit digital switching amplification, audio can be faithfully reproduced to a much greater degree than previous amplifiers.

 

• Key Points in Developing 11.2 MHz 1-Bit Digital Amplifying Technology

 

1. Developing modulation LSI processing 1-Bit digital switching at 11.2 MHz (256 fs) and bridge circuit having ultra high-speed switching performance

 

2. Consolidating more precise and stable amplifying performance achieved by “dynamic feedback method” for high-speed reverse feedback of switching pulse information

 

11.2 MHz 1-Bit Digital Amplifying Technology

As a fundamental technique for 1-Bit digital amplifiers, we employ “7th modulation algorithm” to expand the dynamic range of audio signal response, reproduction frequency band, and secure stability free from a state of oscillation by putting modulated quantization noise control technology used in the process of 1-Bit signal generation in high order.

 

The current technology was able to achieve amplifying performance at the maximum switching frequency of 5.6 MHz (128 fs). With the newly-developed 1-Bit signal generation process, the switching frequency has been doubled, reaching the final level of switching frequency at 11.2 MHz (256 fs).

 

As a result, by setting up a 1-Bit signal generation circuit (LSI) and quick-action switching circuit that enable 11.2 MHz (256 fs) switching control, the following can be realized in principle:

1. Reduction of modulation/quantization noise levels

2. Expansion of band of quantization noise

3. Expansion of output breadth levels in stable performance

 

The above have substantially improved audio performance.

1) Obviously digital systems by definition are *finite* not *infinite*. There is not actually a PCM signal that is nor can be actually infinitely small.

2) When using dithering to represent arbitrarily small digital signals the actual digital "PCM" signal starts to look more and more like a "DSD" signal.

3) Consider a extremely small "analog" signal for which we wish to represent as a stream of electrons (hmmmm....). At some point the signal might be encoded as a single electron / ns (flux). Now representing a smaller signal, say a single electron / ms and so on ... so you see that when dealing with the smallest signals a "DSD" encoding is what you get ... at some point the upper "n" bits will all be 0 and dithering encodes the LSB as 1 or 0 with successively more 0s as the signal gets smaller.

 

At this specific edge case (which he selected) the "PCM" and "DSD" signals are the same.

 

Whether they are lost within the noise, or shaped away, or preserved is actually implementation dependent.

    A converted signal can never be better than the original, that is all.

 

Depends entirely on how you define "better" - if you mean "sound better" then we quite disagree.

 

We also disagree if you mean it cannot have more information than the original signal. Interpolation adds information to the signal every time.

 

DSD can sound better than the PCM it came from. Upsampled PCM can sound better than the original RedBook format.

 

For that matter, MP3's sound better when up sampled.

I'm talking about purely objective measures, not what "sounds better." If a music recording is stored in a format that imposes a certain noise floor, distortion, or other deviation from the production master, no amount of upsampling or conversion afterwards can restore the original.

 

      Paul R said:

    We also disagree if you mean it cannot have more information than the original signal, we also disagree. Interpolation adds information to the signal every time.

 

Only if you use a non-standard definition of information. Although an upsampled signal occupies more storage space, the information content remains the same. A zip file contains the same information as the uncompressed file(s) it was created from. This is no different.

 

      Paul R said:

    DSD can sound better than the PCM it came from. Upsampled PCM can sound better than the original RedBook format.

 

It depends on the DAC. If your DAC has poor upsampling filters and you use a good one, then yes, it will sound better. If the DAC has decent filters and you use a poor one, then it will sound worse. Upsampling in and of itself doesn't alter the sound although specific implementations (might) do.

 

      Paul R said:

    For that matter, MP3's sound better when up sampled.

 

For certain decoder and filter combinations, probably. The MP3 decoding process is rather loosely specified, so there's a lot of room for one decoder to be better than another. For example, a decoder using floating-point processing internally is almost certain to be better than one using 16-bit fixed point.

    But haven't your PCM files already been filtered?

    I can see upsampling first if you want to do additional processing, so that has less effect on the original audio, but I'm not really sure that there are any filtering benefits for files which are already in a PCM format.

 

Yes, it has been decimation filtered already. But for upsampling/oversampling it is in most cases anyway going to happen. It is more about the choice where and how you want that to happen.

 

      Quote

    I can see someone perhaps wanting to upsample if their DAC uses linear-phase upsampling and you want to use minimum-phase upsampling.

 

Or use different kind of linear-phase upsampling, or use different kind of delta-sigma modulator instead of the one inside the DAC...

 

      Quote

    But I feel like most people that prefer DSD like it because it colors the sound.

 

Why do you think similar process inside a $10 DAC chip wouldn't have such effect? Most I've measured are horribly compromised designs because there's just not enough DSP processing power to do things properly.

 

      Quote

    Because DSD is a 1-bit format, there are a lot of compromises to be made so each modulator may have its own type of "sound".

 

So do the modulators inside ADC and DAC chips... For ADC side you have no choice, but at DAC side you may be able to bypass the built-in modulator by using DSD.

 

      Quote

    I hear this argument a lot, and obviously you are very knowledgeable on the subject, but I'd much rather send it a multi-bit input than convert everything to 1-bit first.

 

When you send PCM to the DAC chip, in most cases it will use digital filters to first go to 352.8/384 kHz sampling rate. And then use sample-and-hold (zero-order-hold) oversampling and delta-sigma modualtor to reach 5.6/6.1 MHz D/A conversion rate.

 

When you send DSD to the DAC chip, in best cases no digital processing is applied at all and it goes straight to the D/A conversion stage. Or some minimal digital processing is used (ESS Sabre) before going to the D/A conversion stage.

 

      Quote

    While some people may not consider them to be "multi-bit" DACs, they are not 1-bit DACs either, and do all of their internal processing in a multi-bit format.

 

So does the software that converts PCM to DSD... So yes and no, you can use same "multi-bit" DAC to natively convert DSD to analog. I have demonstrated this with my open DSC1 design.

 

      Quote

    So while you may prevent the DAC from doing any internal upsampling, it's going to go through remodulation which is more likely to have a detrimental effect on the sound, in my opinion.

 

The most detrimental effect on the sound is any SAH/ZOH or linear interpolation oversampling used in a DAC (due to resource shortage), which is most of the DACs.

 

So depends on which DAC you use. If you use for example DAC based TI/BB chips, it is not going through remodulation. Or if you use a DAC based on Cirrus Logic / Wolfson which enables the Direct DSD mode.

 

I've measured pretty nice linearity improvements on ESS Sabre with DSD inputs compared to PCM inputs.

For proper calculations better use floating point format. Better 64-bit.

 

It allow:

 

1. Decrease quantization noise

 

2. Avoid overloading. We can normalize signal only of last stage.

 

Releasing 32-bit floating point math at budget FPGA or microcontrollers is non-trivial task.

 

Need bigger time of development. Possible using C++ (single chip computer) or VHDL (for FPGA) languages. But it is not make this task more easy.

 

Many troubles with debugging. Here possibly using "hardware debugger" - oscillograph :)

 

When you use PC - no problem. Except CPU's resources consuming of course.

    But I feel like most people that prefer DSD like it because it colors the sound.

 

No. My experience is that DSD leads to more focus to instruments, less smudged details - less sibilance, less sharpness/aggresivity in trebles.

 

More "air" and generally better imaging and more resolving timbre of instruments is coming because of higher detail level. That can be only result of less distortion. Higher level of detail is prerequisite to hear more reverb and thus to get better spatial information about depth and the recording "room" in general. With PCM, especially 44.1k, recordings sound me more flat, it's much harder for me to perceive depth.

I guess you don't know much about me or others, at least you are demonstrating lack of knowledge. I would advise not to make statements about other people whom you don't know.

 

You don't need to go further than DAC datasheets or use some measurement gear to see where those fall short. I also publish measurement data and you are free to make your own measurements too. Plus I published my own discrete delta-sigma DAC design. I don't need to use DAC chips from anybody at all. I know how the stuff works inside out.

 

So how about publishing some objective data of your own to back your argument instead of trying to insult other people, thank you.

 

 

P.S. There are number of restrictions with DAC chips in terms of DSP. For example TDP of the chip, low clock speed, synchronous MCLK plus chip area and manufacturing process. And it is just plain stupid to put large DSP less than a millimeter away from a very noise sensitive analog part.

If the core of the DAC chip is DSD then PC software performing the PCM to DSD conversion can use the much higher horsepower in the computer to do a better job of that conversion. And at any rate, provide different algorithms for doing so. The HQPlayer is the most famous example of such software people use.

I plan to perform objective measurements of HQPlayer on ASR Forum one of these days. For now, I would say that while objective differences can be made here, I am extremely doubtful that they are audible differences. Audiophiles are men so they immediately think "more is better." :) In audio more can mean the same many times. But again, we need to measure it and apply psychoacoustics to the results to see if they are or are not audible differences.

This is precisely what's confusing me. If the original source files are usually NOT DSD or else are converted DSD just for the purpose of being played through a DSD DAC, what's the point of getting DACs with native DSD processing chips? Is the converted material still "better" in quality than high resolution PCM source played through a native PCM decoding DAC? Since we're talking about niche audiophile equipment here (not many average consumers are getting mobile or desktop amps and DACs), why are quite a few more recent DACs advertised to play "native" DSD256 and DSD512? That's what I'm trying to figure out. Are those niche DACs better suited for DSD playback and since I don't have any DSD material, should I be actively avoiding "native" DSD DACs?

DSD converted to PCM vs Original PCM

Hello, I am adding to my digital high res music collection. My current DAC does not decode DSD. I bought a handful of DSD downloads, and I converted them with an app made for Mac called DSD Master.  I also have a copy of JRiver which can handle this.

The idea is that if the DSD to PCM conversion is as good as an original PCM file, buying a DSD file may be more future proof.  I may be changing DACs eventually.

Can anyone comment on the quality of an original PCM high res file versus a DSD file, converted to PCM?

Wouldn't the better question be "how much better does DSD to analog sound than DSD to PCM to analog?"

Bits is bits – in the transmission and storage chain. I don’t think it makes much difference how your remember your digital samples – encoded as PCM or encoded as DSD or Apple Lossless or….

Nor should converting between these matter – as long as one doesn’t encounter numeric problems like overflow – because such format conversions don’t affect the data. If you can ’round-trip’ a conversion – say PCM to DSD to PCM – and recover the same bits, the data’s unchanged.

What does matter a great deal is the A/D and D/A converters used. The A/D is largely out of our control (for purchased digital music), but that leaves the D/A.

Pure single-bit D/A should sound better than any multi bit if only because there’s exactly one voltage (or current) source to hold perfect.

So the DAC matters, but not the storage, transmission etc.

Theory is wonderful, eh?

Hi Pete, you say ‘So the DAC matters, but not the storage, transmission etc.’

I’ve had a pair of Lyngdorf TDAi2200 amps in my system for many years now, they are DACs. I’ve had various iterations of source (e.g. Denon, oppo, Mark Levinson CD and many PCs). The TDAs only accept a digital input, XLR has remained my standard carrier ever since I bought them.

Do you include software in your ‘transmission’, because there are clear differences between the programmes I’ve used (e.g. WMC, XXHighend, JRiver) and between system modifiers (for want of a better term e.g. JPlay, Fidelizer, Audiophile Optimiser). I’m currently using Windows Server with no GUI as opposed to the XP/7/8 I have used.

Those are the soft components; I’ve replaced the PC PSUs with Linear PSUs. The ‘storage’ HDDs replaced with with SSDs. I use separate SSDs for OS and music and so on. Good SATA cables and shielded DC lines inside the PCs were a revelation. I’m using a P5 in front of the LPSs (and TDAs) and my PCs are on PowerBases, all physical changes before the DACs.

All the while my DACs have remained exactly the same – but there has been a massive improvement in the sound quality I’m getting; all contributed to by the above changes to ‘storage and transmission’

Hence your statement caused a raised eyebrow here

It all matters but it varies with different dacs.

Cheaper dacs it matters far more thAn much more expensive ones. And this is easy to hear

As good as the pwd mkii is the obvious improvements with changes in the chain matter. As I use an offramp to archive it’s best.

But I do have. Couple of other dacs not so much. Bu what really matters is how much we need to find out. As this is what I do not have. As the new DAC will work with all setups much better and less playing this is far more important than a SS or spin drive

So are we talking about a synerio where we have PCM converted to DSD which removes noise bits and then it gets converted back to PCM minus the noise?

If I covert an MP3 to AIFF, how can that be ?

I assume an MP3 is 10% and an AIFF is equal to 100%, how can you make a better file if you start with less information ?…

That’s random – all of these easily dismissed questions are for NOAH to determine. I am the expert but I cannot say anything yet without experiments. I do know however that SBMD hasn’t been talked of except once, in 2005 forum, and now I know hardly ever is DSD sourced PCM mentioned specifically of being good for getting back DSD. They’re talking about all the other stuff.