Tuesday, September 27, 2016

Jitter

Jitter is probably blamed for bad sound too much, when it isn't to blame, and when there isn't actually bad sound.  Jitter as high as 10nS has been shown to be inaudible.

Anyway, I've always wanted to measure it, and now, with my Sencore DA795, I can.

Unfortunately the Sencore doesn't show numerical values below about 200pS, it just says "Low".  Most of what I'm reporting now is just from watching the logarithmic scale meter, which has markings at 100pS and 200pS.  So I'm reporting highly approximate values.

For the Pioneer PD-75, I see about 150-180pS jitter when playing a CD, which falls to 110pS when stopped.  Looks like some room for improvement in the power supply.  However even 180pS is typical for the best equipment ever made.

For my newest Sonos Connect, I measure 180-220pS jitter, clearly a bit higher than the PD-75 and more variable too, measured at coax spdif output.

(BTW, these are the RMS measurements.  The Peak measurements are over 500pS and very unstable, but the Pioneer is about 40pS better there too.  Also these are all unweighted measurements.  I have no idea how the Sencore measurements would differ from those using J-Test for example, but for sure the Sencore is completely insensitive to jitter in the recording.  It is only looking at the clock that can be recovered from the data, not the data itself.  I suspect John Atkinson uses a J-Test derived weighted peak number.  It's beginning to look like my unweighted RMS numbers are lower, but only about 50%.)

In both cases I used 7 foot Belden RG6 with Canare RCA's as interconnect to the meter.

In their original report, Stereophile reported 388pS jitter for the ZP80, which John Atkinson characterized as Excellent.  I thought that they showed 220pS in a later test, but I have been unable to find that.  In the 388pS test he's using J-Test on the analog output, you would think that to be comparable to the digital outputs.


I'm surprised that a heavy old mechanical device like the PD-75 actually beats Sonos on jitter.  But both numbers are excellent actually.  I would begin to feel worried with jitter above 500pS.  Some of the very best audio devices have jitter around 150pS, and often higher.  I'm not sure if I've ever seen jitter below 100pS for an audio device as such, though external clocks can have jitter spec down to 1pS or below.

Well, even assuming my meter to be perfect, there are 3 sources of jitter measured in the timing of the digital output:

1) The clock jitter of the pioneer, and everything that may ultimately affect the clocking out of data at the SPDIF output.  Audiophiles obsess over things like lack of power supply regulation for the transport section (which may be on a different transformer in this unit anyway) affecting the regulation of the clock.  And so on.

2) The cable losses and smearing (I proved this to be a negligible factor by seeing no difference with a far longer cable connected).

3) The jitter made inevitable by the coding of the spdif signal and practical transmitters and receivers.  I do not believe the clock can be perfectly recovered due to the nature of SPDIF data.  That is why the meter specified range only goes to 150pS with SPDIF input.  It goes to 35pS with a pure clock signal at the clock input.

3a) This varies depending on the actual data.  The Dunn test is worst case for 44.1/16.  Most music is less.  A continuing silence would probably be the least jitter, and for that reason I may record a test case.  I love being able to make my own test recordings btw.  Is this possible at low expense with DSD?  I can't make an SACD for 9000ES but I could apparently make a test disc DSD-File or something like that for my BDP-95.

3b) The actual properties of the Pioneer SPDIF transmitter circuit may also be a factor, say compared to a "perfect" implementation, perhaps an over-implementation.  I already set the baseline at "practical" and presumably the Pioneer isn't at that level perfectly, it may not be as good as something could practically get (say, without an Apollo Project--that's impractical for this) in impedance or risetime, for example.  I imagine some obsessing over such things (and even seeing me writing about them think I obsess over them either...no to my casual observers, I merely pointed such things out, unlike say Lampizator I'm only unsure of how important they are in the overall picture, but I'd expect a Pioneer Reference player like the PD-75 to do a pertty good job already at the powering the SPDIF output, and so it seems too, but who knows, it might be 2pS better with a bigger FET running more ten times more current at the output, on it's own separate transformer of course, blah blah, modifier believe they can may everything better).

What is that "minimum" jitter for SPDIF?  I need to read Dunn's articles.


Toslink out!,  AES/EBU vindicated!


I didn't think this was going to be hard.  Several times this week, usually just before going to bed (already way too late), or going to work, I was tempted to hook the jitter meter right at the end of the Glass Fiber Toslink cable I have connecting the Behringer 2496 DEQ to my Audio G_D DAC19.  This would show the evil Toslink jitter, which I didn't imagine to be very bad--especially with such a good cable--and actually the whole system jitter right at the critical point, the DAC for the panel speakers.  And the whole system is what many people consider complicated, many different boxes connected by digital cables that are mostly AES/EBU, including splitters, converters, EQ's, and a digital preamp.

Finally on Saturday night I was determined to do the measurement.  This would be almost worth the price of the meter, it's what I got the meter for!  (Well, actually among many other things...)

And there were SNAFUs right from the start.  It's as if fate itself did not want me to make this measurement.  First, I discovered my DA-795 meter doesn't accept Toslink inputs to the Jitter test.  It's not really clear because it allows "SPDIF 2" and the Toslink input seems to be on the SPDIF 2 side.  I was also maddened by the fact that I couldn't just select Toslink, and I figured that was the problem, that I had selected SPDIF 2 but needed to make some further selection related to the dozen or so icons at the bottom of the page to specifically enable Toslink.  I tried some of those to no avail.  I tried some other test functions and some of them did enable the Toslink input, but you often needed to have an additional input also, such as for the Transparency test where it compares one stream with another.

Finally, I brought up the manual on my computer and sure enough the Toslink jack that was seemingly on the SPDIF 2 side is specifically identified as being a SPDIF 1 input, and the Jitter test ONLY works with SPDIF 2, either coax or AES/EBU.

So the meter maker wimped out!  They didn't even want to try to let you see the Toslink jitter, for possibly the meter might be blamed!

So I determined to find another path.  So I removed the CO2 toslink converter on my Mac Mini along with its AC adapter and a short cable.  I would do the conversion from Toslink to Coax myself, using a converter which has proven itself time and time to be the best, and perfectly reliable.

Fortunately it's much easier to work on this stuff now that the living room floor back around the DAC is at least partly accessible.  But not perfectly accessible, it still requires trick kneeling and leaning on things and so on, all very carefully so as not to damage anything, including especially myself.

But as I was setting up the CO2, every time I nudged the AC power cables toward the wall, I heard something that sounded very weakly like sparking.  Then I just rotated the big 3-way adapter plug in the lower position of the super heavy duty Pass & Seymour Spec Grade outlet, and my UPS started running on backup power.  I played with this only a second time before panicked, I shut the digital equipment off, I shut the UPS off, and I unplugged the big 3-way adapter plug from the wall.  My girl friend had asked me if the big 3-way adapter was OK at least once.  I had actually asked my electrician about changing the two outlets to 4 outlets and he said it was possible but that was months ago.  An alternative needed to be created Right Now!

So this was another emergency project.  I would not operate my system again until it was safe!  I decided to plug the UPS straight into the wall, then plug the subwoofers into the TV power strip which was right there, and unplug the TV and a TV adapter to make room for those plugs (there are no spare outlets in any of the strips--this is the usual situation).

Now by this time the Jitter Meter had clicked off because of low battery, so I had to put it back on charging for a half hour.  Another delay.  But I was determined!!!

So then a half hour later I see that the Jitter Meter doesn't show any battery indicator like it's supposed to, but it runs OK on battery power, so I'm finally able to set up my test.

And I'm not getting any digital lock!  I reconnect wires, fiddle about, nothing.  Finally I notice once and awhile the jitter arrow shows up at about 500nS for a split second, then disappears.

So then I tried using coax instead of Toslink, but this was complicated because the DEQ's only have Toslink and AES outputs, nothing for Coax, and I haven't seen my AES-to-Coax adapter in years.  So I reclaim the adapter that was currently (and now previously) in use creating the coax line for the supertweeters.  I connected that to the midrange DEQ, and got a short coax cable to measure the jitter at the end of.

So now finally I'm getting lock and jitter measurements.  What at first I see is a bit troubling, the jitter playing a CD on the Pioneer PD-75 being redigitized to 24/96 by the Lavry and through my system (I figured the Lavry clock was about as good as it gets...and that's basically driving all the downstream clocks, so this should be nearly perfect and it has always sounded wonderful) was showing around 500pS jitter.

Then I try switching the digital input to Sonos, and I'm seeing only 280pS jitter.  That's not to bad for a whole pile of boxes connected to Sonos which itself measured about 220pS jitter.  All that extra stuff is only adding about 60pS more jitter.

So then I reset the Lavry to 44.1kHz, and the jitter is now only 290pS, just a tad worse than Sonos.

Playinng non-silence through the Lavry at 96kHz was showing about 380pS jitter.

I find this all to be quite signal dependent.  Digital silence is low jitter, classical music medium, hard rock somewhat higher, and background white noise (the -118dB noise of the Lavry itself) reads very high jitter like that 500pS I saw originally at 24/96.

Perhaps the Lavry show was showing a tad more jitter at 44.1 because the program was slightly different, or because the Lavry itself uses the lower bits more effectively.

Anyway, I'm not going to let higher jitter dissuade me from using High Rez!  I think the Rez is more imporant than 100pS more jitter, which may be averaged out by the DAC anyway.

Also, I'm not going to be planning to use Toslink any more.  My very complex system yields what I now consider surprisingly low jitter using only coax and mostly AES.  Only a tad more jitter is added because of all the complexity of AES DSP units, mostly the jitter seen is mostly only what was present at the source.

All my AES splitters and converters have now vindicated themselves, and I'm getting a second HOSA AES to Coax converter.  Meanwhile, the one unit I have is working for the panels and I've run Toslink to the supertweeter DAC which is less critical I think (can hardly be heard anyway).

My high rez line from the kitchen has always used 2-stage Toslink reconversion, because that is the only practical way to run this from my Mac to several devices.  I wrote about this here years ago.  Now I don't consider that suitable for high rez and I've ordered a New In Box Logitech Transporter.

When I saw the review for the Transporter in 2009 I regretted having gone with Sonos, but still at $1999 it was too expensive I thought and I didn't need much of the functionality, such as the analog outputs, and I would like other functionality, such as analog inputs to be available on the network as Sonos does.

I've had alternatives for handling high resolution discs, but my alternative for high rez files was through the Mac Mini and the dual toslink connection, which I now know needs to be avoided.

With the Transporter I get local control with a remote, no need to buy and maintain a separate computer, and now at the $588 ebay price reasonable.   A more up-to-date system like Sonore would require a new USB DAC...  But I don't want USB!  Or a computer in my living room, constantly crying for updates and maintenance.

Transporter is essentially a little stripped down AV computer which can do a bunch of things, including selecting the music to play, playing it, and putting out analog or digital in AES or Coax.

Of course it won't do DSD (directly anyway, without buying another gizmo, a DSD DAC which can accept DSD over SPDIF).  I don't care about DSD64 files anyway, I don't think I'd ever want any.  But DSD128 has some promise.  I could get a Pono to play those, and get them redigitized through analog at 96kHz, which is basically what I have to do with all DSD files since I have DSP based crossovers.  Maybe tomorrow.  I had been thinking about getting a Pono all week, but the Logitech was more mandatory IMO.












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.



Tuesday, September 20, 2016

Clocks, clocks

We can tell the Pioneer PD-75 and PD-95 clocks are about the same by reading the modification instructions by Octave who makes clock upgrades.  The replacement instructions regarding the parts changes are identical.

So the PD-75 has the same clock circuit as the PD-95.  And BTW it's not the cheapest circuit, it's the step-up circuit.  In the cheapest circuit a crystal is simply connected to the relevant chip with suitable passive parts.  In the step-up circuit, the crystal and passive parts are buffered, in the PD-75/95 by two layers of opamp buffers.  An initial layer which then drives 3 separate buffers to send clocks to different things.

Not to say, there are a bunch of more sophisticated designs than the first step up.

But I think it's clear that these were intended to be in the elite upper category of CD clock performance (there were two categories established by the Red Book).  So we could expect perhaps jitter at least to be less than 300pS, allegedly 1/30 or less of what would be audible, or less.

The PD-95 might benefit even in jitter performance simply by having better power supply.  That has huge effect on simple oscillator clocks.  Good clocks are designed to be immune, though nothing is ever entirely immune from influence, within the same electronics.

Meanwhile it's bugging me that nobody talks about simply replacing the Sonos Connect (formerly known as Zoneplayer ZP80 and ZP90) clock the way one might do for a CD player.  (BTW, in some accounts replacing the clock is essentially the only thing that ultimately determines player-added jitter at the output terminals.  The clock ultimately controls the clocking out of bits, and it doesn't have to vary that rate one iota because servos earlier in the system keep buffers filled with enough data.  If not, there's a skip, which rarely happens anymore with good players and discs.  In other circles, everything counts.)  Instead, it seems the major recommended replacement now adds a whole circuit board which does reclocking on the output.  I would rather avoid reclocking because it only just smoothing the underlying jitter.  It's not really eliminating variation because it can't, it's shifting it somewhat to a lower frequency, so jitter becomes very low frequency wow.  And not only reclocking, but ASRC to 96kHz, or your frequency of choice.  That might actually be good for me, but I think I'd rather do analog resampling than ASRC.  And I've had a longstanding bias against ASRC as simply "buring the jitter in the data, where it cannot be removed".  So why not just replace the clock.  The blurbs all say "multiple lousy clocks."  I'd like to see that for myself, I still think fixing the actual clock is better than any band-aid ASRC.  If it's true there are multiple clocks, why not replace the one that counts most, or all of them, or something?

I suspect that the difference between the existing PD-75 clock and something better would not be by anyone audible in ABX DBT.  But of course, I gotta have the best clock anyway, just to be absolutely sure.

I've asked Kingwa if he still makes his Audio G_D jclock.




Monday, September 19, 2016

Replacement for Sonos

Something like this would be a start, the RoonReady Sonore Sonicorbiter SE.

Article also mentions a bunch of network audio frameworks:

SqueezeLite
ShairPort/AirPlay
MPD
DLNA
HQ Player NAA

Sonos recently has made it hard to select 0dB on a Connect so you can properly adjust level on actual system.  (Sonos is more and more thinking of connecting to all their own devices, rather than serving as a transport layer for music for all devices.  A slippery slope going the wrong way IMO.)

Sonos has never supported Hi Rez digital, or DSD, and probably never will.

Sonos gives no ability to adjust buffering size (and conversely latency) when sharing analog sources over the network.  It seems to be limited due to small buffering to a maximum of 5 uncompressed analog source connections on a Sonos network.

The one good thing that Sonos does that nobody else does is allowing analog connections across the network in the first place.  I cannot switch to any other system until it provides similar or superior analog source connectivity

You might notice this little Sonore device has no analog inputs.

Article comments also mention some other devices:

IQaudIO
Raspberri Pi Audio

In this article about a modded Connect, yet more things are listed:

Denon's Heos Link
AURALiC Aries/Mini

BTW, I don't think RE-clocking and up sampling Sonos output to 96kHz is necessarily the way to go.

Reclocking in general is band-aids, all that re-clocking can ever do is low pass filter the timing variations.

What is needed is much simpler: add a word clock input so external clock can be used!  (OR, bolt high quality clock inside.)





Sunday, September 18, 2016

DVP-9000ES SACD

I'm very much enjoying the sound of my new DVP-9000ES playing SACD's.  I'm feeling that, indeed, Sony did something special, at least with these early SACD players, that I have not heard in non-Sony branded players.

Something different anyway, perhaps not all to the good I'm wondering.  Nowadays even Sony uses highly oversampled multibit DAC chips, sometimes even from the likes of Burr Brown, to implement SACD, just like everyone else.  Back when the first SACD players were introduced, Sony was using their Pulse Converter chips, CXA8042AS.  Those pulse converter chips are used in both the SCD-1/777 and the DVP-9000ES.  Outside of that, and the use of one OPA213, the output circuits are quite different between the super high end and merely upscale, with the super high end showing far more additional stuff, and discrete circuits in the output.

A leading audio engineer, Stanley Lipschutz, and several others, revealed inherent faults in 1 bit delta sigma in AES papers in 2000, just after the public release of the SCD-1.  The principle fault with 1 bit delta sigma is that the background noise isn't gaussian, it's tonal, with idle tones.

Sony then denied, to John Atkinson, that Sony was using 1 bit conversion, in response to a paper by David Rich describing the concerns of Lipschutz and others, published (of all places) in Stereophile.

I'd always wondered if that denial was with regards to future SACD, not past.  In the past machinery, as in all the machines I just mentioned, still for sale but designed years earlier, they might have used 1-bit, but that was now water under the bridge.

I don't know enough about these CXA8042AS chips actually to confirm my version of this, that Sony abandoned true 1-bit when it became apparent that it didn't down scale well, or they hit a brick wall in making further improvements, or were covering up faults all along, or were embarassed by Lipschutz, or something like that.  They had been able to achieve brilliant sound in the 3 early players by perhaps mixing things up a bit (especially in the SCD-1/777ES) so that the 1-bit fundamental character was somewhat obscured.  I think they may have been further oversampling the 1-bit, IIRC 70Mhz was claimed for an earlier Sony ES CD player.  But in this regards, the 9000ES is actually pretty straightforward in the analog section, aka simple, and in that regards very different from SCD-1/777ES, but neverthess (or alternately) similarly good sounding.

The alternative is they never did anything like true 1-bit all along, but that seems to be rewriting history.  1-bit arrived on the scene I think starting with Pioneer machines, and perhaps others, in the late 1980's with great fanfare.  One trademark name was Bitstream, another was BASH.  It may have been Sony, who had been plodding along with Philips chips, that was the follower, coming up with their own 1-bit chips to compete with Pioneer in pushing linearity beyond -80dB down to -100dB and maybe even -110dB and beyond.  So the 1-bit race kept on during the 90's, with Sony ultimately achieving within 1dB of the theoretical possible CD THD+N performance with their 997ES and possibly 707ES and pusing linearity out to -110dB.  Pioneer missed this considerably in their PD-75 but I'm not sure about the higher end ones.

But already, by the time say of the PD-S06, Pioneer seems to have been questioning 1-bit.  That machine and an increasing number of later machines, switched to using full width multibit PCM chips like PCM56 and PCM63 especially.  Or at least they were letting customers have a choice, as early as with the PD-93--which used PCM63, but I think Pioneer had completely abandoned 1-bit by the time of the Lipschutz papers.

But Sony had already tied its hands to 1-bit distribution (if not implementation) with SACD, and kept promoting 1-bit until the launch of SACD, which featured 1-bit based players, but were quietly walking back the principle superiority of SACD from being 1-bit to being, well, whatever the marketing agency could think of.

But why all this fuss and bother if the first SACD players sounded just fine?  Well perhaps it's not just about making things good, it's about making them better and better year after year.  It was too hard to keep making 1-bit better (through various trickery and overengineering?).

Anyway let me also say that DSD is perhaps not as bad as I've written about here-to-fore.  While a pure 2.8Mhz 1-bit delta sigma system would be horribly information-lossy, and correspondingly noisy, that isn't at all like what DSD is.  DSD achieves low audible noise, AND low information inefficiency, by noise shaping.  Noise shaping is not just an add on, it is fundamentally what makes SACD possible, and that may be what I was not thinking about properly.  It's improper perhaps to call DSD a Delta Sigma system, it's a Delta Sigma Noiseshaping system.  Noiseshaping takes the place of the structuring imposed in PCM by coding.  That's the difference here: noiseshaping vs coding.

Noiseshaping compensates for the information loss below 20kHz in a comparable 1-bit delta sigma system, and somewhat more.  If it weren't for noiseshaping, the noise would be the clear sign of information loss.  But actually DSD achives better noise performance than CD in the midband.  Therefore, it is transmitting more information there.

And here, I'm not sure how to make the full comparison.  One way would be to consider the full band 20-20k noise performance as showing the "information loss" of the system.  And the other way would be to consider something like the A-weighted noise performance as showing "the effective information loss" of the system.  By the latter measure for sure, and I'm uncertain for the former, DSD arguably achieves not information loss compared to 16bit 44.1kHz PCM, but information increase.

There's no question still that DSD is an inefficient system in transmitting information, PCM is far more efficient.  But it could be said to be satisfactory in total information transmission, having reached CD quality and somewhat besting it, and having different character in the details, in ways that could be audibly pleasing.

Anyway, I'm thinking that the noise performance is a good indication of information performance, and one can just look at the noise curve of DSD and say it has more information than 44.1/16 in the midband.  I'm only not sure to which the increasing noise in the upper octave cancels that anymore, it might not cancel all of the advantage.  But anyway, the noiseshaping component means I can't simply apply my previous simplistic calculations.  I have to account for the effect of the noise shaping in shifting available information bandwidth from super high frequencies down to useable ones (in the reverse direction to the noise).  So another term might be information shifting.

And I wouldn't worry about the gaussian noise as much as tonal or correlated noise, and noise whose proporition increases or changes character at lower levels.

One way to explore "the details" in the sound would be to make a PCM or DSD recording at artificially low recorded level, say -60dB.  Then playback with lots of gain, and see which sounds better.  (This is actually more complicated than it sounds.  Should the gain used be PCM-like or analog, for example.)

Of course we don't really have to be cynical to see that the big win for Sony with SACD/DSD would have been IP and the big industry selling point was DRM.  Meanwhile I don't doubt that the lack of openness was a big hinderance in the end.

I don't think it's possible to consumers to make SACD's the way they could make CD-R's and CD-RW's, and that's part of the plan.  Interesting that the earliest SACD machines from Sony may not even have supported CD-R and/or CD-RW, though most CD machines of the time did.

The lack of CD-R/RW capability may have been a subtle hint to industry.  We own this, and we're not going to let consumers destroy your profit margins by permitting consumers to copy.

Anyway, WRT Sony, they visibly dropped any major concern for SACD and DSD sometime around 2006 when the Blu Ray vs DVD HD war was heating up.  SACD had not gone the right way, they might have figured, but lessons had been learned to win the next corporate battle, which Sony did.






Saturday, September 17, 2016

Legato Link

Legato Link is Pioneer's trademarked oversampling.  In constrast to most oversampling with brick wall filtering, Legato Link deliberately seeks to capture high frequency information above the Nyquist Frequency (?).  Some say it's merely a relatively slow rather than sharp brick wall digital filter--which might roll off highs earlier (but no evidence of that) or let some aliases through, and that appears to be possibly true, distortion levels would appear inconsequential to specification readers (the PD-75 for example specifies 0.0018, which would be -95dB, very respectible, out of -98.08 perfection).

Pioneer was doing interesting work in digital filtering from the beginning.  And so were others.  Pioneer used the term Legato Linear to refer to the highly refined filters in the PD-95 and comparable transport and recorder.  In my mere PD-75, they used no such term, however as it was the A version and the PD-95 got the B version one suspects it was already developed along a similar direction.  What Legato Linear means is not well explained, though we can guess the Linear has to do with things like linear phase response and/or linear filtering.  There's little evidence that Pioneer was attempting to reconstruct high frequencies, as claimed for Legato Link, though the distortion seems to be slightly higher than absolutely necessary, 0.0018% is, so one expects a compromise is being made with that and pulse accuracy, and if that's the case then only 0.0018% is doing very well, say compared with those now who would go to 1% and beyond with NOS.

The actual usage of the term Legato Link and the hype about reconstructing high frequencies only occurrs much later, possibly with the CXD-500, and then becomes commonplace (with additional adjectives, like Super) in the Pioneer Elite DVD players (and, sadly perhaps, one reason I never bought one, I didn't trust them, as I should have).

However, it appears by some measurements, that actual performance fell a bit short of those inconsequential looking specs, still somewhat insequential looking:

Peter Aczel, for example measure -92dB above 800Hz, but down to -78dB in one channel, or 0.012%, around 250 Hz and below.  He got identical performance in two units.  Published in The Audio Critic.

Julian Hirsch got 0.004%, or -88dB, 20-20kHz, which he seemed to think was quite good.

(These of course inconsequential compared with NOS designs, which may get 10% distortion in some audible frequency ranges.)

One interesting question here, is this digital distortion, which might be much worse as it would increase proportionately at lower amplitudes?

Perhaps the use of Spline interpolation prevents the digital distortion from increasing so much at low frequencies.  Anyway, now I'm a bit curious, and not at all turned off as I was by a friend in 1991 who showed me the pioneer advertisement:

I remember being shown this advert flyer during the early 2000's and I thought it looked pretty stupid.  I didn't then think that the high information retained would be anything but noise, aliases, and distortion.

I looked at that and said, that looks like aliasing distortion they're including.  They can't include high frequency information.  (Maybe they make it up by reasonable interpolation?)  So it must be aliasing and grundge.  I was highly skeptical, and moreso after reading various comments, reviews, and such regarding Pioneer machines.  I mean, I wasn't into buying machines like that much anyway, I was satisfied with the performance of my Sony 507ES I purchased in 1989 and still works today (thanks to a bit of ingenuity on my part, partly).  Anyway, Legato Link seemed like some sort of dangerous nonsense.

Meanwhile, in a somewhat pricier machine, also using bitstream dacs, Sony was getting -96dB 20-20k, and -97dB midband.  (That was for the X997ES, one of the best measuring machines ever, since then it's been downhill mostly for CD performance.  Perfection, again is -98.08dB and 0.001247% for THD+N.)

So color me as believing distortion perfection was being sacrificed in some way for time perfection.  Ok, I'm much more interested now than I was in 1991.  Very interested indeed, a lot of things from HDCD to SACD to HiRezPcm to MQA are all (or partly) about improving time accuracy, such as transient response.

For that reason and others I have some interest in the sound of the PD-75 with Legato Link and Pioneer Bitstream DAC's (which, btw have essentially perfect linearity down to anything you want, and have shown in some comparative measurements to have been, generally speaking models of comparative perfection) even if I've been (and will continue to be) doubtful of the full fidelity of anything 1-bit.

I really got the PD-75 as a transport, because the big mechanism, Me Tarzan or something, now it appears the clock is all important, wimpy transport with good clock is fine.

Still may be a decent transport, esp., with upgraded clock.  But perhaps just use Onkyo RDV-1 already having fairly decent clock, and (for sonics) 1704's.  And tis transport for anything allowing digital output.  So this may be my goto-transport for now, actually for the best sound, until and if I keep (probably) and modify the Pioneer (I may just play the Pioneer as transport, as it works, for it's "vintage" sound, jitter and all, probably still at low levels, maybe even unmeasurable contribution to jitter from the transport itself, just enjoy the vintage feel, and sometimes even the perhaps slightly amplitude distorted sound--but with better time response.)

I have an Oppo BDP-95 available through about 100 ft of high end coax wire (my home wiring network plus interconnects).  I'm not sure how much the spdif degrades, but I get perfect connectivity through my old Tact at 96/24.  (88kHz on the Tact died even with short cables long ago, but 96kHz works fine)




SACD and DSD in DVP-9000ES vs HDCD

I take this mostly as representative as SACD the system, not my particular unit or the DVP-9000ES more specivically.  Because it's how I'm thinking of it all.  I think this unit represents DSD as it was first implemented, with something very much like the 1-bit converters Sony had been using for years, and not the oversampled multibit implementations, which have been the norm ever since the 3rd generation or so.  And it's characteristic of DSD itself, in any implementation, but more purely in its original implementation, where the limitations are more obvious.  So I'm going to say DSD and not "my DVP-9000ES" but you can do the translation back if you don't accept that equivalence.

DSD has a smoothing effect mostly.  That's actually the subjective effect of adding noise above 10kHz, and possibly that's the cause of it.  But it can also open up dynamically, and suddenly get more bite than PCM.

I've recently noticed HDCD has an effect like that, apparently under the producer's control.  The producer can apparently control the sharpness of the digital filter in your system.  Some make it sharp, some make it soft, RR recordings you never notice because they do it frame by frame with the PMI recorder.  I think this is a good idea, to give the producer some control over the playback digital filter, and I like HDCD, perhaps an open standard system would be better than a proprietary one, but HDCD recordings as they exist are generally quite good and advantaged slightly over PCM 44.1/16 as a result.  There actually is a tiny information loss to power the coding system, but it's negligible compared with DSD.  Only a few bits are necessarily stolen by HDCD in full decoding (of course in non-decoding you are possibly getting a compressed form, so that part was not helpful to gaining wide acceptance--the "compatiblity" claim is actually overstated...then again the story in standard CD masterings being all over the map and endless remasterings to endlessly bilk consumers is no pretty picture either...there's really no compatibility in CD's every version is different, and one might be suspicious in hybrid CD/SACD discs that the CD is somewhat dumbed down to make the SACD shine, and in fact if they use 1-bit conversion throughout that would probably be the result).

DSD has this opening up and closing down aspect...but it appears seamlessly as part of the system, apparently driven by the demand for it in the music.  So that's good in some ways, less possibility of error, but I don't think it's much if any better than 44.1/16 PCM and I'd rather have HDCD or High Rez...but perhaps the same actual mastering used in the high rez recordings would work at 44.1/16 too, if they'd just do that.  Mostly the whole "High Rez" thing is to force people who want better masterings to pay more for everything.  It's a bribe to the Engineer not to crank up the compression machine.

I still feel the information loss in SACD (which I admit, is equivalent to the HF noise above 10kHz*, so well you might say, why worry, but there it is, that's what information loss looks like) gives the sound a simplified character generally, like MP3,  only very very slightly, just enough to make things sound better a bit, like blur on a photo.  Generally.

(*And I'm not exactly sure how to count the noise shaping, it hides the information loss within the audible range by creating vastly greater information loss above, but it can't add lost information.  Information loss need be applied to the raw S/N curve, not the noise-shaped one.  Actually, in a wide band, that would probably make it seem better anyway, but my information loss argument applies strictly speaking to the known audible range, and in that case I think it's pretty big.  I'm not exactly sure how to square this with apparent low noise in the midrange in measurements of DSD...but once again I think that's primary the bogus result of noise shaping, without which there would be noticeably more noise in the midrange and upper midrange.  In principle, despite noise shaping I think, DSD adds information up to the actual nyquist frequency, say 1.4Mhz more than equal to it's loss within the audible range, it uses lotsa bits.  But who cares about added information above 100kHz, which is filtered away in the end anyway?  Am I missing something here?)

So I wouldn't recommend DSD, but if a producer has chosen it, it's just another artistic brush, may not kill things for the first few hundred listenings, there's really so much information in music anyway, most of the time we're throwing away almost all of it, only sometimes reaching more than a tiny fraction.