Question for Mr. Smith please. About toslink


#41

Yes, dvorak is substantively correct (all three lines of I2S are needed because the duty factor of the data bits aren’t defined so the FPGA samples the clock and data at the same time and then remembers the data if the clock went from zero to one.)

But more fundamentally all inputs are always being processed all of the time, nothing is saved on one input by using another. Internally data is shuttled around in a 24 - 30 bit bus (with DSD in DoP). All inputs get translated to 24 bit PCM (or DoP) as early as possible.

People should, of course, be guided by their experience, but assuming an input sounds the same on all DACs may keep you from experiencing something unexpectedly good.


#42

I suppose Toslink, Coax is whatever sounds better to you. That is the only guide. Different means to an end. Neither is wrong. Whichever sounds better to you. Seeing as I have a $3,195 Coax cable I am not running out to get a Toslink cable. Although I could get a good one used for $20 to try and experiment. Toslink cables are much less money than Coax. Which indeed might suggest something. Not sure. Although A/B such minor differences could drive one crazy which I do not need. I am very pleased with the sound of the DS in this system. It was a good fit. After “shooting out” Many DAC’s. Some at 5x the price. Even the MSB did not sound as good in this system. To resolve that you need big bucks in other equipment. Although the system it is in is well into six figures. Just goes to show. Your ears and your ears alone.


#43

I just A/B/C’d toslink, AES6 and Coax. Coax might sound a drop better maybe due to quality of the cable. It is too close to call. I say just choose whichever one is convenient for you. Unless anyone can actually hear the difference. Apparently they can but I do not. MR. smith I believe prefers Toslink but others preferred Coax. I2S may indeed be better but I do not have it. I thought AES6 would be better but I can’t really tell much difference between any of them. I imagine the DS is designed to handle them all well. However high end CD transport sounds vastly better than music server on bridge. Or PC on USB for that matter. Whichever connection I use on it.


#44

Thanks @tedsmith and @dvorak. I can’t claim to fully grasp all of this, but I think I’m getting the gist.

The question I have now (and the reason I had the impression I had earlier) is - Why the hype about i2s keeping the signal unconverted? At least, that’s what I had thought - that the advantage of the DMP/DS via HDMI cable was that it remained i2s throughout. And that i2s was the protocol used between boards in the DS and so on, so the less conversion between formats, the better.

Marketing for the less informed or technically inclined? Something no one would ever have known in a typical company because they never would have discussed the workings in such detail ; ) ?

And everything is turned into 24 bit PCM prior to being converted to 20x DSD? And that’s all in the FPGA vs. being one device connected to another via i2s (or whatever protocol)?


#45

Many people repeat a simple explanation if it seems to be reasonable, but that doesn’t make it right. Also the DS rejects more input jitter than many DACs so any effects that are based on jitter may be different with the DS (hence my continuing recommendation(s) that people try at least once connections, cords, tweaks anew with the DS. Some things that made things better with some other DACs may make things worse with the DS and conversely.)

Anyway, in many systems I2S could have been better in systems because it separates out the clock from the signal and hence doesn’t add as much jitter to the system as S/PDIF or AES/EBU might. That’s probably not the reason it’s different in some systems with the DS - instead the differences with I2S in the DS are probably grounding related. I2S (over an HDMI cable) is a set of differential pairs of conductors which lowers common mode noise along the cable. But probably more important HDMI cables have significant grounding area on their connectors and significant shielding and grounding along the cable.

Personally I think they made a mistake not supporting TOSLink as another output on each pair of channels on the DMP :slight_smile:


#46

Uhhh…I won’t mention where the notion originated ; )

Please understand this is not a diss in your direction, more utter ignorance on my part. Thanks for the reply.


#47

I’m pretty sure I know at least one probable source :slight_smile: I’m not taking anything personally, and I hope no-one feels like I’m picking on them.


#48

Not exactly. There are two concepts to grasp here.

First, DoP isn’t really PCM. I mean you can treat the bits as if they’re PCM but all you get is a mild square wave tone with some low level white noise. The square tone pattern is what clues a DoP receiver in to the fact that the lower 16 bits of the 24 bit “PCM samples” are actually DSD data. DoP is just a way to get raw DSD data through a PCM-oriented transport chain. So it’s not really “turned into PCM”, just packaged for moving to the next place.

Second, DSD is really just a special case of PCM. Instead of 65 thousand amplitude values (16-bit) or 16.7 million amplitude values (24-bit) it has 2 (1-bit). The waveform it describes is either fully positive or fully negative in each sample, which seems weird, but as a waveform you can treat it in exactly the same way that you treat multi-bit PCM. You can upsample it, low-pass filter it, scale it. Which is precisely what the DS FPGA does to all input signals before re-modulating everything down to a new 1-bit stream for output.

So DSD isn’t “turned into PCM” in the way that most people would interpret that phrase. It’s momentarily packaged as DoP for convenience but then processed in a way that blows minds because everybody thinks DSD is something totally different to PCM when like I said earlier it’s really just a special case of PCM.

The USB receiver chip outputs on I2S lines, as does the bridge. The HDMI connector is exposing I2S lines to the outside world. All of those run directly into individual input pins on the FPGA. And likewise the Toslink, XLR and coax single wire SPDIF inputs run directly to their own FPGA input pins. There’s nothing else digitally processing the input signals in any way.

I was reading a thread here the other day and noticed Ted saying that I2S success as an input to the DS was probably because of shielding and grounding. That post was from mid 2016 :slight_smile:


#49

This is getting comical, all this gobledeegook; we still all can’t make our minds up the right way to hook things up. It’s getting clearer all the time why vinyl still sounds best.


#50

…ever been on a cartridge alignment forum?


#51

Yup - get that, sorry. DoP is not DSD.

Uh - that does not compute. It may be in the case of the DS DACs as you and Ted have explained here, but DSD has never been PCM, except when converted to or from it. This is the thing - I’ve used PCM ever since it was created, and DSD has never been PCM, and DSD has always been better, IMO.


#52

I so totally feel that lately - Vinyl makes the majority of this stuff sound…OK-ish.

The thing is that there is no “correct” way to hook any digital stuff up at this point.


#53

Beef, my daughter’s BF is a math post doc. He did a bit of math when he was here overThanksgiving and told me DSD and 24/96 are pretty close in total numbers of bits. Unless I’m forgetting the comparison, which is possible, because he doesn’t drink.


#54

omg, no.


#55

Oh, for sure - but it has never been about the total bits, but how they are recorded and played back. Ask Cookie Marenco.


#56

Could well be. His point was that SACD was a big step up but in theory at least 24/96 has caught up.


#57

What a great name


#58

The Cookster is a Hi-fi and DSD hero. Props, gurlfriend! : )

https://otrstudios.com/founder

One I have been listening to lately is her recording back in the day of the Tony Furtado Band. Still gorgeous sounding. She knew what she was doing back then. May have something to do with why Blue Coast sounds good.

@cookie

:elf:‍♀:exploding_head:


#59

Yeah this is the part that breaks people’s heads.

Both PCM and DSD describe a waveform by plotting amplitude over a fixed sampling time interval. They are both subject to Nyqvist’s rule. They both have quantisation noise. From a mathematical perspective they are the same kind of animal.

In a practical sense, for the purpose of encoding and reproducing audio, they have some important differences. By using 16 or more bits of resolution per sample, PCM aims to capture all (or nearly all) of the sub-Nyqvist frequencies with an effectively inaudible amount of quantisation noise. DSD uses ultra high sampling rates to exploit the fact that the extremely high quantisation noise of 1-bit samples can be shaped so that its energy falls in frequencies which are far higher than the actual passband of interest. (Shaped dither is also used in PCM to improve the noise floor at lower frequencies and move the energy of quantisation noise into higher, less audible parts of the spectrum.)

So while 16/44.1 PCM with random dither has a flat -96dB noise floor from 0 to 22.05kHz, 1/2822.4 PCM aka DSD64 would have a noise floor of -6dB from 0 to 1.4MHz without noise shaping! Using noise shaping we can get a very good SNR in the audio passband, with an increasing amount of noise pushing up towards 0dB at much higher frequencies. Then we just get rid of that unwanted noise using a low-pass analog filter.

There are therefore some differences in how you create useful 1-bit PCM audio samples vs 16- or 24-bit audio samples. Most ADCs are actually high sample rate low-resolution devices these days anyway, using noise shaping and sigma-delta modulation to create something more like a DSD stream that then gets mathematically filtered and converted to PCM. In that sense a pure DSD recording is closer to the original than most PCM recordings.

There are differences in how you process the data digitally. High resolution PCM can be scaled, mixed and processed in quite a linear fashion with relatively slight increases in noise. DSD can’t be treated that way because even the slightest modification gives you back your -6dB noise floor.

Maybe most importantly there are differences in the machinery needed to do D-to-A conversion. Ted’s core insight with the DS is that a DSD stream needs nothing more than a positive/negative voltage switched according to the data stream, followed by a gentle analog low pass filter. That’s so different to a traditional audio DAC that it’s probably why people think DSD and PCM are completely different things… the physical machinery for D-to-A conversion is so different that they are mostly one or the other.

But at the mathematical core, this is the truth: DSD is a special case of PCM.


#60

So - I appreciate your perspective, however… shoot me. Sorta like saying one sort of analog device is the same as another. Which may or may not be true if observed from a purely mathematical perspective. Which however, has never had anything to do with what happens on the User end…sonically.

“Ted’s Core Insight”, and apologies to Ted, which he did not create, is the essential difference between DSD and PCM.

Argue amongst yourselves.

You would seem to be arguing Theory vs. Reality.