Blue or red pill?

[mansr]

2 hours ago, STC said:

Again, it looks like the signal streams that entered Tascam and the Altmann were not identical. 

Whatever makes you say such a thing? There is no reason whatsoever to believe that the data transfer between PC and DAC/recorder was compromised in any way.

[adamdea]

40 minutes ago, PeterSt said:

 

He will next ask YOU. Watch him ...

no he won't: seriously you just misread my post. 

[fas42]

3 hours ago, semente said:

But apparently it's not the circuit, just a change in software settings, because one of the options sounds good/better...

 

The change in software settings alters how the hardware which is sourcing the recording operates - the nature of its electrical activity, over time, is different - the circuitry which is concerned with creating and processing the analogue representation should be completely oblivious to that change; but in the real world that is not the case.

 

It is the circuit, in the sense that it is too sensitive to its environment - the, obvious, solution is improve the engineering, after the fact if necessary, to stop this being the case.

[mansr]

13 hours ago, manisandher said:

"I've taken some 24/176.4 captures from the output of my Phasure DAC. I thought you might find the following two useful in determining the performance of the Tascam ADC..."

Same FFT parameters as before:

 

 

The noise floor is considerably lower. There are spikes every 400 Hz, which is still not normal.

[adamdea]

42 minutes ago, Audiophile Neuroscience said:

 

Please enlighten us with the correct methodology without referring to "or,or,or,or" and you know, 'all that sci-ency stuff.

I would have to have a good think about that. Obviously it would be necessary to inspect the set up and see whether there were any odd bits of behaviour. It's not rocket science, nor I think is it unusual when looking at an experiment. I don't know why you are making such a fuss about this; it's an obvious line of inquiry.

[PeterSt]

5 minutes ago, mansr said:

 

The noise floor is considerably lower. There are spikes every 400 Hz, which is still not normal.

 

Hmm ... FYI : I don't recognize those spikes either. Should be a Tascam thing ?

All I can make of this is that it is jitter related. You ?

(looking at 10KHz more closed in is not my daily work)

[mansr]

10 minutes ago, PeterSt said:

Hmm ... FYI : I don't recognize those spikes either. Should be a Tascam thing ?

My Tascam (different model) does nothing of the kind.

 

10 minutes ago, PeterSt said:

All I can make of this is that it is jitter related. You ?

I don't think so. The spikes are present throughout the entire spectrum.

[STC]

1 hour ago, PeterSt said:

 

There is still a problem with your reasoning. 

With this statement from you, you suppose that the both playbacks of the same file, the same bits entering the DAC, change when passing some BNC to RCA bridge. Well, fine. There you have it.

Well done, right ? finished. Software with two times bit identical data can change the interpretation by a BNC adapter each occasion from the other.

haha

 

The bits are the same but how the bits reached the recorder or DAC can be different. For an example, A CD drive could take one minute to write a track while a newer one could take about 50 seconds. The timing of the data accumulated at the recorder varies but they should be bit-identical to the original file. Correct me if I am wrong, isn't a DAC receives data in real time stream and the timing of how the data is converted to analogue need to be precise compared to how the data is written to a drive digitally. I am asking since I do not know.

 

 

 

39 minutes ago, mansr said:

Whatever makes you say such a thing? There is no reason whatsoever to believe that the data transfer between PC and DAC/recorder was compromised in any way.

I have been a audiophile for long enough to believe that there should be audible difference with SPDIF vs Toslink Vs BNC vs XLR connection. Are you telling there cannot be any difference here.

 

 

This chart was taken from https://audiosciencereview.com/forum/index.php?threads/understanding-hdmi-and-potential-for-hdmi-cable-differences.104/ 

 

Amir said "This is showing the *analog* output of the AVR DAC. Purple is S/PDIF. Yellow is HDMI. Identical digital data was sent to each input. Yet what came out of the DAC was much more distorted in the case of HDMI."

 

Assuming that the data reached the DAC were identical to what arrived at the Tascam the noise or jitter cannot be conclusively said to be similar before going through the BNC to RCA adapter. I am sure many here would argue that they could hear a difference with BNC and SPDIF itself but that is not the point of contention here. It would be easier to concentrate on the difference alone that Mani heard by eliminating the different cable used for Tascam and DAC. Ideally, both should have used BNC or RCA.

 

[mansr]

2 minutes ago, STC said:

Assuming that the data reached the DAC were identical to what arrived at the Tascam the noise or jitter cannot be conclusively said to be similar before going through the BNC to RCA adapter.

The purpose of the Tascam recorder was to verify that identical data was sent by the PC each time, nothing else.

[PeterSt]

2 minutes ago, mansr said:

The spikes are present throughout the entire spectrum.

 

Ah, so one of them is precisely at 10KHz. So this means the first one is at 400Hz. Strange ...

(so the only thing we see from jitter here is at 10330 and 9645 or so (and that the frequency is not exactly at 10.000 we already know / saw elsewhere - just saying)

 

@manisandher, Do you perhaps recognize anything of your isolation transformer here (the one providing ~240V) ?

... and somewhere at the back of my mind something is recognizable from 400 Hz, but it won't pop up ...

 

[PeterSt]

OK, I like to see ghosts ... Modern TV's can have a refresh rate of 400Hz ...

Mani ?

[PeterSt]

Lots of hits on 400Hz (isolation) transformers ...

[lmitche]

6 hours ago, PeterSt said:

 

Actually this is not correct in the context you (seem to) put it :

 

 

There is no such thing as an "SFS various filter" and SFS is not a filter as such as well. It is a buffer with a virtual "analogue" means of lengths (to set), hence within a predetermined range the number of settings is virtually infinite. It also doesn't work out as a filter. It implies "noise signature" (together will all the other dials).

None of these alter the stream, so all these millions of settings possibilities will produce bit identical streams.

 

This is different from upsampling filters (which are for reconstruction), which do change the streams. But these were left out of the equation (were always set the same). Just like volume was not changed in any event. This is all the most obvious for "us" and therefore is not explicitly mentioned.

Peter,

 

So is the SFS parameter kinda like the thickness dial in a meat slicer?  You get the same amount if meat at the output, but the numbers of times the machine makes the slice (frequency) changes at the two settings?

[mansr]

8 minutes ago, PeterSt said:

OK, I like to see ghosts ... Modern TV's can have a refresh rate of 400Hz ...

Mani ?

There's no TV anywhere near the equipment involved here.

[jabbr]

6 hours ago, esldude said:

Two software settings.   Both put out the same identical bit stream we are told.  Yet sound different we are told.  

Oh! 

 

Should not be so controversial. Of course they could sound different — and they provably do!

 

Forget what I wrote about fiberoptic Ethernet etc. 

 

@PeterSt says it is “noise signature” so end of story — no reason at all why noise transmitted along with digital data cannot affect SQ and here we have evidence that it does!

 

I wouldn’t have bet against @manisandher now that I know what is being tested.

[jabbr]

43 minutes ago, STC said:

Amir said "This is showing the *analog* output of the AVR DAC. Purple is S/PDIF. Yellow is HDMI. Identical digital data was sent to each input. Yet what came out of the DAC was much more distorted in the case of HDMI."

This is classic for close-in phase noise: the base of the peak is widened.

[pkane2001]

30 minutes ago, PeterSt said:

Lots of hits on 400Hz (isolation) transformers ...

 

So what explains the regular interval 100Hz spikes in the older DAC captures? Same cause as 400Hz with NOS1? Strange, especially that the magnitude of the spikes is about the same in both cases.

 

EDIT: interesting that 176.4KHz is exactly 4 times 44.1Khz, isn’t it??? 

[PeterSt]

23 minutes ago, mansr said:

32 minutes ago, PeterSt said:

OK, I like to see ghosts ... Modern TV's can have a refresh rate of 400Hz ...

Mani ?

There's no TV anywhere near the equipment involved here.

 

I hope it was clear I wasn't thinking in the direction of your environment (but it could have).

 

Btw, this is why WiFi is forbidden in the Audio PC. You will see that too (and worse).

[PeterSt]

27 minutes ago, lmitche said:

So is the SFS parameter kinda like the thickness dial in a meat slicer?  You get the same amount if meat at the output, but the numbers of times the machine makes the slice (frequency) changes at the two settings?

 

Nice analogy.

Yes, and I suppose that the less slices the more full your mouth is, the more burping will come from it (side-noise). 

The smaller the slices, the more (profound) your teeth will screech (side noise again but of different flavor).

[PeterSt]

13 minutes ago, jabbr said:

 

@PeterSt says it is “noise signature” so end of story — no reason at all why noise transmitted along with digital data cannot affect SQ and here we have evidence that it does!

 

Jonathan, careful - because what I claim is that the noise emerges in the DAC('s receiver parts).

This is also how isolation (like with Toslink) does not help (and is even worse because of the receiver being worse (implies more noise - although this is a different matter)).

[jabbr]

2 minutes ago, PeterSt said:

 

Jonathan, careful - because what I claim is that the noise emerges in the DAC('s receiver parts).

This is also how isolation (like with Toslink) does not help (and is even worse because of the receiver being worse (implies more noise - although this is a different matter)).

 

No problem, there are numerous ways a digital signal could be identical yet electrical differences affect the DAC receiver.

 

In order to sort that out you’d need a high speed scope, perhaps phase analyzer etc to look at what is happening in the wire. 

[pkane2001]

1 hour ago, STC said:

 

The bits are the same but how the bits reached the recorder or DAC can be different. For an example, A CD drive could take one minute to write a track while a newer one could take about 50 seconds. The timing of the data accumulated at the recorder varies but they should be bit-identical to the original file. Correct me if I am wrong, isn't a DAC receives data in real time stream and the timing of how the data is converted to analogue need to be precise compared to how the data is written to a drive digitally. I am asking since I do not know.

 

 

 

I have been a audiophile for long enough to believe that there should be audible difference with SPDIF vs Toslink Vs BNC vs XLR connection. Are you telling there cannot be any difference here.

 

 

This chart was taken from https://audiosciencereview.com/forum/index.php?threads/understanding-hdmi-and-potential-for-hdmi-cable-differences.104/ 

 

Amir said "This is showing the *analog* output of the AVR DAC. Purple is S/PDIF. Yellow is HDMI. Identical digital data was sent to each input. Yet what came out of the DAC was much more distorted in the case of HDMI."

 

Assuming that the data reached the DAC were identical to what arrived at the Tascam the noise or jitter cannot be conclusively said to be similar before going through the BNC to RCA adapter. I am sure many here would argue that they could hear a difference with BNC and SPDIF itself but that is not the point of contention here. It would be easier to concentrate on the difference alone that Mani heard by eliminating the different cable used for Tascam and DAC. Ideally, both should have used BNC or RCA.

 

 

Jitter can be audible and can distort playback, there's no doubt in my mind that this can happen as I spent years in the 90's measuring, testing, and trying to eliminate jitter from a CD-based system. 

 

In the HDMI vs SPDIF jitter is responsible for the large difference in the width of the 12KHz spike. This is because HDMI does not carry a separate audio clock. The receiver derives the timing for audio from the video signal, which often adds a ton of jitter to the process, so this plot is not surprising at all.

 

 

[STC]

3 minutes ago, pkane2001 said:

The receiver derives the timing for audio from the video signal, which often adds a ton of jitter to the process, so this plot is not surprising at all.

 

How about the Altmann DAC? Can the data after the SFS changes reaching the buffer before the DAC subject to timing difference?

[mansr]

I would have liked to examine the signals entering the DAC chip in order to see the jitter levels there. Unfortunately, I don't think Mani's scope is up to the task and I don't particularly like the idea of hauling mine around on trains.

[PeterSt]

28 minutes ago, pkane2001 said:

So what explains the regular interval 100Hz spikes in the older DAC captures? Same cause as 400Hz with NOS1?

 

100Hz can be easily explained (think how (half) rectifiers may work). But no such thing in the NOS1. There's hardly any spur of 50Hz (or 60Hz for the USA) even at the low end.

 

This is from a random NOS1a/G3 which currently is under test and just received the G3 upgrade :

 

 

Yes, with some effort we may see 400Hz just the same.

But it is way lower and not throughout and as it looks, not mains related :