Hi JK2,
Yes, this card shows a much better dynamic range and the software shows a much resolution, so suddenly what is revelead? Measurable jitter, not e lot, but not good enough for the general analogue resolution of the system. And this being loopback mode (rather than using two seperate EMU Audio Systems - which could be done - I instead used EMU to AP2) we probably still are masking some jitter as well. That and the gretaer resolution is why I prefer the AP2 for testing.
However, I would say this whole setup is around 20 times better to use as test gear than what some others have proposed here, to we have a 20 times better level of confidence that we can measure something that is really present.
You would not perchance also have a Behringer DEQ or DCX at hand to to use via SPDIF out from the 1616m and then back through the 1616m inputs, just to illustrate exactly how impeccable it's analogue outputs are?
Ciao T
PS, Sy, in case you wonder, what JK2 shows is a J-Test...
Yes, this card shows a much better dynamic range and the software shows a much resolution, so suddenly what is revelead? Measurable jitter, not e lot, but not good enough for the general analogue resolution of the system. And this being loopback mode (rather than using two seperate EMU Audio Systems - which could be done - I instead used EMU to AP2) we probably still are masking some jitter as well. That and the gretaer resolution is why I prefer the AP2 for testing.
However, I would say this whole setup is around 20 times better to use as test gear than what some others have proposed here, to we have a 20 times better level of confidence that we can measure something that is really present.
You would not perchance also have a Behringer DEQ or DCX at hand to to use via SPDIF out from the 1616m and then back through the 1616m inputs, just to illustrate exactly how impeccable it's analogue outputs are?
Ciao T
PS, Sy, in case you wonder, what JK2 shows is a J-Test...
Ah, thank you. I thought he was actually getting spectra like that out of actual test signals and was wondering what kind of crap setup he was measuring. That method is quite relevant for measuring the sensitivity of a DAC to jitter. Of course, that's not what I'm trying to measure, but it's interesting nonetheless- I have read some of Dunn's papers.
No. The *only* important parameter is what comes out of the analog output. That's where the music is. George's stuff is very interesting, but it's irrelevant to the question of the changes in the analog signal output with differing USB to spdif converters, which is what I'm looking at. The analog output is the bottom line of the music storage and delivery- for the nth time, if that doesn't change, the sound presumably (in the absence of any good, controlled sensory evidence) doesn't change.
I take it by your focus on the Behringer that you'd rather me use that than the Echo?
I take it by your focus on the Behringer that you'd rather me use that than the Echo?
You continually reject the possibility that your DAC (DCX or Echo) may not be up to the job at hand. You put up some loopback tests to prove it had which George advised was of not relevant as the ADC & DAC will self cancel any inherent jitter in the system. You claim the DAC has no inherent jitter as it would show up as spurs on the analogue out plot but you have now been told that the loopback test will not show this. Why do you avoid proper testing of your DACs? Have you done a cross-linked configuration as George suggested? Can you post the results?
Use whatever DAC that can be shown to be lower jitter than the unit under test - is this not the normal approach - the measuring device needs to be of a higher spec than the DUT?
Last edited:
Hi,
Yes, of course. You may damage you long held beliefs in the correctness of the simplified and crippled electronics theory thought at engineering school and so beloved by certain objectivists in this place. It may expand your mind in unknown ways and to unknown degrees.
Or you may decide that you acquired these beliefs and prejudices too dearly and it would require by far too much re-arranging of mental furniture to allow this new information to be accepted and integrated and you may turn into a bitter negativist that feels the need to attack and ridicule anything out of line with his (already invalidated) worldview, not because you suspect anyone is pulling a fast one, but because you can of course not allow to be real what cannot be real, even though you already know it is real.
Or you may feel that the whole world is a surrealist experiment and you may join Csicon (the Committee for the Surrealist Investigation of Claims Of the Normal).
So such a test is extremely dangerous and should be attempted only in the right frame of mind and never by degreed engineers and any guardians of academic orthodoxy, unless they are sure they have right open frame of mind.
Remember, all our perception are filtered through our mind and our unconscious. So we never actually percieve reality, but instead a carefully edited and filtered version.
Can you see the face in this image of african tribesmen?
Salvador Dahli saw it. I have as well.
Ciao T
PS.
The Committee for the Surrealist Investigation of Claims Of the Normal invite you to join them
If your I.Q. is over 150, and you have $3,125.00 (including handling charge), you might be eligible for a trial membership in CSICON. If you think you qualify, put the money in a cigar box in front of all your neighbours and bury it in your frontyard in broad daylight (not the backyard secretly at night, that is the place and way to bury it if you want to join the Illuminati). Following one of our Underground Agents will contact you shortly.
Alternatively publicly burning Federal Reserve Notes in the amount of at least $ 3,000 (aka instant demurage) may make you eligible if carried out on national TV, while wearing a funny costume. Other pieces of surrealist performance art (such as shaving in public, showering in a public fountain or living seven days in the window of an art-gallery may also qualify.
Yes, of course. You may damage you long held beliefs in the correctness of the simplified and crippled electronics theory thought at engineering school and so beloved by certain objectivists in this place. It may expand your mind in unknown ways and to unknown degrees.
Or you may decide that you acquired these beliefs and prejudices too dearly and it would require by far too much re-arranging of mental furniture to allow this new information to be accepted and integrated and you may turn into a bitter negativist that feels the need to attack and ridicule anything out of line with his (already invalidated) worldview, not because you suspect anyone is pulling a fast one, but because you can of course not allow to be real what cannot be real, even though you already know it is real.
Or you may feel that the whole world is a surrealist experiment and you may join Csicon (the Committee for the Surrealist Investigation of Claims Of the Normal).
So such a test is extremely dangerous and should be attempted only in the right frame of mind and never by degreed engineers and any guardians of academic orthodoxy, unless they are sure they have right open frame of mind.
Remember, all our perception are filtered through our mind and our unconscious. So we never actually percieve reality, but instead a carefully edited and filtered version.
Can you see the face in this image of african tribesmen?
Salvador Dahli saw it. I have as well.
Ciao T
PS.
The Committee for the Surrealist Investigation of Claims Of the Normal invite you to join them
If your I.Q. is over 150, and you have $3,125.00 (including handling charge), you might be eligible for a trial membership in CSICON. If you think you qualify, put the money in a cigar box in front of all your neighbours and bury it in your frontyard in broad daylight (not the backyard secretly at night, that is the place and way to bury it if you want to join the Illuminati). Following one of our Underground Agents will contact you shortly.
Alternatively publicly burning Federal Reserve Notes in the amount of at least $ 3,000 (aka instant demurage) may make you eligible if carried out on national TV, while wearing a funny costume. Other pieces of surrealist performance art (such as shaving in public, showering in a public fountain or living seven days in the window of an art-gallery may also qualify.
Have you had any experience yet in using diffmaker in looking for changes? I'd be interested if you had - the difficulties are various, but perhaps the ones I found were my sloppy understanding. I found it hard to get a deep enough null from it. Its rather like the Hafler 'subtract the input from the scaled output' test - its jolly hard to get nothing and then, with what's left, what really matters? Both linear (freq response) and non-linear effects are treated in the same way and I think the linear effects are in the main benign, given that speakers are generally way worse in this department than dacs. Yet in my experience, even ultra-cheap speakers unerringly reveal flaws in the electronics.
In practice its a lot worse than that (worse that is from an objectivist pov) in that the unconscious creates the objects of perception, it does not filter or edit pre-existing objects. Are you up to speed on constructivist epistemology T?
Sy,
I would suggest that it is by far more than that, it is an excellent tool (if combined with an analyser having high enough analogue resolution) to test the jitter sensitivity of the whole system under test.
I know that, as doing a J-Test measurement on the output with a test system having sufficient resolution (AP2, EMU1616m-mod with suitabe software, etc.) would of course show material differences that would take almost any action to avoid illustrating as being in existence.
Maybe JK can send one of his modded units and one stock to JA at stereophile who has actually got sufficiently sensitive test gear as well as experience in making meaningful measurements, instead of bothering sending it to you for testing that clearly seems to involve an agenda that has nothing to do with the real state of affairs?
Here an example of how the topic may be approached, in this case showing three different USB/SPDIF converters (one from an earlier test for reference):
Stereophile: Lindemann & Stello USB-S/PDIF converters
Of interest may be that both the Stelo converter in the test and the Lindeman one use the same silicon, one I estimate as being capable to attain around 150 - 200ps pp jitter with a PC Source set up for low jitter (as this chipset is not asynchronous).
Despite being virtually identical in circuit execution (there are some different components and the layout differs) we see material differences in the DAC's analogue output.
So Sy, here you can see the demonstration of how to test if you want find out what the differences are.
Ciao T
I would suggest that it is by far more than that, it is an excellent tool (if combined with an analyser having high enough analogue resolution) to test the jitter sensitivity of the whole system under test.
I know that, as doing a J-Test measurement on the output with a test system having sufficient resolution (AP2, EMU1616m-mod with suitabe software, etc.) would of course show material differences that would take almost any action to avoid illustrating as being in existence.
Maybe JK can send one of his modded units and one stock to JA at stereophile who has actually got sufficiently sensitive test gear as well as experience in making meaningful measurements, instead of bothering sending it to you for testing that clearly seems to involve an agenda that has nothing to do with the real state of affairs?
Here an example of how the topic may be approached, in this case showing three different USB/SPDIF converters (one from an earlier test for reference):
Stereophile: Lindemann & Stello USB-S/PDIF converters
Of interest may be that both the Stelo converter in the test and the Lindeman one use the same silicon, one I estimate as being capable to attain around 150 - 200ps pp jitter with a PC Source set up for low jitter (as this chipset is not asynchronous).
Despite being virtually identical in circuit execution (there are some different components and the layout differs) we see material differences in the DAC's analogue output.
So Sy, here you can see the demonstration of how to test if you want find out what the differences are.
Ciao T
Those Stereophile plots show the peak widening and spurious harmonics on the devices with jitter. Isn't that what we are looking for? Shouldn't a better spdif interface be cleaner?
I'll do an asynchronous DAC and ADC tomorrow, just for fun.
I'll do an asynchronous DAC and ADC tomorrow, just for fun.
Now I'm kind of happy that I haven't been following this thread in real time. But I just read the last few dozen posts, and then the first hundred or so. And I had a recurring "crazy" idea. Maybe there's a simpler explanation for the perceptible audio effects: Maybe the attenuation of the input level, alone, is responsible. It was, after all, originaly stated that the SPDIF signal level would need to be higher than normal, I think. Maybe not a coincidence. And intermodulation was also mentioned as a possibility, for the cause of the types of noticed effects. Maybe a clue.
Does anyone else recall the technique required for measuring harmonics with a spectrum analyzer? You have to attenuate the input more and more until the ratios of the harmonics to the fundamental stop changing. Why? Because the mixer (a multiplier, i.e. basically an intermodulator) that is built into all spec ans, just after the input attenuator, generates harmonics, just by its nature (even the digitally-implemented ones), but especially for input levels that are "too high". So you have to attenuate the input level until you're sure that the harmonics you're seeing are the real ones from the input, and don't have significant contributions from the mixer-generated harmonics.
That's just an example of the concept. But lots of types of non-linearities cause harmonics and intermods to be generated. But, also, with smaller variations around some operating point they look more linear and behave better, which should lower the severity of the generated intermods and harmonics. So attenuating the signal amplitude at least theoretically "could" improve such things, in some cases.
And (this might be a long shot, but who knows?) if there's also some non-linear threshold-type mechanism involved, or the non-linearity just gets too severe outside of some range, you might be able to get a perceivable qualitative difference relatively easily, by attenuating the signal.
Sorry. Ideas are not very well developed and I don't have any idea by what mechanism they might apply, in the case being discussed. Just a "guess".
Edit: Did I understand correctly that jkeny owns "minicircuits"? Wow. That would be extremely impressive. We love the minicircuits minicircuits.
Does anyone else recall the technique required for measuring harmonics with a spectrum analyzer? You have to attenuate the input more and more until the ratios of the harmonics to the fundamental stop changing. Why? Because the mixer (a multiplier, i.e. basically an intermodulator) that is built into all spec ans, just after the input attenuator, generates harmonics, just by its nature (even the digitally-implemented ones), but especially for input levels that are "too high". So you have to attenuate the input level until you're sure that the harmonics you're seeing are the real ones from the input, and don't have significant contributions from the mixer-generated harmonics.
That's just an example of the concept. But lots of types of non-linearities cause harmonics and intermods to be generated. But, also, with smaller variations around some operating point they look more linear and behave better, which should lower the severity of the generated intermods and harmonics. So attenuating the signal amplitude at least theoretically "could" improve such things, in some cases.
And (this might be a long shot, but who knows?) if there's also some non-linear threshold-type mechanism involved, or the non-linearity just gets too severe outside of some range, you might be able to get a perceivable qualitative difference relatively easily, by attenuating the signal.
Sorry. Ideas are not very well developed and I don't have any idea by what mechanism they might apply, in the case being discussed. Just a "guess".
Edit: Did I understand correctly that jkeny owns "minicircuits"? Wow. That would be extremely impressive. We love the minicircuits minicircuits.
Last edited:
That was my point. A whole list of stark varying DACs; in what way were this 'proven' DACs as jkeny said? OK later he said that for some DACs the effect wasn't there; which ones? How controlled were those tests anyway, better, worse or same as those Stuart is going to do? There's a lot of pressure on Stuart do 'use the best there is', 'do it well controlled'; how about those other 'test' that showed pos effects?
jan didden
Hi Jan,
First, at least from me there is no pressure to use anything at all, to make test at all. What I am demanding is that Sy, who is the one who usually slams others for sloppy testing, making unsupported claims etc. et. al. hold himself to equal if not higher standards if anything he does is to be taken serious.
As he so obviously does not intend to any such thing I have decided to treat this all as an elaborate joke on his part, in the best marxist tradition, but sadly at the expense of JKenny.
I think that given Sy has a long history of extreme bias against anything that does not fit completely and directly in his view how things should be (and this extends of course to such things as the effect of attenuating a SPDIF signal fall seriously into this sort), it is not unreasonable to require a demonstration that the tests he will conduct will represent best practice and will be able to provide relevant information (e.g. are sensitive enough and appropriate) and that they will be conducted fairly.
For now my view is that the results of both measurements and listening are already predetermined, not by the differences or not present in the devices under test, but by selecting irrelevant test methodologies, using sub-standard test equipment and excessive expectation bias in any listening test.
And it is up to Sy to demonstrate that in fact instead his tests are appropriate, his test gear is sufficiently capable and his listening tests will be conducted so as remove know biases and be calibrated so as demonstrate their ability to distinguish known audible phenomenae. And I think Sy has amply demonstrated in this thread, he has no intention or desire to anything like that.
JKenny may be accused of too much gentle naivety in actually offering his device for testing that will be so obviously aimed at returning a "no different" result, but that is his problem and he will no doubt in the future look at this affair as "another round for experience" and be less eager to offer his stuff to biased tests.
Ciao T
First, at least from me there is no pressure to use anything at all, to make test at all. What I am demanding is that Sy, who is the one who usually slams others for sloppy testing, making unsupported claims etc. et. al. hold himself to equal if not higher standards if anything he does is to be taken serious.
As he so obviously does not intend to any such thing I have decided to treat this all as an elaborate joke on his part, in the best marxist tradition, but sadly at the expense of JKenny.
I think that given Sy has a long history of extreme bias against anything that does not fit completely and directly in his view how things should be (and this extends of course to such things as the effect of attenuating a SPDIF signal fall seriously into this sort), it is not unreasonable to require a demonstration that the tests he will conduct will represent best practice and will be able to provide relevant information (e.g. are sensitive enough and appropriate) and that they will be conducted fairly.
For now my view is that the results of both measurements and listening are already predetermined, not by the differences or not present in the devices under test, but by selecting irrelevant test methodologies, using sub-standard test equipment and excessive expectation bias in any listening test.
And it is up to Sy to demonstrate that in fact instead his tests are appropriate, his test gear is sufficiently capable and his listening tests will be conducted so as remove know biases and be calibrated so as demonstrate their ability to distinguish known audible phenomenae. And I think Sy has amply demonstrated in this thread, he has no intention or desire to anything like that.
JKenny may be accused of too much gentle naivety in actually offering his device for testing that will be so obviously aimed at returning a "no different" result, but that is his problem and he will no doubt in the future look at this affair as "another round for experience" and be less eager to offer his stuff to biased tests.
Ciao T
Why not prove first that an introduction of an attenuator effects the (a) PLL circuits control voltage. This would be a much simpler test to conduct and when we have seen the result, let the discussion explode into how a PLL effects an A/D and eventually the sonic signature. Taking the analogue space into the test setup is a huge challenge and introduces a lot of pitfalls. SY - go and see the music performance if it is not too late - we will wait!
/
/
I listed DACs which were used in the past & anecdotally worked in the hope that Sy might consider it a useful list to try select from (I also offered to provide others). Instead he insists on using what he has available without any attempt to prove the equipment is capable. The whole point of the exercise was that my listening results were called into question as not relevant & his measurement tests were going to sort out the answer. It's admirable that you are trying to protect SY from the the harsh realities of proper testing technique but he has put himself forth as the yardstick that will prove or disprove the anecdotal claims. It is up to him to show that his results will be any more relevant than the anecdotal evidence.
I have said as much in the past but SY has the mantra "if it can be heard it will be seen on the analogue out waveform" - maybe you should talk to SY about the problems he will encounter in trying to live up to this or do you think this a fool's errand - as you seem to imply?TNT said:
OK then I admit it, I own Minicircuits but I needed to send my device to SY for "proper" testing as I can't afford the cost of a $35,000 analyser like that Joseph K fellow has 😀gotee said:
SY, I look forward to your posts firstly the SPDIF waveform with/without attenuators that you said you had already done & you said showed nothing like the sort of effect seen in the shots posted here already.
Secondly you said that you were doing a comparison against a $20 Chinese USB DAC - I look forward to seeing these comparative results
Thirdly the with/without attenuator shots should be interesting & of course your listening results!
Last edited:
Jan, the "pressure" is to do a bunch of tests that are irrelevant to the task at hand and to magically materialize different commercial DAC units. I have no interest in that- the question still remains, "Is there a difference in the analog output signal of a DAC between an expensive (relatively) USB/spdif converter and a cheap one?"
No analog output difference, no sonic difference, unless you want to invoke Booga-Booga.
No analog output difference, no sonic difference, unless you want to invoke Booga-Booga.
I certainly also think that proving a fact in the analogue domain is of course the ideal but why not chew the elephant in pieces?
/
/
SY, have you redefined the orignal question... ( question still remains, "Is there a difference in the analog output signal between an expensive (relatively) USB/spdif converter and a cheap one)
I thougth this was about proving that an attenuator reduces jitter and thus, improves the "sound"?
Hence the topic title?
I thougth this was about proving that an attenuator reduces jitter and thus, improves the "sound"?
Hence the topic title?
- Status
- Not open for further replies.