As I said my use of the word transparent was in relation to jitter which is what we were solely talking about - inaudible & transparent mean the same thing to me - if you have a different meaning for these terms, I'm interested in hearing it.
You are now changing the jitter context in which I mentioned "transparent & are therefore mis-quoting me by removing this context.
Clocks being 'off' by a fixed frequency (fixed 1 Hz) is not normal - thy tend to drift (clock wander or drift) so sometimes < 1Hz, sometimes exactly 1Hz & sometimes > 1Hz. Your algorithm dealing with 1Hz fixed freq offset will obviously not be as effective at other frequencies i.e. clock drift or jitter wander will not be attenuated as you specify above.
Really? Please explain. I believe you are confusing the LF phase changes that Bruno states are introduced by the operation of ASRCs - he is saying that some jitter is changed into these LF phase issues which weren't in the original signal. In other words, one will only see these phase issues in increased width of the base of the main frequency spike of Jtest signals.
Well, all I can tell you is what I stated in my first post about ESS DACs - turning off ASRC is well known to produce better SQ when using a low jitter source. This is a known characteristic of the DAC chip & AFAIK there is nothing Benchmark or anybody can have done externally to change this characteristic of their DACs unless they have changed their ASRC which I have not seen any word of.
None of that addresses the question, or the potential issue.
4Ω is only nominal, it is only a gross approximation. High efficiency is not near as important to a SET as a flat impedance curve.
Your answers imply a lack of knowledge on the subject i find very surprising.
dave
Mark, I agree with most of what you say but I prefer to call it auditory processing & that's one of the recognized phrases you will find in this field of research as well as psychophysics, psychoacoustics, auditory scene analysis, & lots of other terms - some associated with hearing aid research, some with music synthesis, some with computer modelling of auditory perception, etc.
Some of the early questions (1930s, 40s) about auditory perception still don't have a working theory that explains all aspects of their functionality- relative pitch perception, for instance. The two main competing ideas - roughly, one involves the position(s) of the vibration along the length of the basilar membrane & the competing one involves the timing correlation between nerve fibre groups - the place & temporal code theories - neither of which fully deal with all observed phenomena of relative pitch perception.
So even for something that we do all the time when listening to music - perceiving the changing pitch - there is no agreed theory that explains all it's observed characteristics.
Last edited:
Hilarious, high percentage distortion amplification but of course it's all euphonic unlike virtually unmeasurable jitter. Don't bother responding I think I'll drop out of these threads permanently.
Your point being mentioned system very unlikely to be able to reveal differences attributed to DAC's?
The PLL (phase locked loop) is the attempted solution to clock drift so I'm not sure what your question means?
If no attempt was made at synchronising between clocks then clocks would drift apart so much over long playback time that digital samples would be dropped & this would be very noticeable.
If no attempt was made at synchronising between clocks then clocks would drift apart so much over long playback time that digital samples would be dropped & this would be very noticeable.
I think he was saying something similar. This is an overly pedantic, opinionated thread, and most here have agendas, so wading through it all for anything meaningful has become a waste of time
I think it is a good thread.
Just beware that there are "Commercial Vendors large & small hawking their wares" present, and this person (an archetype "tube-whatever rolling" type of guy) believe we should trust hes hearing and hes beliefs in "boutique component" superiority.
I dont.
Otherwise this is one of the more interesting ones lately and was glad seeing it resurrected form the "closed".
//
Just beware that there are "Commercial Vendors large & small hawking their wares" present, and this person (an archetype "tube-whatever rolling" type of guy) believe we should trust hes hearing and hes beliefs in "boutique component" superiority.
I dont.
Otherwise this is one of the more interesting ones lately and was glad seeing it resurrected form the "closed".
//
Last edited:
Interesting comment!
I think you're making it a bit too complicated, though. I agree that A/B tests are probably more sensitive than ABX tests. The blind testing of the Swedish Audio engineering society has already been mentioned in the thread. They have consistently been able to detect coloration in many dacs, by using elaborate a/b testing. They were not able to detect the Benchmark Dac2, however, and as mentioned they couldn't detect any coloration in the cheap Yamaha WXC-50.
I double checked their method. What they do is to send the signal through the DAC under test, and convert it back to digital through the best AD converter they know of. Then they compare this signal to the original signal, listened to through an additional dac +amp/speakers. Is the extra DA/AD conversion audible or not? (it is actually more elaborate than this, but this is the basic idea).
Their complete setup is difficult to attain for others, as it involves software which volume matches and randomizes the A/B tracks, and more. But it shouldn't be so difficult to do a simple single blind A/B testing. Say that you want to test whether two different dacs indeed have sonic characters, in your setup and for your ears. If both of them have volume controls, it's not difficult. You can connect both to a mechanical switch, and connect this switch to your pre or power amp. Level match them as much as much as you can. Enlist an assistant. Ask the assistant to play the same song 12 times, and switch at random between A and B, without you knowing which is which (you can use simple online software to generate a random sequence that the assistant will follow).
Are you able to identify the correct dac in, say, 9 or 10 out of 12 instances? Now this isn't a perfect method. But if you are able to identify the dacs correctly, it would consitute some actual evidence that these differences are real and not imagined, and that you are able to hear them.
Alternatively, you can record the output of the two dacs, volume match it digitally, send it to somebody else, and ask them to make a randomized sequence which you then listen to, and try to identify which is which. This would also make it possible to involve other listeners in the test.
One may also do it as a preference test, which is even closer to real life listening.
Again, this is not impossible to do. We shouldn't move the goal post too much. If you can pick out the correct dac when an assistant switches between them without you knowing which is which, it is already something. And I do mean it when I say that it's an extraordinary claim that the Benchmark Dac3 has a sonic character. According to specs/measurements, it has THD+N which is 110-115 db below the signal, jitter levels which approach zero, etc. If it does the change the signal in any way whatsover, it means that there's something going on here which is fully and completely absent from the measurements. That would be extraordinary. And it would be an important thing to be aware of for all of us who are interested in sound reproduction.
Let it be said that I'm fully aware that blind tests can mask differences and create false negatives. Still, it's the best thing we have for discerning whether very small differences are audible for humans. Those who do it best at the moment, IMHO, are the Swedes. They were unable to detect the DAC2, and now we're talking about the DAC3. So until you or others come up with firm evidence that it has a sonic character, I will maintain that the most likely explanation for any perceived differences between this and other dacs of similar quality is that it's caused by listener bias.
Last edited:
This is not logical. If the volume control is not transparent, it will create a difference in the file. If the assumption is that two files are identical for the human ear to start with, volume controlling them with a non-transparent volume control might make them different. If anything, this will increase the likelihood that two files which in fact are identical get perceived as different.
How many times have such sane words been uttered since the 1940-s?
And yet, measurements in high end audio, even today, do not correlate to sound quality. At all.
Until there is an underlying theory explaining how a 5cm piece of wire, or a 1mm² point of contact influence perceived sound we will continue getting null results.
Quite true. The sore spot in most systems is the active preamp. This is why I don't use one.
My DAC volume control changes the D/A reference voltage. It does not change gain, resistance in the analog path or impedance. It does not decimate the digital words. Very transparent.
Possible, but not happening in my case.
Steve N.
Empirical Audio