One prophecy fulfilled: 
As was clearly pointed out in that unfortunate thread, signal phase is NOT a reason for two files sounding different, and those files are NOT an example of humans hearing sensitivity to the sound phase.
Primary reason is that first file contains amplitude modulated signal and our hearing is very sensitive to that. Other file contains phase modulated signal and our hearing is mildly sensitive to the phase modulation. Of course, they sound different.
As was clearly pointed out in that unfortunate thread, signal phase is NOT a reason for two files sounding different, and those files are NOT an example of humans hearing sensitivity to the sound phase.
Primary reason is that first file contains amplitude modulated signal and our hearing is very sensitive to that. Other file contains phase modulated signal and our hearing is mildly sensitive to the phase modulation. Of course, they sound different.
Well, you are the epitome of all kinds of imaginary audio distortion myth spinning.
Asserting intangible matters being tangible, instilling insignificant matters being significant, repetitive hacking on FFT and other veritable measurement techniques for not being able to pick up your imaginary nonsense ad nauseam yet never providing any novel concept or measure on how to quantify it so as to keep the myth alive is all you do, virtually every day, year in year out, it's all a circular augmentation crap show about... nothing.
Oh yes I do!
And I understand trivial nitpicking too.
They both are woven from the same cloth.
You a Fourier denier? If not, then the time-domain and frequency-domain views are equivalent ways of looking at the exact same signal. The only problem we are talking about here is if you throw away phase information and then believe you still have all the information that matters.
EDIT: If this is hard to understand, maybe it will help if I say that phase and amplitude (e.g. crest factor) in complex waveforms are intimately related. If you change phase then you may change amplitude (depending on the particular phase change). They are describing the same exact same signal. This is very easy to graphically if you have a sine wave and its 3rd harmonic. If you change the phase of the harmonic, the crest factor and thus the amplitude changes. If you hear any difference then you are hearing both amplitude and phase. The only difference is whether you describe the signal in the time domain or the frequency domain. Its the same signal either way.
Last edited:
It certainly looks like you are desperately trying to turn this place into a cult of imaginary audio nonsense.To quote a respectable member of this forum, "I don't do cults."
In case anyone is feeling confused about the point I'm trying make, maybe I'm not the problem. Just in case though, please let me try to explain once again what everyone here should probably already know and understand.
In the time domain, the independent variable is time (seen in and x direction on a graph). A signal as viewed in the time domain is a real-valued function of time (e.g. amplitude as a function of time).
In the frequency domain, the independent variable is frequency. It is a complex-valued function with a real and imaginary part. When we talk about the frequency domain we sometimes use the idea of phasors. They are graphed in a way that looks like a vector, but they are interpreted differently, as I will return to shortly. A phasor is graphed as point in a cartesian plane (x/y plane). The horizontal axis is the cosine amplitude. The vertical axis is the sine amplitude. There is one phasor point for each frequency.
Okay, I just described the phasor point in rectangular coordinates. The point on the graph that represents the phasor can also be described in polar coordinates. In that case its properties are in terms of magnitude and phase. The phase is relative phase compared to some reference (e.g. a fundamental frequency).
What phase is not, is it is not a property of the time domain. It is a property of frequency domain as seen in the polar representation of a phasor.
It seems like some people have been conflating the time domain and the frequency domain as if they are all the same. Its like some people seem to think phase is a property of time. Not exactly, the time domain view is amplitude as function of time. Talking about phase takes us back to the frequency domain where it has a defined meaning.
The transformation of rectangular to polar coordinates is high school level math, so there should be no confusion that both coordinate systems can describe the same point.
There should also be no confusion that changing phase in the frequency domain can sometimes produce what looks like amplitude modulation or frequency modulation. Both types of waveforms can be viewed as periodic signals that can be represented the frequency domain or in the time domain. Just don't conflate the two.
In the case of the example waveforms at the Purifi website, the two waveforms when viewed in the frequency domain differ only in the phase of some of frequencies. The frequencies and their amplitudes are exactly the same for both signals. Again, only phase differs.
Now, when we look at the signals in the time domain our intuition may suggest that something more than phase has changed. However, the time domain only has amplitude not phase. Again, when we use the term phase we are implicitly referring the frequency domain view of the signals.
Hopefully this is starting to make more sense. Nobody is making anything up or trying to fool you. Maybe you are confused because you don't remember some of what you learned a long time ago in school.
Another factor that can be confusing is that a lot of engineers and authors seem to repeat the myth that humans are insensitive to phase. That was the finding of early psychoacoustics research which used simple waveforms and or waveforms in which phase didn't happen to have audible effects. That was extrapolated to human hearing in general based on the mistaken assumption that the human auditory system is linear, time-invariant, and or stationary. If that were true, we could extend the findings to all waveforms of any complexity. We now know from more recent research that human auditory system (ear/brain) is much more complex than once assumed. It is non-linear, non-time invariant, and non stationary. Links have been posted before to relevant literature before in case anyone has forgotten.
Okay, then. If there are serious questions I will be happy to respond to PMs. Otherwise, there is probably no further point in talking about this stuff at this time.
In the time domain, the independent variable is time (seen in and x direction on a graph). A signal as viewed in the time domain is a real-valued function of time (e.g. amplitude as a function of time).
In the frequency domain, the independent variable is frequency. It is a complex-valued function with a real and imaginary part. When we talk about the frequency domain we sometimes use the idea of phasors. They are graphed in a way that looks like a vector, but they are interpreted differently, as I will return to shortly. A phasor is graphed as point in a cartesian plane (x/y plane). The horizontal axis is the cosine amplitude. The vertical axis is the sine amplitude. There is one phasor point for each frequency.
Okay, I just described the phasor point in rectangular coordinates. The point on the graph that represents the phasor can also be described in polar coordinates. In that case its properties are in terms of magnitude and phase. The phase is relative phase compared to some reference (e.g. a fundamental frequency).
What phase is not, is it is not a property of the time domain. It is a property of frequency domain as seen in the polar representation of a phasor.
It seems like some people have been conflating the time domain and the frequency domain as if they are all the same. Its like some people seem to think phase is a property of time. Not exactly, the time domain view is amplitude as function of time. Talking about phase takes us back to the frequency domain where it has a defined meaning.
The transformation of rectangular to polar coordinates is high school level math, so there should be no confusion that both coordinate systems can describe the same point.
There should also be no confusion that changing phase in the frequency domain can sometimes produce what looks like amplitude modulation or frequency modulation. Both types of waveforms can be viewed as periodic signals that can be represented the frequency domain or in the time domain. Just don't conflate the two.
In the case of the example waveforms at the Purifi website, the two waveforms when viewed in the frequency domain differ only in the phase of some of frequencies. The frequencies and their amplitudes are exactly the same for both signals. Again, only phase differs.
Now, when we look at the signals in the time domain our intuition may suggest that something more than phase has changed. However, the time domain only has amplitude not phase. Again, when we use the term phase we are implicitly referring the frequency domain view of the signals.
Hopefully this is starting to make more sense. Nobody is making anything up or trying to fool you. Maybe you are confused because you don't remember some of what you learned a long time ago in school.
Another factor that can be confusing is that a lot of engineers and authors seem to repeat the myth that humans are insensitive to phase. That was the finding of early psychoacoustics research which used simple waveforms and or waveforms in which phase didn't happen to have audible effects. That was extrapolated to human hearing in general based on the mistaken assumption that the human auditory system is linear, time-invariant, and or stationary. If that were true, we could extend the findings to all waveforms of any complexity. We now know from more recent research that human auditory system (ear/brain) is much more complex than once assumed. It is non-linear, non-time invariant, and non stationary. Links have been posted before to relevant literature before in case anyone has forgotten.
Okay, then. If there are serious questions I will be happy to respond to PMs. Otherwise, there is probably no further point in talking about this stuff at this time.
I just can't understand why somebody would waste so much time on that silly marketing stuff. With all that wasted effort you probably could have solved the dreaded intermittent packing flooding issue in your dac.
As already said, don't stream over ethernet so I don't have a problem with it here. Do however have some experience with how RFI/EMI can affect dac sound, usually subtly, yet sometimes audible to an attentive listener in A/B testing.
Regarding listening tests, I have blind tested people but don't normally have enough people on hand to double blind. Even using double blind, blinding efforts have been known to fail anyway, such as medical research (more details available if anyone cares). Occasionally DBT occurs here by accident such as when Windows suddenly starts messing with my audio stream. Nobody hinted to me that something about the sound changed, but I could hear it anyway. Similarly PMA's listening test was effectively DBT. Anyway, using available practical methods to help with listening test reliability as best we can, IME very small noise problems can affect sound in a way that most people don't even realize isn't quite as clean as it should be. Other people may notice some difference, even if very small.
Therefore, as a matter of policy I don't automatically reject claims that based on experience with EMI/RFI noise seem potentially plausible. That of course doesn't mean there is any confidence that the product works as claimed. Not is there is confidence the product can't work. IMHO a wait and see attitude (deferral of final judgment) is not unreasonable.
(1) Isolators are really attenuators which vary in effectiveness with frequency.
(2) Packet congestion (or whatever one want's to call it) might be another factor. I asked a high end audio designer about those ethernet switches and he said he knows people who use them. He said part of what they do is reclock the packets so they arrive more evenly in time. I asked if it made any difference in sound quality, to which he replied, "I don't know." This is a guy who tells it like it is in his experience. He as told me about people who buy expensive audio gear they don't need, even if it makes their system sound worse. What he didn't say in this case is that the L2 switches can't possibly work. IOW his point of view on this is similar to mine. If the boxes do have an effect on sound, then maybe we need to look deeper to understand why.
(2) Packet congestion (or whatever one want's to call it) might be another factor. I asked a high end audio designer about those ethernet switches and he said he knows people who use them. He said part of what they do is reclock the packets so they arrive more evenly in time. I asked if it made any difference in sound quality, to which he replied, "I don't know." This is a guy who tells it like it is in his experience. He as told me about people who buy expensive audio gear they don't need, even if it makes their system sound worse. What he didn't say in this case is that the L2 switches can't possibly work. IOW his point of view on this is similar to mine. If the boxes do have an effect on sound, then maybe we need to look deeper to understand why.
1) what do you mean by isolator?
2) if you have various ethernet appliances what are you going to plug them into other than a switch? The expensive switch will have the same chip on it as the $15 one from best buy. so there can be no difference in the packet switching. The timing doesn't matter as the packets are received and go thought a microprocessor before being clocked out to the DAC. In the case of our crystal healer turned reviewer he goes from the switch to a cheap fibre converter, back to ethernet into an uptone regen and then into the DAC. And you are really willing to give the nordost switch the benefit of the doubt?
2) if you have various ethernet appliances what are you going to plug them into other than a switch? The expensive switch will have the same chip on it as the $15 one from best buy. so there can be no difference in the packet switching. The timing doesn't matter as the packets are received and go thought a microprocessor before being clocked out to the DAC. In the case of our crystal healer turned reviewer he goes from the switch to a cheap fibre converter, back to ethernet into an uptone regen and then into the DAC. And you are really willing to give the nordost switch the benefit of the doubt?
Isolators like discussed at:
https://electronics.stackexchange.c...re-ethernet-rj45-sockets-magnetically-coupled
https://networkengineering.stackexc...et-magnetic-transformer-and-how-are-they-used
Some discussion about possibly faulty network equipment:
https://electronics.stackexchange.com/questions/53388/ground-loop-problem-with-power-over-ethernet
"...looks like TP-Link are cutting corners with their PoE devices and omitting isolation circuitry to reduce cost, violating the standards."
Still feel I don't know of all possible problems that could potentially occur without a specific case to look at where a L2 switch does make an audible difference.
https://electronics.stackexchange.c...re-ethernet-rj45-sockets-magnetically-coupled
https://networkengineering.stackexc...et-magnetic-transformer-and-how-are-they-used
Some discussion about possibly faulty network equipment:
https://electronics.stackexchange.com/questions/53388/ground-loop-problem-with-power-over-ethernet
"...looks like TP-Link are cutting corners with their PoE devices and omitting isolation circuitry to reduce cost, violating the standards."
Still feel I don't know of all possible problems that could potentially occur without a specific case to look at where a L2 switch does make an audible difference.
My point is that there could be pathologically faulty network equipment involved in a case where switch helps. You saying that's impossible?
If it could happen, and the sound is affected, and if an expensive L2 switch bandaids over the problem, then maybe we have an explanation? Just saying, if the sound is affected, then we may need to look deeper for an explanation. Whatever we might find may only make sense after we find it.
Bill, I have seen stuff like that from a big, reputable manufacturer where some engineering problem remained and the engineers were in denial that it could happen. Have seen it more than once from more than one manufacturer. Not saying its common though.
If it could happen, and the sound is affected, and if an expensive L2 switch bandaids over the problem, then maybe we have an explanation? Just saying, if the sound is affected, then we may need to look deeper for an explanation. Whatever we might find may only make sense after we find it.
Bill, I have seen stuff like that from a big, reputable manufacturer where some engineering problem remained and the engineers were in denial that it could happen. Have seen it more than once from more than one manufacturer. Not saying its common though.
Last edited:
Just jumping in here, but I really don't see any way a local switch, at any price point, is going to make a difference - it just basically receives a packet of data and sends it back out. TCP/IP is not designed to be seamless in the time domain, it is designed to be accurate and ensure that the data gets through. As such, packets are not even guaranteed to be received in the order in which they are sent so delays or losses from being bounced around from one satellite to another are common. The receiving equipment it responsible for assembling the data back into the correct format and order - and for audio, I would assume that a certain minimum buffer length would be required to ensure this happens. None of which is affected by the local switch - it has no way of knowing the packet content type in a way that matters to the audio integrity; even a cheapo D-Link from China-mart has more than enough processing capability to not be a source of significant slow downs for any likely home network and should not contribute to any delay issues.
Hal
Hal
Last edited:
Well, that is certainly possible - but just replacing it with something at a more normal cost point would work just as well.
- Status
- Not open for further replies.