Please understand I am trying to explain the ideal case where the dac is designed to accept data from the USB board but strongly attenuate any USB noise effects that may come with the data. Its doable for a diy dac, but usually out of the buyer's control with a commercial dac.
i agree here, tho even in a "perfect dac implementation world" i would be curious if this usb cable still does something, even if the difference is way smaller than on cheaper implementations
i have to say, im not doing it because i think this "THE usb fix" but it may be very well a improvement with many dacs which atleast should be explored 🙂 specially the filtering on each cable end made me curious since its not often seen, actually the only time i seen it was in these neutrik emf filtering connectors and i think they did this for a reason, not because its fixable inside a device but additional filtering might just help "more", specially with bad implementations but where we draw the line between good and bad here?
no i get it, you are probably right but i havent had much expierence with diy dacs yet
^That type of filter can be good when the main problem is USB +5v power rail noise relative to USB ground.
However, there can also be a problem with USB power ground noise forming a ground loop once the dac is connected to the AC power ground. Even if there is not a ground loop in the sense of a wiring path back through the AC line ground, there can still be a problem where the USB ground acts like a noise antenna and radiates noise into dac or other downstream circuitry.
So, the overall best solution is wideband galvanic isolation in the dac, with care to also avoid radiated EMI/RFI coupling. As said before, it can all be done pretty well in the diy case. However, a USB isolator by itself may or may not be better. Turns out Topping makes a USB isolator product. My sad experience with that one is described in a post at: https://www.diyaudio.com/community/threads/problematic-usb-isolator-unit.399247/post-7356655
However, there can also be a problem with USB power ground noise forming a ground loop once the dac is connected to the AC power ground. Even if there is not a ground loop in the sense of a wiring path back through the AC line ground, there can still be a problem where the USB ground acts like a noise antenna and radiates noise into dac or other downstream circuitry.
So, the overall best solution is wideband galvanic isolation in the dac, with care to also avoid radiated EMI/RFI coupling. As said before, it can all be done pretty well in the diy case. However, a USB isolator by itself may or may not be better. Turns out Topping makes a USB isolator product. My sad experience with that one is described in a post at: https://www.diyaudio.com/community/threads/problematic-usb-isolator-unit.399247/post-7356655
had a pretty similar expierence with the HS01 unfortunaly...
imo galvanic isolation is the way to go, the question is if additional filtering afterwards can be beneficial, in the case of the HS01 i pretty much think so, might not be the case for all isolation circuits tho
From what I have seen so far some people don't seem to know how to take measurements and so have to resort to unsubstantiated claims based on purely sighted subjective listening.
which measurements do you wanna see ? im pretty sure this filter "does something" as the datasheet/common sense suggests, exactly, we end up arguing wether its audible either case, just move along please...
a measurement will prove if there is noise or not , and if there is some it will show what kind of and how much
this is normaly the very first step when you want to make a filter , or I miss something !!
.
What if the noise is no longer separate from the audio signal, what if some noise has become part of the audio signal (if no audio signal, then no noise)? How do you measure how much of an audio signal is correlated-noise?
What if noise comes in occasional high-intensity bursts (like frying or popping), but it looks like low level white noise on an FFT?
What if the noise is only at RF frequencies, then how do you relate that to SQ?
The point is that noise comes in all guises, and sometimes its not known, or there is no standard way, to make sense out of whatever can be measured relative to what may be heard.
Even Bill Whitlock doesn't seem know how to measure all the noise effects he has found in audio equipment. He said the evidence of a noise problem can merely be the presence of, “Veiled" or "grainy" quality in audio. How do you measure veiled and or grainy?
And BTW, Whitlock did not publish any controlled listening test proof of his claim in his document I quoted from. Nor did he publish such proof anywhere else I am aware of.
Quote taken from:
What if noise comes in occasional high-intensity bursts (like frying or popping), but it looks like low level white noise on an FFT?
What if the noise is only at RF frequencies, then how do you relate that to SQ?
The point is that noise comes in all guises, and sometimes its not known, or there is no standard way, to make sense out of whatever can be measured relative to what may be heard.
Even Bill Whitlock doesn't seem know how to measure all the noise effects he has found in audio equipment. He said the evidence of a noise problem can merely be the presence of, “Veiled" or "grainy" quality in audio. How do you measure veiled and or grainy?
And BTW, Whitlock did not publish any controlled listening test proof of his claim in his document I quoted from. Nor did he publish such proof anywhere else I am aware of.
Quote taken from:
Last edited:
The trouble is it's very hard to do a proper blind comparison as you need to unplug things to chenge over. So you've done a subjective comparison which is inherently flawed.
thats where not only a short listen plays a role but how it performs longterm compared to other solutions...
preferably not only my opinion in the end but from others too, so we might get a bigger picture of how it really performs, since this is quite unusual in the audiophile usb cable landscape this will be a pretty interesting test, i tested a few "straight usb cables" already
on the other side, i say nothing "against" measurement, if you do them please share them, maybe i do them in the future, (but i find it more interesting to see the measurement correlating on how it sounded, to maybe draw a conclusion from it)
but imo (unless we talk "really audible levels") its pointless to discuss whether its inaudible purely from measurements, its not like you "actively HEAR that noise" ... thats in regions where people come up with phrases like "blacker background/less hash/less harsh/smoother/holographic" etc
It's not pointless at all. It's telling you it's most likely not going to make an audible difference. The only way to test that is then with blind testing, preferably with instantaneous switching.
What exactly is telling him "it" is most likely not going to make an audible difference?
@Ghoostknight only said it was pointless to discuss something, most likely meaning pointless to discuss with people who would denigrate and mock people like Bill Whitlock (or other people making similar claims), and describe Whitlock's (and other people's) work as flawed because someone dared to describe a sound without controlled blind listening test "proof."
Otherwise, the people doing the mocking and denigrating must not be operating from principle, but rather for some more personal animus based motivation.
Last edited:
I love that word - "filter" the noise. It's used so often these days because there are noise generators/ noise factories left, front, and right. Some even plug them into their audio systems - directly by means of various types of interconnect cables (analog and/or digital), and indirectly through the shared mains outlet. Here, the indirect is actually - direct, but this way of noise infestation is commonly overlooked - nevertheless as harmful as the first one.
... but how do you filter the noise? That noise has to go somewhere (basic physics - the law of conservation of energy).
So... what happens with the noise?
Last edited:
Conducted to earth. Reflected to the aether. Converted to heat.
As usual the easiest solution to solve the problem is to not let it being born but that would be too effective.
As usual the easiest solution to solve the problem is to not let it being born but that would be too effective.
If you are sure your filter "does something" then just show any measurement that shows a difference. Otherwise how can you be sure that there is a difference if you don't measure? Human hearing is not suited for the task especially since no proper comparison is made.
Just like anything else. If you are sure there is an audible difference you take various measurements until you find a measurement that shows a difference. No need to stick with FFT or SINAD. Anything that is audible to humans can be measured. Audible differences in controlled listening test show differences in measurements. Uncontrolled sighted subjective listening test involves other senses and brain which makes any findings unreliable.
Yes, this is basic physics. Read here how ferrite beads filter noise:
https://www.analog.com/en/resources/analog-dialogue/articles/ferrite-beads-demystified.html
"A ferrite bead is a passive device that filters high frequency noise energy over a broad frequency range. It becomes resistive over its intended frequency range and dissipates the noise energy in the form of heat."
Can you give a complete list of what needs to be measured and with what equipment ? It is not a joke, those SINAD numbers are just a some part of sound portrait