Popular Post mansr Posted June 3, 2017 Popular Post Share Posted June 3, 2017 It doesn't matter if Amir is the biggest jerk to ever walk the face of the earth. By refusing to sell him a device, Alex still looks like he is hiding something. I can see three possible reasons for this refusal: Alex knows it's a scam and is afraid he'll be found out. Alex has no confidence in his product and is afraid it will be shown useless. Alex is a petty, vindictive dick. I'd also like to point out that the approved reviewers who'd be permitted to measure the device are known for never publishing a negative review. Even when the measured performance is questionable, they find some way to praise the product. esldude, sarvsa, plissken and 3 others 6 Link to comment
Popular Post mansr Posted June 3, 2017 Popular Post Share Posted June 3, 2017 4 minutes ago, Superdad said: That means a lot in the face of the barrage that my act of cancelling Amir's order (for personal reasons) has caused You know, you could have spared yourself all that grief by simply selling Amir the damn thing. esldude and tmtomh 2 Link to comment
mansr Posted June 3, 2017 Share Posted June 3, 2017 39 minutes ago, PeterSt said: This is for those who like challenges. This is 15 minutes long, the phosphor of the scope never deleting. Btw mind the horizontal scale which is meant to show the length of the audio word (so this is the output of Jabbr's DAC, so to speak). The below shows that this is 13ps of p-p jitter but it's a moot thing because I use the trace width to measure, and the width is 13ps "thick" and can't be more thin. Call it scope limits. But this should really be well under 1ps p-p. Btw this is not from USB isolation; this is the internal isolation in I2S. One of these days I will make some plots from USB itself, behind the USB isolation (our own). People, if this would be so easy to measure, why don't we see these plots anywhere from manufacturers ? Answer : not per se because the outcome is bad. But the gear is expensive ... Here's a bonus : 210ps p-p of jitter. What changed ? this is without in-DAC isolation ... (this is data impeded jitter) What exactly is being measured in those graphs? What scope are you using? Link to comment
mansr Posted June 3, 2017 Share Posted June 3, 2017 1 hour ago, wgscott said: What would be of great benefit would be to depersonalize this issue. Alex made it personal. He can fix that by selling a device to Amir. Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 10 hours ago, Jud said: Importantly, I clearly heard it in a blind test. Several folks pushing for measurements in this thread have, in other threads involving measured differences, said the real acid test would be the ability to tell a difference in a blind test. Here we have that. No disrespect intended, but what you have provided amounts to nothing more than an anecdote. We don't know the difference between the devices you tested, and your setup wasn't documented. This means nobody can accurately repeat it or even assess how well you accounted for incidental variables that might have affected the outcome. The number of participants was also far too small to say anything with confidence. I realise you didn't set out to perform a scientifically rigorous test, so please try not to pass it off as one. sarvsa 1 Link to comment
Popular Post mansr Posted June 4, 2017 Popular Post Share Posted June 4, 2017 8 minutes ago, Jud said: Certainly it wasn't a scientific test. Now will everyone who doesn't have scientifically conducted blind tests verifying a piece of their audio systems functions as claimed, and who does not completely understand the circuitry and function of that piece of equipment on an engineering level with verifying measurements, please get rid of that unscientific piece of no-good junk. Once we've cleaned out our audio systems, we can start on our computers, cars, and homes. Nice straw man. I've never suggested that everybody must understand and personally validate the design of everything they use. It is quite sufficient that somebody does this. The problem with many audiophile widgets is that nobody has done anything to validate their operation. In some cases, the manufacturer even actively prevents others from testing it. Would you buy a car from a manufacturer who refused to let it undergo standard testing? What of a car manufacturer who was found to be rigging tests in its favour? Do they get a free pass, or are they hauled in front of court? esldude, tmtomh and sarvsa 3 Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 3 minutes ago, Jud said: I believe both Bill and Dennis use digital room correction/equalization. There is substantial academic literature regarding human insensitivity to the overall system frequency response this is designed to optimize. Nearly all these systems use minimum or intermediate phase filters. Is the favorable response to these systems due to overall frequency response optimization, or to a slight euphonic "reverb effect" in the audible range from the post-ringing of these filters? I'm unaware of academic/scientific studies proving one or the other. Why Bill and Dennis, you cock-eyed subjectivists! I don't know about Bill and Dennis, but I use DSP mainly to tame a few low-frequency resonance peaks and for subwoofer integration. The difference is readily measured and easily audible with some music (anything hitting those peaks gets a boomy and uneven bass). I doubt it makes much audible difference above a few kHz, but the effect is again measurable and it can't hurt to have a more correct response. Link to comment
Popular Post mansr Posted June 4, 2017 Popular Post Share Posted June 4, 2017 Just now, Jud said: Are you quite certain, through scientific testing you or someone else has conducted, that it is a more correct response via-a-vis any possible reverb effect in the mid and upper portion of the frequency band? I'm quite certain that removing, say, a 5 dB hump at 100 Hz is better than not doing this. plissken and tmtomh 2 Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 1 minute ago, Jud said: Better than not doing it if you factor in potentially adding euphonic reverb in the mid and upper frequencies? You said the difference in bass was "clearly audible" to you. Did you test this blind so as to remove any expectation bias, or even better, research the academic literature to confirm scientifically the favorable effects you thought were clearly audible? No, I haven't done rigorous testing. However, I'm not the one selling it, so that burden isn't on me. I also have no reason to suspect them of secretly adding reverb rather than doing what they say they're doing. tmtomh 1 Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 24 minutes ago, Jud said: Of course with regard to minimum phase and intermediate phase filters, there's no "secretly added" reverb. Rather there is the potential for reverb if post-ringing has that audible effect, something various people have said is so, but that hasn't at least to my knowledge been established in the scientific literature one way or the other. Filter ringing occurs at inaudible frequencies. Reverb is something else entirely. tmtomh 1 Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 28 minutes ago, Jud said: I've read several people quite familiar with digital filtering (as I recognize you are yourself), including Miska, Peter, and whoever did the filter design work for Ayre and Charles Hansen, say that ringing, though ultrasonic, smears the audible signal over a greater period of elapsed time, creating an audible effect similar to reverb. Those people all sell their own various magic potions intended to remedy this supposed problem. I won't go so far as to call them liars, but they have more than a little motive to exaggerate the severity of the issue. It is trivially true that band-limiting a signal involves some amount of "smearing" in time if the input has high-frequency content that gets removed. Low-pass filtering and smearing are synonymous, insofar smearing can be considered a technical term at all. tmtomh 1 Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 3 minutes ago, PeterSt said: No-one in the world sells USB cables with specs Anything carrying the official USB logo is required to conform to the standard specification. Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 Just now, PeterSt said: Yes, but that's the other way around. I see few audio cables with USB logo tags. So I guess they must all be off-spec. Absence of the logo doesn't necessarily mean it isn't compliant, only that nobody has had it formally certified. Just now, PeterSt said: And you know what ? they are. They have to be, because the one we carry is so much on all specs possible that it sounds unequivocally the best of them all. I am even serious. But ours too, no USB logo on it. Only our own. I'm not sure what you're trying say there. Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 4 minutes ago, firedog said: And I guess in some alternate universe that actually happens. Sure, there are "fakes" that don't actually meet the spec. That's a problem. The point is that by using the official logo, you are stating that the device/cable conforms to the spec. Whether or not that is true is a different matter. Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 15 minutes ago, PeterSt said: No USB spec has been made for audio. All USB devices share the same electrical spec. There is no special version for audio just like there is no special version for storage or networking devices. The document titled "USB Device Class Definition for Audio Devices" details a protocol for use with audio devices. What exactly do you suggest is missing? Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 12 minutes ago, mmerrill99 said: Are you saying that USB audio has bit errors? In my testing, I had to try pretty hard to provoke bit errors over USB. When they did occur, it was blatantly obvious. tmtomh 1 Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 1 minute ago, PeterSt said: The way we use USB Audio is not so much error corrective, like the normal usage. So a hdd connection is allowed some error rate (though don't ask me where and how this is written out) because it will be corrected. Well, really no need to elaborate further. Yes, as stated in the spec, USB audio uses isochronous transfers with error detection but without retransmission on error. Why do you insist there is no spec? Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 8 minutes ago, PeterSt said: How did you know you had USB errors ? I captured the digital output of a USB to S/PDIF interface. With a proper USB cable, there were no errors. With a tangled mess of wires in place of a cable, errors started showing up. tmtomh 1 Link to comment
mansr Posted June 4, 2017 Share Posted June 4, 2017 4 minutes ago, pkane2001 said: Please explain how an optical isolator will not stop ground loops or leakage currents.. By the way, current audio standards require well under 100Mbps bandwidth. USB wire speed is 1.5 Mbps (low speed), 12 Mbps (full speed), or 480 Mbps (high speed). If 12 is insufficient, you must step up to 480. That said, optical technology capable of far higher speeds is readily available (Ethernet). I don't know what USB products exist, if any, but that doesn't mean there are none. Link to comment
mansr Posted June 5, 2017 Share Posted June 5, 2017 5 minutes ago, plissken said: I mean I'm pretty sure I could feed a DAC something out of spec that it isn't expecting / designed for, say 120V, and wouldn't even need any speakers connected to tell the difference. I was fooling around with an old Cambridge Audio AVR recently. As it turned out, feeding it an unexpected, but in-spec, S/PDIF signal made it emit smoke. Link to comment
mansr Posted June 5, 2017 Share Posted June 5, 2017 Just to put things into some perspective, here are a couple of scope captures showing extreme jitter levels in a Steinberg UR242 audio interface. The scope is watching the LR clock at the DAC chip. This image shows a full LR clock cycle at 48 kHz. The measured cycle time has a standard deviation of 2.75 ns. This shows the rising edge of the LR clock with a fixed delay of one period after the trigger. The peak-to-peak jitter is about 15 ns. Horrific. This is the same capture but with longer (10 seconds) waveform persistence. Clearly, this level of performance is considered acceptable by the designers of this device as well as those using it. Is it really necessary to fret over femtosecond level jitter? esldude 1 Link to comment
mansr Posted June 5, 2017 Share Posted June 5, 2017 Here's an FFT plot of a 1 kHz sine tone from that horrid Steinberg device and from the iFi Nano recorded at 192 kHz on a Tascam UH-7000. The FFT size is 32M which gives a frequency resolution of about 6 mHz. Note the much wider skirt exhibited by the Steinberg. Same plot zoomed out a bit: esldude 1 Link to comment
mansr Posted June 5, 2017 Share Posted June 5, 2017 3 minutes ago, PeterSt said: What can you say about the difference in "noise line" between the two ? Is the Steinberg so way more averaged or something ? The Steinberg has much lower noise overall. This is a wider plot with ~1 Hz resolution: Link to comment
mansr Posted June 5, 2017 Share Posted June 5, 2017 5 minutes ago, plissken said: That kills the audiophiIe styled DAC. The UR242 is $139. Is this with balanced / XLR on the Steinberg? If so this lines up with the testing Amir started doing with the $79 Behringer and he didn't even capture the balanced output, only the singled ended. The Steinberg has balanced TRS jacks. In fairness to iFi, their more expensive models are significantly better than the Nano. Link to comment
mansr Posted June 5, 2017 Share Posted June 5, 2017 1 minute ago, jabbr said: Thanks. Interesting isn't it how these plots look very different than the zoomed out plots. So perhaps the difference in noise baseline between -120 vs -150 db isn't so important There's more to it than that. At higher frequencies, the Steinberg starts doing decidedly weird things: Link to comment
Recommended Posts