I use a Sysconcept now mainly.
Modern DACs should be jitter immune, and Rob Watts has stated that the Hugo 2 is from memory. That initially made it puzzling that a cheap cable should affect it. The only possible reason I can think of for this is that, like USB, in the process of dealing with the jitter-heavy signal, the electronics of the Hugo 2 itself end up generating enough noise to cause distortion in the output.
My experience is that more complex USB gear makes the sound brighter; this also applies to audiophile USB cables too; the technical explanation is that this increases RF noise, making noise floor modulation worse. And a small amount of noise floor modulation spices the sound up like MSG. So my advice is to keep the source bit perfect and as simple as possible; also your reference benchmark should be optical, and your target would be to get the USB to sound the same as optical - I have managed to do this with Hugo 2 on headphones with a large powerful MSI lap-top, which has USB and optical out. One of the targets with Hugo 2 was to close the SQ difference from these two inputs, and that I have done (even though I was not expecting them to sound identical!).
Wow, lots of questions, I will try my best to answer.
1. Inter sample peaks never have been an issue as the first WTA filter has an embedded 2.75 dB overload margin; this ensures that inter-sample overloads will never happen.
2. Yes, the M scaler has about 630 milli-sec (0.63S) latency with 48k source, but it has a video mode where the latency is the same as Dave - 104 mS - which is normally OK using modern projectors and TV's (but set your source latency to 0)
3. OK the M scaler has the same as an ideal sinc function (which will perfectly recover the original bandwidth limited analogue signal) to 16.6 bits accuracy. So in effect this means that the interpolation under all conditions will not get any worse than 16.6 bits accuracy, and in practice would be better than this. I hope to have some measurements soon confirming the actual transient timing error. Now the volume control is after the WTA1, so volume has not impact on this.
4. Yes, in the case of HD recordings, there may be some small benefit in more filtering. But for regular 44.1, no I think that is unlikely. In fact, with 44.1k, the stopband attenuation is -135 dB, and most of the filtering is better than 150dB - so the WTA filter already effectively removes out of band images and noise.
5. No Hugo 2 certainly does not have issues with demanding impedance swings. Also, other measurements published either use test equipment that is not capable of measuring Hugo 2, or have applied methodology that fails to take into account the large latency of the WTA filters.
As to your final point - I am not sure your sums are right! 1 bit is a 6dB change. And perception measurements have detected a sensitivity of about 1 degree. Also, although 4uS is the limit to interaural delay, the ear/brain is actually much more sensitive to timing errors than just the 4uS - otherwise we could not hear the difference between the 16FS to 256FS WTA filter, as this works well below the 4uS limit. And there is more to timing errors than just the interaural delay, as these errors create more problems than just imaging (which is what interaural delay is used for - left right imagery), such as the perception of timbre, instrument separation and low frequency pitch.
Thank you for the well written answer.
1. So the Hugo 2 has 2.75 dB overload margin. The reason I asked was because Benchmark Media state their DACs have around 3.01 to 3.5dB of headroom. Which do you think is the optimal amount and does Hugo 2's "headroom" increase as you lower the digital volume control?
2. Does the video mode reduce quality at all? If future DACs have this sort of latency, I don't think they could be used in a studio or gaming situations - as they need millisecond latency in the single digits.
3. That sounds good, thank you for clarifying. Would be interesting to see measurements, do you think using an ETC is a good way of testing temporal performance? There's a website called "Advanced Audio Analyser" that tests this way, which was the reason I switched to Chord as I compared standard filter DACs and NOS DACs to the Hugo - Hugo appeared to have almost identical timing measurements to NOS DACs, only without the aliasing and distortion issues.
4. Yes, I think with HD recordings there could be a sound quality benefit to extreme filtering of ultrasonics. Would you consider making such a HD filter in the future?
5. Yeah, I don't really trust their measurements tbh. They differ remarkably in output impedance from what the official specifications state. I know there can be margins of error, but not that large and I trust you and Chord to not falsely advertise. Maybe if Stereophile gets a hold of a Hugo 2 it could shut them up lol.
True Rob, regarding the last point. I thought I was onto something but you can't really calculate something in isolation of other metrics, I knew it was a failed attempt lol. I think if one were to formulate some sort of solid timing spec to aim for, we would need to do some sort of testing regarding human hearings ability to detect low-level transient error. It's possible the psychoacoustics scientific tests have already been done on this perhaps, I am not aware of them though.
I have a few more technical questions I would like to ask, if you are happy to answer them:
6. Regarding noise floor modulation, would LED displays dimming zones be a good comparison to explain its effects on audio (as they dynamically lower the absolute darkest shade it can reproduce)? I have always thought that LED displays look flat, with a lack of depth compared to emissive displays such as Plasma and OLED - even if say, an LED TV and Plasma TV has a similar black level.
Also, please can you confirm my understanding of noise floor modulations effect on the outputted waveform. As noise when outputting sinusoidal wave will be visible as small peaks and valleys on the "curved lines", even at its peaks. If the noise outputted is dynamic (noise floor modulation), I assume it will result in an alteration of the waves peak, aka the leading edge in musical terms? Is this why, perhaps, that noise floor modulation affects depth as it affects the consistency of waveform peak fidelity? Aka non-linear leading edge performance?
7. What do you make of "linear distortion shaping" I hear of some manufactures attempting to implement now? What I mean by "linear distortion" is that each harmonic component (2nd, 3rd, 4th etc) are at the same level at all power levels before clipping, while also being in the same ratio to one another. Do you think having linear distortion characteristics is something that is audible and one of the goals for neutral sound?
8. I'm sure I saw you mentioned somewhere that under a certain power rating, both the Hugo 2 and DAVE have genuinely zero distortion. Is this true and which number on the digital volume control?
9. I was talking to another manufacturer the other day about amplifiers and he mentioned, the wrong type or too much feedback can cause TIM (transient intermodular distortion). Do you measure your DACs for this?
10. What's the very maximum outputted power in VRMS from the Hugo 2s RCA output? The reason I ask is I want to know what gain setting my power amp should be set to.
11. Would you ever consider making a standalone headphone amp? For example, for people that want Pulse Array transparent sound but with their turntable, or more power for inefficient headphones such as HiFiMan Susvara?