@Rob Watts
Hi Rob, I'm kind of curious about a few technical aspects about your DAC designs and have a few questions I've been thinking about:
1. What have you done to combat inter-sample peaks in your DACs?
2. I have read little snippets of the DAVE + Blu MK2 combo having high latency. Is this true? If so, how many milliseconds latency? The reason I ask is when M-Scalers are built-in even your budget level DACs (future Mojo?) and the technology will be available to more people. How will musicians use it or even gamers if there's noticable lag? I don't notice anything with my Hugo 2 but that only has a tap length of around 50k.
3. You have calculated that the DAVE + Blu MK2 combo has transient reproduction accuracy to 16-bits. Does this change as you use the digital volume control? For example, if one reduces the volume from 0dB to -12dB, does that reduce transient accuracy down to 14-bits?
4. Would you ever consider creating a more extreme version of your green Hugo 2 filter in the future? What I mean by that is, instead of a slight roll-off for higher sample rate content, how about complete elimination of content above say 22KHz a la brick wall? I say this as some external amplifiers have trouble with ultrasonic content and generate huge amounts of intermodular distortion. Plus the advantage of higher sample rates is the improvement in timing precision and not the ultrasonic garbage we can't hear.
5. I know that this has been brought up before and even I refuted it based on the system they used to measure the Hugo 2 with. But what do you make of the output impedance measurements and headphone load THD measurements posted here before? Does the Hugo 2 have feedback issues with demanding impedance swings?
I have one more question too regarding transient accuracy converted to bits. It's 2am right now and I'm half asleep, so I may not be thinking straight about this. Please correct me if I'm wrong. But if the human hearing system has temporal acuity to as low as 4μS. If one converts that to arcseconds (I used an online calculator so may be wrong) that results in 0.00006 arcseconds. If one then converts that to degrees that results in 1.66666667e-8. This may be another crude conversion, but as there's only 360 degrees in 1 full turn, aka it's 100% you could sort of calculate a bit level out if it (every half is -3dB etc). 360 halved around 35 times would bring you close 1.66666667e-8. That's 105dB which is around 17.5-bits. If my calculations are correct (which I doubt, I have probably embarssed myself), could that mean a 4 million tap length filter would be the golden number?