This is my D-6 with DSD512. I only get the uncorrelated DSD noise ?
Yes, there are, see above in this thread. If you want to use 20 kHz SINAD rating as measure, A26 is very good on that front.
Yes, so no worries that you would get any audio band distortion from that one.
Ya, and in band is similarly clean to Cyan 2 . D-6 at DSD512 is very good, not just for the price.
But the DSC1 transients familiarity of the Cyan 2 is something different (but with much cleaner performance and zero PSU interference)
Not in mine, see for example this 44.1k PCM input case, around 100 kHz:
But could be similar to Topping, that it is more or less random how good unit you happen to get. Like my two Topping E30’s, with SNR difference of about 60 dB.
I don’t know anyone that has an A26 or AH90 so can’t share the full suite of measurements like I shared for Cyan 2
Not available to home demo either
Those Cyan2 measurements are my favourite to do. Give a good overall picture on technical strengths and weaknesses
1 Like
Yes but I wrote with DSD512 input
Need to compare Apples with Apples
Can’t compare Cyan2 DSD256 with D-6 PCM44.1k
I had ugly wideband with D-6 at first
And then I played with grounding to the Pico and it cleaned right up
All that crap vanished
So what goes into your amplifier will depend on overall system grounding but that’s the same with any DAC
If I played with my laptop and Pico grounding during Cyan2 measurement I’m sure I can introduce something funky to show
I use the same setup always…
Cyan 2 has LPS, so it cannot have such whine from SMPS.
Personally I think D-6 is very good and not just for the price. Very good of course doesn’t mean perfect. We can nit pick one thing about each DAC
The main problem I have with SMSL is every AKM DSD Direct DAC they released after D-6 has been sh!t with DSD input
So once D-6 goes out of production maybe this year, that’s the end of any recommendations 
I found their D-6 issue at DSD512 and they fixed it immediately but I don’t have time and motivation to go through that for every DAC they have
1 Like
For 4499EX DACs it’s about bugs the AKM chipset has (and in D-6 case it was too). AKM catched up with ESS in number of bugs since they dramatically increased configuration options.
Gustard took the time to figure out how to work around this. Time costs…
2 Likes
It’s nice that they fixed the obvious thing
But I haven’t seen a full suite of A26 or AH90 measurements, so not sure on its differences vs D-6 and vs Cyan 2 .
I have to wait for someone locally to offer me a loan.
Anyway my biggest decision is now whether to buy a Cyan 2 
I wonder how long my friend will let me borrow it to keep listening
Beyond measurements, I have no interest in reading other people’s subjective impressions.
So beyond measurements, it is best to get one to demo and decide with own ears 
I can’t see anything from the measurements to make me want to try I2S. USB is very well sorted by Holo. I don’t have a separate I2S box and no motivation to find one really.
Late last night I wasn’t happy with my IMD vs level plot.
From Multitone Analyzer on the laptop I was outputting to RPi4 NAA input and then it back from HQPlayer into my NAA and into my ADC, also connected to laptop.
Something in this loop was causing a blip in the IMD vs level plotting, getting results was really inconsistent.
So very late last night , I used Multitone’s Analyzer’s built in DSD upsampling to DSD256, bypassing all network (just for IMD vs level plot). Just straight out to the DAC and straight into the ADC. ASIO in and out.
I got the same result each time, so this result is reproducible.
Nothing weird about this result. Small turns up and down are minor. Mainly looking for a big rise up in IMD and we don’t get that. Nice.
1 Like
Part of the reason is varying relatively large delay on that path. Input side clock is only virtual. (which is not actually used to clock anything)
1 Like
Ah i see. Actually all the REW measurements above (exception is the level sweep) are:
iPad (Onkyo HF 24bit / 44.1kHz test files created in REW) → RPi4 NAA input → HQPlayer DSD256 output → UpBoardGateway NAA → Cyan 2 → ADC → REW RTA on laptop
While measuring Cyan 2 I also wanted to confirm iPad as source and feeding RPi4 NAA input , that there was no funny business happening. Indeed there is no funny business.
It’s also good to reproduce main listening chain as much as practically possible.
It’s really only the level sweeps where I think its best to use Multitone Analyzer’s built in DSD upsampling and ASIO out and ASIO in, because any blip means you need to restart the level sweeps again (I lost 45 minutes last night )
Even my Pico measurement is from test file stored in Onkyo HF app on iPad.
tests my normal listening chain
For static test signals it is easy, because there’s no expectation between particular output position vs input position.
1 Like
Cyan 2 direct USB connected to Mac Mini M1 with RPi4 NAA input (Apple Music Hires on iPad)
DoP 256 enabled. Auto family rate enabled to help lower CPU loading
And new 5EC-ul (Ultra Light) modulator also to help keep M1 CPU lower also.
With Cyan 2’s built-in USB galvanic isolation, no worries.
Nice office system !
1 Like