I may be wrong, but I think what you see there is the analog signal output by the DAC, not the DSD signal applied to it, so you can’t tell what the latter looks like.
@Marian in post #79 above Jussi mentioned:
I think you are talking purely about the digital domain but isn’t the practical sense more important?
i.e. measurements of the analogue output of D to A ?
Especially when we are talking about the “sound of DSD” ?
In the ADI-2 example above, we clearly see improved objective performance upsampling to 1bit ~12MHz
That’s correct, I am talking about the digital domain.
It surely is, and I did say I was not concerned about the audibility of the DSD noise, just about the correlated nature of it, for the sake of correctness. If a DAC supports DSD and measures better with it, by all means, use it. I would however expect a good DAC to handle PCM properly too. After all, DSD can’t provide any information that’s not there already.
The RME ADI-2 is certainly a great DAC. It handles PCM very well… but you can see it performs even better when fed DSD256 rates.
Btw Jussi has shown this for many many DACs over the years. I’ve only shown you one example.
While you are only focused here on the digital domain, his main focus is on improving the analogue output of DACs using HQPlayer - whether PCM or DSD. By measuring the analogue output. The most important thing for sound quality.
In terms of the ADI-2 it’s max PCM input rate is PCM768kHz. It’s analogue output performance is still best with DSD256 rate regardless of any PCM rate.
Anybody is free to knock themselves out at the objective performance (analogue output!) with PCM test signals but also keep in mind performance is elevated again with DSD256 input (for the AKM model running in DSD Direct mode):
Let me bring the practical aspect of this and ask: can you hear those differences? First, I would not bother too much with anything above 20kHz - other than to make sure it’s not damaging-high. Below 20kHz, if you bring some peaks from -140dB to -160dB, does it really matter? I seriously doubt it. Software can always be made to perform better than hardware, but beyond a certain threshold, those numbers don’t mean anything anymore beyond marketing.
In a blind test possibly not but I live to feel good. I don’t live to just do blind tests.
I feel warm and fuzzy and good on the inside knowing that HQPlayer allows me to objectively improve my already very very good D to A’s…
And simultaneously I enjoy reading the subjective impressions of others like this thread was initially partially about. I definitely don’t dismiss somebody’s subjective impressions.
1 Like
Can’t argue with ASMR.
Or as you previously wrote… 
Can’t argue with myself either 
Not the first time I’ve read these posts, but they’re a welcome revision 
They are one mans way of seeing things. I can’t say he’s “wrong” though, and thats beside the point.
I do like his constructions though, they sound good to my ears!