Differences in sound: DAC vs. DAC + Pre-amplifier

[ecwl]

Forgive me for not having read all 18 pages here. But I did a quick search for some keywords and didn’t see them so I presume (and was surprised) this wasn’t discussed. Although I think @barrows sort of touched on this a little

 

I have always thought that one of the reasons why digital volume can sometimes sound inferior to an analog preamp is because a lot of DACs don’t have great low-level linearity. This is measurable except I only see Hi Fi News and Audio Science Review measure this.

 

Randomly I selected one DAC with good linearity and one with poor linearity from both sites:

https://www.audio “science” review/forum/index.php?threads/topping-d90-balanced-usb-dac-review.10519/

https://www.audio “science” review/forum/index.php?threads/measurements-of-musician-pegasus-r2r-dac.18786/

 

https://www.hifinews.com/content/simaudio-moon-780d-v2-network-attached-dac-lab-report

https://www.hifinews.com/content/ps-audio-stellar-gain-cell-dac-stellar-m700-dacpreampmonoblocks-lab-report

 

So you can imagine as you reduce the digital volume, even if mathematically your digital audio signal is accurately reduced, the DAC simply cannot accurately recreate the low-level analog signals because of the poor linearity. Of course, playing the DAC at full power and then using a good analog preamp to lower the volume level would mask the low-level nonlinearity issue.

 

And of course, there are other reasons why people prefer analog preamps.

 

 

[ecwl]

45 minutes ago, AudioDoctor said:

 

I think this gets to why HQPlayer has, in my opinion, a great software volume control. Perfect, or near as mathematically possible, linearity and as the volume control can be done in conjunction with upsampling, no resolution is lost in the process.

While this is somewhat true, but even if you’re to use HQPlayer to send excellently upsampled and noise-shaped DSD512 signals to your DAC, your DAC hardware still has to be able to produce that low level of linearity. A simple thought experiment is that if your DSD DAC has lots and lots of jitter, it doesn’t matter how great the noise-shaping is for the DSD512 signal, the actually analog output would still have poor low-level linearity.

[ecwl]

11 minutes ago, sandyk said:

Nevertheless, it still comes down to how electrically quiet the source device is, and this will also apply  with conversions to DSD.

If you doubt this, and are using a Desktop PC try doing what one and a half (Gary) is recommending here.

https://audiophilestyle.com/forums/topic/60381-hdd-to-case-bonding-uptick-in-sq/?tab=comments#comment-1084599

 It won't cost much at all to at least try it.😉

 

I don’t think that’s totally true. Obviously, the ASR measurements are presumably done on the same PC’s and the Hi Fi News measurements are don’t on a different PC that is used for all the measurements.

But I don’t think introducing a high-end desktop PC with no noise injected into the DAC is going to suddenly make a DAC hardware with low poor low-level linearity into a DAC with great low-level linearity. By all means, anyone who is interested can send your favorite low-noise PC to ASR or Hi Fi News for them to use.
I think a lot of this is fundamental to the DAC’s hardware architecture. But of course, you can somewhat bypass that architecture if you use your PC to upsample and noise shape to DSD256/512/1024 but then you’re now reliant on the DAC’s DSD hardware architecture.