From the Denefrips Terminator specifications:
Matched Precision Thin Film 0.005% R-2R Network Arrays
The precision of resistors is quite good at the present.
And precision is not the only factor in sound quality.
CDs were from the beginning more precise than vinyls, but did they sound better?
That statement is just plain crazy and needs more contextā¦ 
The digital sound of the CD player is more precise, as it is delivered with the regularity of its clock, while the rotation of a vinyl can not reach that regularity, even with a very good turntable with a very good power supply.
Everything is dependent on outside influenceā¦ You are describing one element of precisionā¦ the definition of the term āprecisionā in a system, is dependent on several factors that generally define a measureable average āprecisionā relative to a contextual experienceā¦
Precision can always be measured with the right tools.
But itās not the only factor in sound quality. For instance, itās well established that solid state amps are more precise than tube amps. Nevertheless, the sound of tube amps remains more pleasant.
Why do you think Denafrips uses FPGAs?
Yet another subjective statementā¦
How precise is oneās hearing acuity? The measurement (by analysis) of this is quantified and general, and is dependent on the acuity of the subjectās hearing neurology and physical elements.
I presume this is whyā¦ I wanted to make sure these things were made evident to @Sjcrowther and anybody questioning the capabilities of their R2R DACās.
The human hearing is very sensitive.
The fluctuations of jitter, for instance, are measured in nanosecond, if Iām not mistaken. Itās very littleā¦ but human hearing is sensitive to it.
ā¦ This all plays into the construct of what is āPrecisionā?.. and why your statement needed more contextā¦ Where is precision specifically measured? What is the contextual connotative definition of āprecisionā in the comparison? Obviously, āprecisionā is not one thingā¦ it is dependent on a variety of contextual elements.
At the present state of technology, the resistors of the R2R DAC convert the digital sound into an analog sound (analog electrical currents) with a lesser precision than a good Delta-Sigma DAC. The measurements prove it. Just as the measurements prove that a tubes amp amplifies the sound with a lesser precision than a solid state amp.
This is how I understand precision in this context.
Magnetic 
Attraction
Can you send us a copy of the āDebug Infoā when you have this issue? You can get it in Audirvana Settings>My account
Note: By clicking on the Debug Info button, Audirvana automatically copies all its content, all you have to do is paste it in response to this message.
Yes.
The interesting parts of the thread are the graph showing the deviation beyond 16 bits and the following Jussi Lasko comment:
āIt is technically impossible to make R2R DAC that is is linear to 24-bits. This is just based on reality and laws of physics. IOW, where LSB have accuracy better than Ā±Ā½LSB. For Denafrips, 16-bit figure is based on the measurements Iāve seen.ā
When I use HQPlayer to force 16 bit output for 24 bit files the sound is slightly fuller and the image more precise. Switching back to 24 bits the sound is slightly more diffuse.
Anyway I apologise for taking the thread away from Audirvana to Denafrips. I just wanted to know whether I could force 16 bit output in Audirvana and this has been answered.
No problem, you always learn something. I was kind of thinking of getting the Ares II just to experiment a bit with discrete R2R DAC. If itās true that the linearity is not preserved beyond 16bits to the point itās audible, Iām not going to waste my money and stick to the 16bit ladder DACs I already have.
Yes and no.
I have a D/S DAC as well - Pro-Ject DAC Box RS. I used to have the Chord TT and have spent time with the Dave.
The Pro-Ject is good. I donāt like the Chord house sound: very clear, detailed, revealing but cold and very āsiliconeā.
The Denafrips sounds far more appealing to my ears, linear or not. It is just that to me it sounds even better restricted to 16 bits (by a small amount) for PCM. The DSD is just great. Surprisingly close in sound presentation to highest res. PCM.
Not many DACS do both.
Have you tried to covert the file to 16bit using XLD or something like that?
Iām saying this because maybe HQPlayer is doing some monkey business when you select 16 bit output. It could be that what you like.
Do you also upsample at the same time?
Have you also tried to upsample (with Audirvana) everything to DSD? How does that sound with your Denafrips?
Yes I up sample to 705.6kh and do other processing as well but that has nothing to do with the linearity of R2R ladder DACS.