Hardware vs software over/up sampling?

[HansHolland]

Hello all,

Most DAC's do over/up sampling before doing the digital to analog conversion. If I understand the specifications of my DAC correctly then it does 8x over/up sampling with a resolution of 32-bit.

I am using Roon. Roon is able to over/up sample. It is claimed with 64-bit floating point precision.

The limit of my system for letting Roon doing the over/up sampling is the USB interface. My DAC has a highish 768kHz 32-bit USB input. Thus 16x over/up sampling is possible. And the calculations are (in theory) preciser.

Now I have the feeling that I hear that over/up sampling with Roon sounds better than letting the DAC doing the over/up sampling. But that might be just my brain that wants that it sounds better because I did some effort. Now my question: can somebody measure both situations and compare them?

I think that the difference is the largest when:
-the over/up sampling factor is large: choose a 44.1kHz track and a DAC with a high specification DAC input
-the DAC has NOT a very high specification over/up sampling filter (if the DAC has 64-bit floating point itself, then it makes no sense)

And do not forget to choose the same filter.
B.t.w. I use a RME ADI-2 DAC FS 1st generation.

Kind regards, Hans

 

[Jimbob54]

Hello all,

Most DAC's do over/up sampling before doing the digital to analog conversion. If I understand the specifications of my DAC correctly then it does 8x over/up sampling with a resolution of 32-bit.

I am using Roon. Roon is able to over/up sample. It is claimed with 64-bit floating point precision.

The limit of my system for letting Roon doing the over/up sampling is the USB interface. My DAC has a highish 768kHz 32-bit USB input. Thus 16x over/up sampling is possible. And the calculations are (in theory) preciser.

Now I have the feeling that I hear that over/up sampling with Roon sounds better than letting the DAC doing the over/up sampling. But that might be just my brain that wants that it sounds better because I did some effort. Now my question: can somebody measure both situations and compare them?

I think that the difference is the largest when:
-the over/up sampling factor is large: choose a 44.1kHz track and a DAC with a high specification DAC input
-the DAC has NOT a very high specification over/up sampling filter (if the DAC has 64-bit floating point itself, then it makes no sense)

And do not forget to choose the same filter.
B.t.w. I use a RME ADI-2 DAC FS 1st generation.

Kind regards, Hans

Do a blind test. Have a friend switch the roon upsampling on and off noting whether you think it's on /off before and after a switch, you blinded.

I think unfortunately it adds a delay into proceedings so not an instant A/B switch, but it's the only way to tell if there's a difference