What digital filters are Schiit and Fiio using? And does it make a difference?

[Fluffy]

I Have a Schiit Modi 3 and a Fiio X3 2nd gen. the Fiio doesn't have enough power to drive my Focal Clear or Audeze LCD2C, but it's doing a great job with the Meze 99 Classics. The thing is, I always preferred listening to the Meze via the Fiio instead of the Schiit, who's feeding an Arcam rHead (a much better amplifier overall). I never really given it much thought, and in AB testing I couldn’t really tell them apart, so I guessed it's some expectation bias. Anyway, I continued to use the Fiio for the Meze because why not.

Recently, I felt quite… fatigued, let's say, from the Schiit/Arcam configuration. I don't know why, it just makes my head hurt lately. So I switched over to listening through the Fiio and as always, it felt much more pleasant. This time it really struck me as weird, so I decided to test something I have yet to try – the digital filter of each DAC. I did it by creating a 1 sample impulse in audacity at 44khz, and playing it thorough both DACs, while recording into my Focusrite 2i2 set to capture at 192khz. And here are the results (normalized):

So it appears that Schiit is using a linear phase while Fiio is using something close to minimum phase. I actually never try to test on myself if this is something that I can actually hear, but now I think maybe it could have an effect. The thing is I don't really have a way of blind testing this – it would require a DAC that can switch between these two filters, so to eliminate any other factor.

Can anybody corroborate that these are the filters used in these devices? If so, is this a brand-level decision to go with a specific filter, or it's just due to these particular DACs? And could it be the reason I prefer the Fiio over the Schiit? I'm saying again, I'm really not certain that this is a valid claim and it is probably some placebo effect, but I have no way to blind test this.

 

[solderdude]

In the FiiO you can select 2 filter options.
The Schiit just uses the standard filter in the DAC chip.

You have to realize the impulse is an illegal signal which cannot exist in any audio file.

https://reference-audio-analyzer.pro/en/report/dac/fiio-x3-ii.php

I suspect the device was measured before the firmware update that brought 2 filter modes, steep and gradual

 

[Fluffy]

In the FiiO you can select 2 filter options.
The Schiit just uses the standard filter in the DAC chip.

You have to realize the impulse is an illegal signal which cannot exist in any audio file.

https://reference-audio-analyzer.pro/en/report/dac/fiio-x3-ii.php

I suspect the device was measured before the firmware update that brought 2 filter modes, steep and gradual

I actually tested the two types of filters in the Fiio – Graduated and Steep. They both give this minimum phase response; the difference is in how sharp is the filter after 20khz. The difference is pretty small actually, nothing that should be considered audible.

I know that these impulse responses represent an unrealistic scenario. But still, these filters have an effect on regular music because of the phase shift, don't they? And at least in theory they should make transients have more attack because of the lack of pre-ringing - not that I'm saying it should be audible, I don't really know.

 

[solderdude]

The pre-ringing is not present in the 3-5kHz range, that's where the 'attack' in music is. Not in the >10kHz range.
Regardless of the filter a 3kHz at 0dBFS will be the same in both players.

The Cirrus chip has a 'pseudo linear phase filter' so would expect the phase to be close to linear phase.

 

[dc655321]

And at least in theory they should make transients have more attack because of the lack of pre-ringing

What theory predicts that?

But still, these filters have an effect on regular music because of the phase shift, don't they?

Sure. The linear phase filter shifts all phases the same amount (i.e. an inaudible delay), and the minimum-phase-ish filter likely may have some phase shift as its corner frequency is approached (i.e. at high frequency where one may not notice or won't with music).

Since you're recording in a loopback, play a track through your Audacity filters, record into separate files and null-test the results (with Delta Wave?). No extra gear required.

 

[Fluffy]

Since you're recording in a loopback, play a track through your Audacity filters, record into separate files and null-test the results (with Delta Wave?). No extra gear required.

I didn't use loopback to record the impulses mentioned above. I recorded directly from the headphone out of the fiio\schiit-arcam, into the mic inputs of the Focusrite. In other words, it came out of the DACs in analog form, and came into the ADC of the Focusrite.

Recording in aucadity in loopback just gives me the original digital impulse with no filters applied…

 

[dc655321]

I didn't use loopback to record the impulses mentioned above.

Ok, my mistake.
In that case, simply level match output from both dacs, record the same level-matched track through each dac into the Focusrite, null-test each file using Delta Wave to see the differences.

My hypothesis is that any differences would be minimal.