• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Gustard x16 roll off fast starting around 8khz

dsnyder0cnn

Active Member
Joined
Dec 14, 2019
Messages
216
Likes
215
Okay, I found some time to measure my Gustard X16 using REW and a Focusrite Scarlett 2i2 (gen2). Not the best measurement gear, but adequate to get an idea of what's going on.

Here's the frequency response given eleven different input formats:
Gustard X16 - Formats.png


Note: the scale is only 1 dB (very zoomed in). Also, there's no compensation for the frequency response of my $149 audio interface, which is responsible for most if not all of the 0.2 dB rolloff at each end of the spectrum. This is about the best I've seen.

A number of my ESS DACs exhibit alternating behavior in their impulse response between 44.1. kHz and 48 kHz time-bases. By this, I mean that all multiples of 44.1 kHz look a certain way, and multiples of 48 kHz are sort-of flipped. Not sure how audible this is, but I'd love to know why this is.

Gustard X16 - 44.1 kHz L-FAST Impulse.png

So, this looks "normal" to me. Being a linear filter, we see a balanced amount of pre and post ringing. The step response rises basically straight up and stays essentially flat for at least ~2 ms.

However, the 48 kHz plot looks wonky to me:
Gustard X16 - 48 kHz L-FAST Impulse.png

It almost looks like the phase of the impulse is inverted, but it does "change its mind," making the step positive after the initial flip-flop. This pattern repeats itself in alternating fashion as I cycle through the 44.1, 48, 88.2, 96, ... sequence.

Moving on to the frequency response of the three different filter settings, they are about what I was expecting to see:
Gustard X16 - 44.1 kHz Filters.png

Here, I've separated the plots a bit in REW to make the differences in shape easier to see. Mostly, what's there to see is that M-SLOW rolls off faster than the other two. H-FAST is more extended than it looks in this plot, but not quite as extended as L-FAST. The response anomalies look messy, but keep in mind that we're looking at a range of less than 1 dB here.

However, I find the time-domain effects of different filter options more interesting than their frequency responses. You've already seen L-FAST, so here are the other two, taken at 44.1 kHz:
Gustard X16 - 44.1 kHz M-SLOW Impulse.png

Again, about what I'd expect…minimal pre-ringing and more post-ringing. Now for H-FAST:
Gustard X16 - 44.1 kHz H-FAST Impulse.png

Post-ringing is a little less severe, and no pre-ringing before -200 micro-seconds. I tend to prefer the sound of H-FAST, for whatever reason.

On the question of "NOS" mode, I enabled it and fed the DAC a 352.8 kHz signal. NOS mode just disables oversampling in the DAC with the assumption that you'll be performing upsampling in your media player (eg., HQPlayer, Roon, or whatever). Predictably, the different filters have no effect on either the frequency or time-domain response, but for completeness, here they are:
Gustard X16 - 352.8 kHz Filters.png

Gustard X16 - 352.8 kHz Impulse.png

Here, the pre/post ringing depends on the algorithms used to perform the upsampling.

Testing Process
I used REW's generate function to create 4M length sweeps from 0 to 30 kHz at 96 kHz as WAV files. I included a timing reference in these files. I then used dBpoweramp to convert these 32-bit WAV files to resampled FLAC and DSD files at common rates (44.1 to 384 kHz and DSD64 - DSD256). For a transport, I used a Raspberry Pi 4 running VitOS and Roon Bridge. I started the measurements in REW with acoustic timing reference enabled and then hit "Play Now" in Roon for each of the different formats. I used no calibration file, which explains the upside-down "horseshoe" shape to all of the response graphs…this is just the natural roll-off of the ADC in my audio interface.

Here's a link to my REW *.mdat files in case you want to load them up in REW and look around yourself: Gustard X16 Measurements

Edit: if you look at the distortion plots, you'll see elevated noise in a few. I usually take measurements on battery power, but the battery was running low, so I had to plug-in to finish.
 

Attachments

  • Gustard X16 - 44.1 kHz H-FAST Impulse.png
    Gustard X16 - 44.1 kHz H-FAST Impulse.png
    99.6 KB · Views: 20
Last edited:
Top Bottom