Innerfidelity's measurement graphs of GS-X MK II - am I reading this right?

[riceFET]

TL;DR below wall of text.

I'm a noob when it comes to actually reading the graphs with equipment measurements so I'm hoping some of the folks here can enlighten me.

For the past few days (well, weeks and months actually, been lurking here since November last year) I've been looking at various measurements available at different sites and found Innerfidelity's measurements of the GS-X MK II, I haven't done a super deep search so it's the only one I could find so far.

If I'm not mistaken this is basically a high end implementation of the Dynalo. Its asking price of 3000 USD certainly no laughing matter. Now I know that we're at a point where excellent products are available for much less so I compared it to the O2's measurement and found the following parts of the single page PDF to be of interest:

The tabulated data, if I'm not mistaken, lists the clipping point as volt/watts THD+N @ 1%, which, when looking at the O2's graphs, seem nothing noteworthy. The voltage right before clipping happens below 0.01% THD+N.

A similar thing can be seen in ASR's graphs for the JDS Atom, Massdrop THX 789, Neurochrome HP-1, and the RME ADI-2 DAC, where the line goes down gently then rises steeply, indicating the device has clipped.

Not the same story for the GS-X graphs. I highlighted the areas of concern for me. We can see that distortion rises logarithmically to above 0.01% past 2 V and on 16 ohms, the steep rise happens close to 0.1%. Higher impedance loads seem to do better, but this doesn't look good to me especially for a device that costs 2.999 kilobucks.

So my questions are:

1. Is there anything wrong with my interpretation of the graphs - that the GS-X MK II has distortion values that are quite high past 2 Volts? And that such values are unacceptable at its price level?

2. Am I even reading the graphs right and tables right? Are there any nuances that I might have missed?

Thanks to anyone who answers.

TL;DR:
Am I correct in my interpretation that Innerfidelity's measurements show that the GS-X MK II has relatively high distortion values past 2V? Refer to the image posted above.

 

[amirm]

That's correct. While not common, a number of amps have a primary source of distortion that kicks in before they ultimately clip. This is not good, especially at high levels in the GS-X MK II. No way any good amplifier should have that kind of distortion.

If you want to spend $3000, then the Benchmark HPA4 is for you. If you want to spend less, Massdrop THX AAA 789 although it is not available all the time. Finally at $99, the JDS Labs way outperforms both of the amps above and also represents state of the art.

 

[riceFET]

Thanks Amir! I actually saw the PDF summary of the measurements months ago but never really bothered to look at the graphs since I couldn't understand them at the time. This place certainly helped alleviate that.

When I bothered to actually read the graphs and understand them I was shocked because reading reviews of that thing one would think it has immaculate performance (I know subjective impressions don't mean much, but still). Learn something everyday, huh?

And no, I'm not looking to dump 3K on a headphone amp, at least not right now, but I have been looking around all over the place.

 

[amirm]

Yeh, it is common for other sites to praise products despite poor objective performance.

 

[stevenswall]

^This is especially common with Rafe who took over InnerFidelity when Tyll retired. Tyll took measurements, related them to reality, and gave his impressions. Rafe... Does not do that. (It's more like a story that gives you a certain feel for the 'audiophile article' aesthetic more than data.)

Glad some measurements are being done though.