In this case, he specifically said the following in the conclusions:
I agree with him in general on this, but unfortunately, there will be some who missed the underlined (I did that) part and thought the X3800H's higher distortions would not be good audibly speaking. It would have been nice if he would do like what you had done in the past with D+M, and Yamaha, that is, while letting them know there are things they needed to improve on, also emphasized to the readers that sometimes the relatively higher distortions would still be below the threshold of audibility and still made appropriate recommendations in a more balanced way.
As
@lovinthehd alluded to before, Amir did often mention something like, "...not audible..", or "...but thankfully levels are not high enough to be an audible concern...", but people tend to focus on the more dramatic comments such as "..not acceptable.." and missed the nuance.
Since you guys have a communication channel with the mass producers like D+M and Yamaha, they will listen to you more and it is never a bad thing for them to set their goals higher in terms of FR, THD, DR, XT and IMD and not simply aim to clear the bar of threshold of audibility, unless in doing so they had to raise their price point much higher. In this case, if they were to even just use the ES9006 that was widely used in Yamaha's non flag ship AVRs, SINAD would have no trouble clearing at least 90 dB and much fewer people, even Amir would have complained. So I do think D+M dropped the ball on this. They likely know it too, because they are using the ES9018Q2M (a real reference class Sabre DAC) in their flagship Denon AVR and the $7,000 Marantz A10. ES9018Q2M: 116 dB SINAD, 127 dB SNR (probably mono, so expect a few dB worse for stereo). Their bench performance will most likely be limited by the volume chip that cannot do better than 100 dB SINAD.
The biggest point I disagreed with him was the wording on "clipped at 1.4 V..." sort of remarks on his SINAD vs pre out voltage. To me that is false, he used the term "clip" simply because at that point SINAD (Denon/Marantz AVRs etc.) dropped rapidly but missed the fact that the rate of degradation actually would level off and still managed above 75 dB (0.018%) at 2 V and higher, indicating that wasn't due to clipping, but as you explained before (if I understood your point correctly) that the drop from that point was due to the internal power amps clipping, injecting much higher distortions. I further assume that the levelling off at higher voltage (about 1.7 V iirc) might have been because the internal power amp's influence were via some sort of feedback loop or capacitive coupling so the pre out itself won't be influenced in an unlimited sense.
