The gain issue is still not solved adequately to my mind. I cannot fathom why the designer is introducing such significant potential problems for the sake of a few dB of SNR.
I don't think this is nearly as big a deal as people are making it out to be, but, I do think there are two different issues that have been aggregated into one. The first issue is whether having an input limit is an inherently bad thing, and the second is whether the input limit of this amp is too low.
1. To address the first, all amps have an input limit. This is particularly noticeable in an amp with an input transformer where it will saturate if the signal level is too high. But, it is also true with something like a B22 where a very high level input could damage the pot (and in some cases, such as with a stepper, that level is exceedingly high, but it does exist).
So, input limits are inherent in designing a headphone amp. But, with an IPT, or the opamp before the volume control, there are concrete objective benefits to doing it such as CMRR, isolation from ground loops, and lower SNR. There are also trade-offs, such as the fact that badly behaved sources (those with too high of an output) may not be compatible. The fact that not all sources are compatible with all amps should not be a surprise to anyone, particularly in an area where there are no hard and fast standards. Indeed, I try to keep my sources at 1Vrms or lower which works well as I tend to transformer couple most amps. Further, while this amp may not be able to handle a high source level, it seems to be, with perhaps some tweaking, able handle a higher source impedance than many other amps.
2. The second issue is that this amp may have a particularly low signal input limit. OK. So, as I say, not all amps are useful in all situations. Use something else. This is a different and new-ish topology to the headphone world. Like all circuits, it has benefits and drawbacks, and will work in only limited situations. This does not make it any different than any other amplifier out there, AMB's designs included. The flaw here seems to be not in having a limit, but in the designer's contention that it is perfect and better than everything else in every situation. Clearly this is not the case, so, use it where it is appropriate. In places where it is not, use a resistor divider (or another pot) in front of the first opamp, or use a different transformer to increase the power supply.
Finally, so I can avoid posting in this thread again, for an amp with a lot of feedback, the THD numbers don't strike me as being off the chart. They are fine, but not spectacular. My soundcard does better than that, and it sounds terrible. And, that's the last issue -- contrary to what at least one poster believes, not much can be gleaned from the numbers posted in the first post. All you can say is that the amp is working within expected specs. There is a lot about audio design, how an amp reacts with a real world load (back EMF injected into a feedback loop, for instance) and human perception that is not captured in those measurements.