The entire system used during our cable evaluation was new Benchmark products, from the source to the amps, which I understand to be ( prove me wrong ), the standard AES48. And if I am wrong, so be it. I have been wrong before. So, given the " cable differences might be greatly reduced ", which is your wording, are you suggesting, that differences in cables, under these conditions, are NOT audible at all ?There are very slight differences between the very best cables and the very worst- not enough to write home about; the balanced standard is very effective!
As best I can make out the Benchmark line drive unit does not support AES48. If a preamp has both balanced and single ended outputs that run at the same time, without a switch to go between the two, then its unlikely to support the standard.
This is because you can't operate both balanced and single ended off of the same output: they are mutually incompatible. What's happening with the Benchmark line drive (which is similar to their other units that make a balanced output) is that the balanced output references ground. That's a violation of the standard (and is quite common throughout high end audio).
The following article is from the Benchmark website. If you read through it you will see that AES48 is not mentioned, although near the end of the article AES3 *is*; clearly John is familiar with the AES:
https://benchmarkmedia.com/blogs/application_notes/balanced-vs-unbalanced-analog-interfaces
It is also apparent from this article that John knows his stuff! I suspect he is aware that his line stages don't support AES48; otherwise he would mention that somewhere on his site as otherwise he makes all the right arguments supporting differential amplifiers and the like.
Now the argument has been made that if both outputs of the XLR connection reference ground, but the ground currents are identical, then they cancel. This is true but now the cancellation has a number value attached, because in the real world they won't/can't be identical. This is the same argument that some people mistakenly make for an output transformer (driving a balanced line) needing a center tap to reference it to ground. The reason this is not actually a practice is because the center tap can't be perfectly placed and so will degrade the CMRR performance. It also means that there will be ground currents in the cable, and further means that now the construction of the cable has become critical to its sound.
So I'm not doubting what you heard- I've experienced exactly the same thing. I think what's going on here is you were under the impression that the Benchmark line drives support AES48. That would explain about 95% of our exchanges on this thread!
BTW I miss Al too- he retired a few months back and has been far less active here ever since. I really appreciated his comments!