Since the modifications so far have been confined to just the removal of C13/C15 and new opamps (NE5532's), I didn't expect such dramatic changes but testing during the weekend went much better than expected.
1 - The noise level of the unit is very low now. There is no audible hiss throughout the volume settings. I verified my impression by recording the output of the PR8 into my MOTU 1224 and 2408-mk2, then checking the noise profile in Adobe Audition. Results were sound-checked with Audio-Technica ATH-40fs headphones and Alesis M1 active MK2 speakers with a 15-inch subwoofer crossed over at 75Hz through a DBX 223XL active crossover.
2- I recorded my singer through the Roland MMP-2 using a Studio Projects B3, and through the PR8 using an ADK A51 type V set up next to one another for simultaneous capture into the MOTU. I'm familiar with the sound of these two mics and it allowed for fast/simple impressions. The results were very good. The PR8 in its current state is easily a match for the MMP-2 in terms of clarity and sparkle ( I hate using these wine-tasting terms to describe sound).
3- Without recording, I had a friend play his acoustic guitar with built-in pickups through a Behringer DI20 running as a splitter to send output to both the PR8 and the modified ART-TPS. As I switched the custom room speakers (I'll describe these at another time) between the two preamps, there were subtle tonal differences in the 200-500Hz range that I think people describe as "roundness". Occassionally we'd perceive a "roundness" advantage in the PR8 over the TPS, but it was never anything that woud push a strong preference either way.
It's very encouraging to see that such a simple change can yield such satisying results. Now that I'm happy with the first steps of the journey, I'm emboldened for more! I ordered the Cinemag transformers on June 24th, and I'm thinking hard about using a gain-setting network between Q1/Q2.
I learned here http://www.geofex.com/Article_Folders/potsecrets/potscret.htm that a reverse-log pot has a curve that can be emulated by properly selecting a resistor across two of its terminals. If someone can guide me with selecting values for the pot and resistors, I'm ready to buy parts!
Concurrently with installing the gain network between Q1/Q2, I guess the output of C8 gets connected directly to R25 (elimnating R24, R27) and the value of R26 gets reduced to something like 10-20K, correct?
I'm also thinking that I'd like to optimize two channels to run ribbon microphones (the Apex210/Nady RSM-2) that I hope to buy through another member of this forum. I suppose that those channels need a very low value resistor between Q1/Q2, and some other values around U2B. All comments appreciated here!
Thanks to all, this is going very well.