I should point out that for tinybox (and most other amps as well), the equivalent input noise is best at maximum gain; it's just that it's easier to hear the noise floor because it's louder. But you record at lower gain levels and increase level in post-processing, the signal to noise ratio will be worse; not necessarily because of digital dynamic range, but because of the analog signal-to-noise.
Also note that only adrianf and one other person I forget have high sensitivity couplings, which yields the best possible self-noise. And as I have said many times, PFAs always have 3dB worse noise than actives to tinybox--that is true for both my actives and the Nbobs, due to decreased polarization voltage.
I should add that my experience was not the norm. I've got pulls done before the high sensitivity couplings were done that were pristine. The show in question was a non-PA'd event where an artist was in a room playing and I had to be about 8-10 feet from the "stage." With +32dB gain on the TinyBox and an input level of 10 on the PCM-M10, I was still getting peaks that were relatively low for the performance part versus the applause.
At my last show, with the revised couplings, I had my M10 set at 4.5 with low gain/high power (leaves the couplings in a 'high sensitivity' state) and got a smoking pull. I ended up having to boost the gain on the deck to 5.5 to get peaks around -9/-10dB during the recording but mid-gain would've been too much for the show.
Again, setting the TinyBox to the appropriate gain setting BEFORE recording is always the best practice and the Jon's work is simply AMAZING. In fact, I'm that impressed by the TinyBox/Active solution that I also picked up a PFA active set with higher sensitivity couplings as well because the product is THAT GOOD and I've now got a P48 capable deck (which I can run two pairs on or board plus TinyBox). Thanks Jon for all of your hard work on this beast of a product and for coming up with a quieter solution.