Some new information from Roland below this reminder quote...
This is odd tho, I noticed some faint banding in the R09HR 100% mic spectrogram...
Looking at the histogram the spikes are visible, but on closer inspection they fall on the exact frequency intervals (1000 Hz increments).
So that's really wierd ...any suggestions? I can't imagine any real digital noise that would fall so neatly on the frequency divisions? An engineering mode in the firmware left on? I am using 1.04 ... what am I overlooking here?
digifish
AND guysonic's review...
http://taperssection.com/index.php/topic,106364.0.html show some strange low-level digi-noise in the mic-in path.
So I have heard back from Roland R&D Japan -
They note (as we know) that this noise is only an issue on the mic-in. They say that the specifications of the unit take this into account and that due to the small circuit board and design constraints (cost?) they found it impossible to remove this signal from the mic path. Further, it is mainly outside the audible range (> 20 kHz) and so is not considered a big issue in a recorder of this nature.
The statement that it's mainly outside the audible range is true if you record at 88.2/96 kHz but not at 44.1/48kHz, so I thought that a bit odd. Similarly why release a recorder with > 48 kHz sampling if you don't believe it is audible up there, ahhhhh the pressures of marketing requirements on the R&D department
Just a reminder the line-in is free of this noise (and that on the HR mic-in, it is at a very low level, see guysonics graph above and so not audible under any circumstances I can find, including normalizing silence). As is clear the R09HR is far better overall than the R09 for mic-in specs that also include similar digi-noise.
digifish