Have yet to off-load and check sync of the files, so stay tuned..
Back to report. I transferred stereo files from both PR2s last night. tl;dr- Recording and control of multiple units by the Sidus app works. Sync is not ideal, but seems manageable
Upon transfer I noted the corresponding files have the same time stamp but very slightly different sizes. Hmmm, will that mean I'll need to align by timecode rather than simply aligning the start of both files?
Prior to recording using both PR2's together last Friday night I set the timecode on each recorder to 60 frames per second, figuring that choosing the highest available frame rate may serve to increase the potential resolution of alignment. I also changed the DID device name of each PR2 to reflect the channels being recorded, and thankfully that device renaming carried over to the saved file names. So I had two stereo files from the separate recorders: LR_4061.wav and CB_4061.wav
I don't have editing software loaded on the laptop I was using to do the file transfer, so figured I'd try an old trick I've long used to use to achieve quick, easy stereo playback of 4-channel baffled omni LRCB recordings previously made to Tascam DR2d. I renamed the stereo files: 1.wav, 2.wav, placed them in a folder with a .pjt extension, and copied it to the SD card in the Roland R44. Despite the slightly different file sizes, the files were successfully recognized for four channel playback by the R44.
The trick I use for quick 2ch stereo playback of LRCB recordings is to engage Mid/Side playback decoding on the R44 for the Center/Back microphone pair. The Center microphone serves as Mid, while the Back (rear-facing) microphone serves as Side, and that pseudo stereo pair output gets mixed in to taste with the dedicated L/R channels. Adjustment of the M/S ratio all the way down to zero mutes the Back channel entirely, so that when combined with the L/R stereo pair the result is a 3 channel L/C/R mix with C correctly panned to center. At that point I can adjust the balance between LR and C for best effect. That's 3 channels total. Once that 3 channel L/C/R balance is about right, adjusting the M/S ratio brings in Back rear-facing channel content, with its polarity inverted in the Right channel by way of the M/Ss stereo decoder. That diffuses and spreads out the single rear-facing mic channel content, effectively "stereoizing" it. Thats useful but not perfect because polarity inversion of the single rear-facing omni routed to the Right channel causes some low frequency cancellation on the Right and reinforcement on the Left, but its pretty effective in terms of producing a richer stereo image with a greatly improved sensation of the fore/aft depth dimension. Its effective and works well enough as long as not too much Back channel content is needed. To do it correctly I instead import the files into an audio editor and use a VST phase-manipulation plugin prior to the polarity inversion of the Back channel. That plugin allows for making any arbitrary phase rotation across the entire audible frequency range. Choosing a 90 degree rotation of phase, in combination with a polarity flip on the copy routed to the Right produces a symmetrical 180 degree phase shift between Left and Right channels, that actually consists of a +90 degree phase shift routed Left and a -90 degree phase shift routed Right. And keeps all low frequency cancellation/reinforcement symmetrical across both channels in addition to keeping the higher-frequency phase interaction symmetrical as well. Sometimes different phase rotations (symmetrical or not, making them fully cancelling or not when collapsed to mono) can be used to advantage when dealing with off-center low frequency content. Whatever works out best.
I'm pleased the files play as a 4-ch recording in the R44, even though the two files are slightly different length. With the rear channel muted (ratio set to 100% Mid / 0% Side) its the a straight 3-channel L/C/R input mixed down to 2-ch stereo output. Listening carefully I could detect a slight flanging like delay during the applause. The slight delay isn't obviously notable during the music, although improvement after full alignment might make it so. That slight delay indicates to me that the starting points of the two files are very close yet are not identical in time like they would be if recorded to a single 4-channel recorder. More critically, the slight miss-alignment does not seem to drift or change over the course of the recording. So..
My initial conclusion is that aligning the files in an editor by way of their shared timecode frame values, rather than simply aligning the start of the files, may be all that is required to fully sync files recorded across multiple PR2s which are being controlled by the same phone app.. But even if not, because there seemed to be no or at lease very minimal drift, the two files should at least remain phase-locked after a manual alignment of the two files been performed.
Will update again after getting them into an editor and aligning via timecode values..
