(24-04-2017, 09:44 PM)mixogen Wrote: Well before you go too far with this ideology, it should be understood that the AES guys don't always understand things.
The monobloc guys won't appreciate it much either! (edit: I hope I don't sound rude here - I'm thankful we can have this discussion!)
Think of close tom mic and close snare mic. Those 2 mics BOTH pick up snare AND tom (as both mics pickup tons of bleed).
When you setup the mics you listen and make sure that snare sounds good and the tom sound good. Doesn't have to be exact perfect phase, just needs to sound great. Leave the mics all on during tracking and if everything sounds great, you have a good enough phase relationship.
After tracking you can decide to time align snare to overheads and toms to overheads. If you do this notice how the snare's tone changes and tom's tone change. That fat full thing is gone, and often toms sound terrible now and the snare sounds like crap due to the tom bleed. Zoom in on the tom's track and notice the snare transient is not aligned to the snare close mic. Nothing you can do, your tom bleed is EQ-ing the snare track in nasty way, and vice/versa. THE FIX: Gate the snares and toms. Its sounding much better, but that fatness you had is now gone, its weaker, so you go looking for samples to replace/assist in things. This works because we gate, and only one mic at a time is on (and also the blend with overheads isn't often same volume, ie the close mic is often louder and jumps over the OH).
Now imagine XY mic. One mic is closer to the left side, one to the right side. The bass will reach one mic before it reaches the other. Whatever instrument is on the other side of the bassist will reach the other mic first. That's just the way it is, same as close tom/ close snare. But XY the distance is much shorter, so the high frequencies are what to watch out for when adjusting time.
Sample rate 44.1:
A frequency of 5,512.5 has exactly 8 samples to represent the actual wave shape when it was captured. At this frequency you move one of the XY then you have a 1in8 chance of landing in perfect phase, and if you are only 1 sample off then you are 45degrees out of phase. If you are 4 samples off then you are at a null. I chose that frequency because its easy math. Each sample represents 12% of the wave shape.
Now the rest of the frequencies around 5k are not easy math. The samples for 5k or 6k or 10k or 15k are all fractions of a repeating wav, each cycle is gonna give you a different peak and null. This means you will probably never find perfect phase again when you shift the file over. The vast majority of frequencies up high will be out of phase minimum of 12% or much more, in a random weird way.
This will represent itself at highs that sounds out of phase, non-harmonic, ugly, dull, smeared, and simply un-natural. And that's what I heard when I tried it on this project, though the bass imaging was more centered, just didn't seem worth it.
Now if you look at lows, time aligning is not messed up because the capture allows so many more samples.
A 44Hz wave has 1000 samples. If you are off a few samples, won't be noticeable at all, you can dial in the phase to less than 1%. You can line it up in very detailed manner.
Trying to line up the higher frequencies as good as they were when they were captured, its just not gonna happen. The original capture took from full resolution actual waveforms, and the engineer could get it better than we ever can.
Anyway, it gives me an idea for this track.
Maybe one could time-align the XY and low pass it at around 150Hz? This would effectively make the XY a bass mic and use the close mics for high end sounds?
This might work!
As I said before, if it works for you use it, if not move on. This thread has helped me realize that my knowledge has grown far beyond the average mix engineers knowledge. I think my Benchmark DAC2, B&W 805n, and 300-watt mono amps allowing me to hear much deeper into the music helped that growth. Money well spent for sure.
ALX