IraSeigel wrote:I think I remember reading some time ago that the typical listener can't distinguish less than 20ms of delay.
But for the couple ms you're discussing here, I can't see what difference it makes, unless you're recording sounds separately in an anechoic chamber, 1 source and 1 mic at a time. Any other scenario will have lots of intrinsic delay and phasing.
I really wish whenever people would make blanket statements like these they would put in the proper qualifiers. Yes I would agree that you need at least 15ms (I say 15 rather than 20, as I can perceive that most days, if you're doing an A and B comparison anyway) before most humans can perceive the time shift. But that is really only the case if your'e talking about sound coming out of two separate speakers or two separate sources (think actors mouth and speaker here). If you want to test this out, take a mono source, play it out of your studio monitors at home and delay the signal in one of the two monitors. You will hear the image shift left or right. Typically when I'm at home in a controlled environment I can start to perceive that image shift around 12ms or so, depending on the source material, speaker placement and the phase of the moon.
It's the Haas effect in action.
Change the situation to a singer and their monitors, particularly with In Ear Monitors on and it won't take 20ms for them to hear what sounds like comb filtering, but it's combining in their heads. They hear the natural voice, the transmission through the bones in their skull and the monitors all combining. I'm not a voice performer so I have never bothered to experiment on myself in these kinds of situations. But I can tell you that if I combine a dry source with a delay source and listen in headphones or In Ear Monitors, it takes a lot less than 20ms for me to actually hear the phase cancellation happening. I'm talking about purposely combining a non processed and processed signal in some manner here. If you're smart and careful obviously most days you're not going to do that except for things like time based effects (i.e. reverb or delays) in which case you want that because it's part of the effect.
Of course having said all that, I remember sitting in on a session in the Live Sound Reinforcement workshop (used to be hosted by Syn Aud Con and Live Sound Mag) where Dave Robb was demonstrating this, using the exact example above, two speakers, using a time delay in one channel to "pan" a signal. I was blown away by how many in the room said they could not perceive the shift in image at all. Even some that were right in the middle of the listening area like myself. It was very obvious to me. Just shows how many "cloth ear'd" there are running around in the herd.