I can see this being a small issue for singers w/ IEM's ("head sound" combined with IEM's), but any instrument and "feel" is not likely to be influenced by such a small delay (equivalent to standing 2' in front of your guitar amp).
Try flipping the phase and see which way the singer's prefer their vocal mic's sound in their IEM's. Might be all it takes. A variable phase tool might even help to get their vocal "head sound" working constructively (in phase) with their IEM's. I believe this phenomenon is generally more about phase relationships (frequency domain) than latency (time domain and "feel/timing").
If you consider that people often mix monitors through a digital console, the latency is likely to be greater than 2ms even with something like a Venue or PM5D. Then add floor monitors that will inherently add more delay (speed of sound through the air to your ears) and 2ms is practically nill by comparison (but floor monitors aren't blocking the singer's acoustic sound from reaching their own ears like high-isolation IEM's generally will)...
MADIface-XT+ARC / 3x HDSP MADI / ADI648
2x SSL Alphalink MADI AX
2x Multiface / 2x Digiface /2x ADI8