Hacker News new | ask | show | jobs
by deckiedan 1385 days ago
It gets really messy though when you're working with a band (usual case when a drum kit is involved), and multiple mics.

A drum kit may well have 3 or more mics able to "hear" each drum. The direct mic on the drum, possibly a phase reverse mic on the opposite side of the drum, and any overhead mics. Also, any nearby vocal and other instrument mics (if the drums aren't inside a massive isolation box), etc...

And then the speaker distance and direction come into play. Any stage monitors ("foldbacks" in UK English) are usually pointed back at the drum kit - but aren't silent in their reverse (eg towards the audience) and so their sound will be out of phase with at least something. Then the distances - the sound from the drum will be (given a stage size of a few metres) several ms delayed actually getting to the audience. How far are the speakers (and subs) from the audience too? And since subs are almost omnidirectional, what about spill from them back into mics on stage?

Also, if you're working with digital mixing, you have a few fixed ms implicit extra delay everywhere... So getting the mic and drum head to "actually" move in sync is... Unlikely. But then again, with digital you usually have delays available on every channel and can guddle around with that until it's sounds good enough...

Messsssy. In-ear-monitors can help a lot keeping the stage sound cleaner - but all IEMs I've used introduce some more delay, due to processing, analogue-digital-analogue conversion, etc. Which is only a few ms, but for a vocalist, can give the unnerving experience of feeling like their voice is coming from a couple metres away. (Technically the same as coming from a stage monitor, but when the sound is right inside your ears, it can freak some people out).

Live sound is fun! It's all compromise, all fighting the tensions of time against perfectionism against theory against budget against the show must go on, and keeping the musicians and audience happy.

That all said, in my experience, often phase things can be fixed by either flipping phase, or moving a speaker or mic around until it magically works...

1 comments

> Also, if you're working with digital mixing, you have a few fixed ms implicit extra delay everywhere...

This is just shoddy implementation... It's totally possible for delays in digital systems to be measured in microseconds... So to have delays of milliseconds is just some software/firmware engineer being lazy and buffering a load of samples together somewhere.

I think the issue is, at least in part, that a lot of audio gear is existing. I mean if a recording studio was built around a million dollar digital console in 2006, there aren’t a lot of business cases for replacement when the engineering solved the problem years ago and when the hardware for the solutions is already at hand.

My engineering intuition is that millisecond latencies would be easier to solve because there the hardware is easier to build than sub-millisecond delays, and because millisecond delays are common for keeping microphones in phase.

A standard solution for avoiding AD to DA latency is an analog mixer because it provides a signal path running at close to the speed of light.

YMMV.