Phase-Coherant Drum Mic Placement

[christiannokes]

I'm not sure that I can tell if any number of my drum mics are less than severely out of phase.

I have been reading countless articles from professional enginneers talking about how correct phase between drum mics [

is your drum sound

.

It sounds like they flip phase reversal swithches and listen to what sounds best. Is this what a lot of you guys do?

What if You didn't have phase reversal switches? I don't have one, and I don't know what to do. I am using garageband, and it looks like the only thing I can do is time alignment. But in increments of like 20 ms or something. Any ideas?[/u]

[kakumei47]

make a mic cable and wire one end normally but wire the other one with pin 2 and 3 switched. or just take an existing cable and open up one end and switch them. you can buy phase reversers too. they're good to have around. mark this cable clearly as it will f you up later if you don't.

set up drums in mono first. bad phase issues will be more obvious than if everything is panned.

see how mics affect other sounds. for instance, start with just overheads. then listen to the cymbals with the tom mics in too. does the cymbal sound change? if so, use phase reverse or move mics. cymbals and snare are the things i usually need to worry about most on drums.

[drumsound]

I like to use one speaker mono. I just pan all the drum mics to one side. It's easiest to hear that way. The thing to listen for is the low end getting weird or disappearing. Often I check just the OH mics and then the room. Then the OH and Room against each other. Next I add in the BD mic(s). If I'm using 2 BD mics (usually) I check those without other mics and then work with them as a unit. So if I have the outside BD phase flipped, when I'm checking the OH to the BD I'll flip both BD mic to keep the BD relationship right. Then I add snare, then toms. Basically however its sounds the biggest and fullest with the most frequency extension is where things stay.

[Professor]

Phase reversal switches will only put one particular back into alignment while shifting others out.
The reason you are having phase issues at all is because the sound from each instrument is arriving at different microphones at different times. The sound travels away from the instrument at the speed of sound but then is converted into electricity at the microphone and travels at closer to the speed of light. Since the speed of sound is relatively slow, that subtle time difference in arriving say at the snare drum mic an inch above the drum and the overhead mic 3-4 feet above the drums appears as a delay of about 3 to 3.5 milliseconds. Let's say it's 3.5ms, in that case the snare drum frequencies right around 142Hz will be out of phase between the two tracks, while the snare frequencies around 284Hz will be in-phase, and the frequencies in between will be, well, in between.

The real, true, best solution is simple time alignment of the recorded microphone tracks which simply means picking the mic that is furthest away (most delayed, usually the overhead) and lining all the others up to match. You can do that with a delay on all the other drum channels or you can do it by visually aligning the waves on a DAW screen.
The catch is that you are using Garageband, and it's not designed to do that. It's designed to let you play around with making music, but not designed to be a full-featured DAW to replace Logic Audio.

I would suggest a couple of more practical cures for your phasing issues:

1- Use fewer microphones.
The phasing is occuring because sounds are arriving at different times at different mics, so if you use fewer mics you will reduce the possible places where phasing will happen. It's not really necessary to mic every tom, every cymbal, have two mics on the snare, or whatever. Perhaps try slimming down to a basic Overhead pair plus Kick and Snare and only work your way back up based on what you need to hear more in the sound.

2- Get the mics closer to their sources and further from each other.
In that example I gave for the snare drum in the snare mic and overheads, I only considered timing, not volume. Yes, the mics would be out of phase at 142-Hz, but the difference in distance from 1-inch to 4-feet means the overheads are about 5.5 powers (doubled 5.5 times) further away from the snare than the close snare mic. The "inverse-square law" (look that up) says the volume will be 1/4 the power for every doubling in distance, or down about -6dB for every doubling in distance. Practically speaking, don't worry about the math, but know that the snare drum sound will be much quieter in the overheads than in the snare mic - although you'll temper that a bit with a less sensitive snare mic (and probably less amplification) and a more sensitive overhead (probably with more amplification). If you get them right back up to the same level, then you're out of luck there, but if you are using similar mics at similar levels, say like snare and 2-3 toms all miced with dynamics that are amplified the same, then this distance thing will really help. You'll find it described as the "3:1 Rule" or the "3 to 1 Rule" when you search for it, but that is the worst possible name because it is very misleading. I call it the "Greater than 3 to 1 Recommendation" because it ain't rule, and while 3 is the magic number, it's not the only number. But I've written a lot about that on the board before.

3- Use different mics and know where they are aiming.
I'm going out on a limb and guessing all you mics are cardioid. You may even be aiming them 'carefully' so the front side hears what you want and the backside ignores something else. Here's the catch(es) - The more directional a microphone is, the more it causes phase issues in stereo mixes, and cardioid mics don't reject everything behind them, just a very small point directly behind them, but they still pick up sound 90? off-axis, and even 135? off-axis. What you might consider (if they are available to you) is to use omni-directional microphones above the drumset as overheads instead of cardioids because they will not sound as phasey (long explanation of how that works) perhaps tighter patterns like hyper-cardioid or fig-8 up close while paying careful attention to what they are rejecting.
This last option takes a bigger mic locker and more experience to work out, but is pretty effective. In the mean time, the first two will serve you best.

Hope that helps you at least some small amount.

-Jeremy

[cgarges]

Sorry Jeremy (and any other time-aligning advocates), but I simply won't ever buy time aligning as a solution to getting a drumkit to sound natural. Especially not in place of simply checking polarity relationships. Time aligning is (as far as I know) a recent fix for an intersting problem, but one that sounds unnatural and sometimes casues more problems than it fixes.

If you align all your drum mics to a room mic that is, say as close as ten feet from the kit, that's still approximately 10ms later that your drum sound is occuring in relation to how it was played in context. That's a noticeable change in the feel and precision of what was actually played.

In addition, nobody hears a drumkit like that in real life. Room sounds are heard at a different time when you're sitting behind a kit and a kit sounds different if you're standing 20 feet away from it. It's those subtle differences in arrival times and locational cues that can give a recorded drumkit a sense of space and dimension, especially when recorded in stereo or when attention is paid to any sort of binaural listening principals.

Getting the polarity relationship of your close mics is a much simpler solution to a natural and full drumkit sound. It's one that we've all been listening to for years. It's a solution that works.

For those of us who have polarity switches somewhere on our input stages, it's a cinch, because it smply takes a few seconds to switch stuff back and forth. Like Tony said, get any relationships between multiple close mics correct (two or more mics on any one instrument), and maintain those relationships while auditioning settings in the context of the entire kit and you're set. If you don't have polarity switches or for some reason it's a bigger pain to do, than take a look at the more obvious candidates. There will most likely be some funniness between the overheads and either the kick drum or the snare drum. It's worth the time to listen to your sources with either the kick drum mic(s) or snare drums mic(s) flipped and chances are, you'll hear something that sounds better than the other choices. Go with that and you'll be fine, but do take the time to listen to those choices so you can make the right one.

I will, as usual, give Jeremy total credit for an awesome set of answers in the options he gave you.

Chris Garges
Charlotte, NC

[dirk_v]

Sorry Jeremy (and any other time-aligning advocates), but I simply won't ever buy time aligning as a solution to getting a drumkit to sound natural. Especially not in place of simply checking polarity relationships. Time aligning is (as far as I know) a recent fix for an intersting problem, but one that sounds unnatural and sometimes casues more problems than it fixes.

If you align all your drum mics to a room mic that is, say as close as ten feet from the kit, that's still approximately 10ms later that your drum sound is occuring in relation to how it was played in context. That's a noticeable change in the feel and precision of what was actually played.

In addition, nobody hears a drumkit like that in real life. Room sounds are heard at a different time when you're sitting behind a kit and a kit sounds different if you're standing 20 feet away from it. It's those subtle differences in arrival times and locational cues that can give a recorded drumkit a sense of space and dimension, especially when recorded in stereo or when attention is paid to any sort of binaural listening principals.

Although I can definitely see both sides of this argument, I think that Chris may have inadvertently reinforced my belief in time-aligning drum tracks here. Ironic, isn't it?

For me, the ideal situation is to try and get a good picture of the kit from the OH mics, and then use close mics to reinforce as necessary. I think it's akin to mic'ing an orchestra from the 12th row, and then spot-mic'ing soloists or other instruments which need reinforcement or focus. As Jeremy and Chris both explained, hearing a drum hit from the same close mic and a mic 10' away = 10msec timing difference. Assuming that you're using your "big picture" mics as the majority of the kit sound, it makes more sense to me to make up that 10msec difference to properly reinforce the drum in the "big picture" sound.

Of course, this all becomes much more complicated when you're relying more on close mics for your drum sound (or if you've got a lot of channels up for drums period), and in those cases, I find that time-aligning tracks becomes a very slippery slope that falls into the category of "causes more problems than it fixes." Once you've got track A slipped in phase with track B, then it's suddenly out of phase with track C. And so on. I believe this is where things can also start to sound unnatural. The phase switch is now your friend, and the advice given above is right on.

I guess it all comes back to how you like to work, what's right for the song/session, and "if it sounds right, it is right." Each time I record drums I learn a little more, and right now I'm pushing myself toward the school of using fewer mics, leaving less flexibility in the sound down the line, but hopefully getting the right sound to begin with.

-dv

[cgarges]

Although I can definitely see both sides of this argument, I think that Chris may have inadvertently reinforced my belief in time-aligning drum tracks here. Ironic, isn't it?

Sweet.

I think it's akin to mic'ing an orchestra from the 12th row, and then spot-mic'ing soloists or other instruments which need reinforcement or focus.

Right. Try time-aligning a close mic and an ambient mic on a classical soloist and ask them if it sounds weird to them. If they say it sounds weird (and I suspect many would) it's probably because there are supposed to be timing differences in close and ambient sounds.

If you see the timing differences as odd, you can either see it as the close sounds being early or the ambient sounds as being late. The ambient sounds are supposed to be late. That's how the world works.

I'm not saying it's wrong, I'm just saying it's wholly unnatural to align that stuff.

While I totally see why someone would think this is a good idea, and it seems good on paper, I think (and this is just my opinion) that it falls apart in practice.

Again, that's just me.

I guess it all comes back to how you like to work, what's right for the song/session, and "if it sounds right, it is right."

Thumbs-up!

Chris Garges
Charlotte, NC

[christiannokes]

I just found a way to time align the tracks on Garageband. I set the Apple delay to 100 percent and tried 1-5 ms for kick and tom mic.

I see what both of you are saying. And to be honest, none of my drum mixes have ever sounded so clear.

I really need to get some phase switches so I can compare. It looks like my kick and maybe my tom was out of phase.

[JWL]

I'm a big fan of the Recorderman method, which is a variation on the Glyn Johns method. This gets you a solid foundation of the entire kit. You can make sure the mics are in phase for both kick and snare, and fine tune for cymbals.

From there, what do you need? Add it in a way that sounds good in context. I'll usually add kick, snare, and a crotch mic, along with a room mic or three. Remember the 3-1 rule.

You might want to time align some tracks, but I agree that time aligning room tracks may not be the best thing to do in some cases. A snare closeup with overheads generally seems to work from aligning. Same with the kick. Room mics, not so much. Generally, I'll get it as good as I can with phase switches. Then I'll save a new version and play with time alignment. I usually go with the time aligned version.

[Professor]

Sorry Jeremy (and any other time-aligning advocates), but I simply won't ever buy time aligning as a solution to getting a drumkit to sound natural. Especially not in place of simply checking polarity relationships. Time aligning is (as far as I know) a recent fix for an intersting problem, but one that sounds unnatural and sometimes casues more problems than it fixes.

Well that's all I can stand of your sassy talk and non-believing there Mr. Garges, and if I ever catch you 'round these here parts again...
well, and if I knew what you looked like...
and if I really took any of this personally...

nah, I guess we'd just go have a beer.


So yeah, it's quite alright to go the route you prefer, and if the gigs are coming in then whatever you're doing is "right".
The time-aligning thing is interesting to me but something I came to in a strange way.
I was certainly familiar with the concept since I started in this game from a classical background. I don't know when (if ever) I get to actually start calling myself a "Tonmeister" but that's kind of the pattern I follow. And I gotta say that kinda like Dirk, your objections actually reinforce why I even consider it. (Oh, and I don't mean the playing classical music, but the recording of it, since I know you were a percussion major at Miami near the same time I was a percussion major in Boca Raton... crazy.)
If I have mics 15-feet above the conductor's head, then even though I've altered my distance ratio of violin-to-mic versus trumpet-to-mic so the inverse square law isn't as much of an issue, I would still have an issue with timing if I put a spot mic in front of a soloist or weak instrument. Yes, I want that perspective of hearing the orchestra from that single point some distance away, but if I add in a little of my spot mic for presence then it will also be arriving earlier than the main pair. So I would be inclined to delay the spot mic not just to align with the main pair, but possibly to be a hair later so the precedence effect kicks in and my brain thinks it hears a stronger soloist from the same distant mic pair, and not an earlier soloist.
I'm not sure if that made sense - I'm typing really late again.
I then consider the same situation on a drumset. I aim for an overall kit sound from the overheads that I goose with a little presence boost from a couple close mics.
Indeed I even eventually got around to noticing that the overheads are out-of-phase for the initial impact of each drum stroke. The snare head moves downward, sucking air away from the close and OH mics and generating a negative voltage in both, but I don't want my speakers to start each snare pop by sucking air away from the listener. The bass drum mic that most guys consider "out-of-phase" with the overheads, is indeed phase-correct for nature, so when I do time-align, I also invert the OH mics.
So yeah, I use the phase switch too.
And actually I don't like aligning on the DAW screen. I measure the timing there, then dial in delays on the Yamaha console, and then listen to the results in stereo & mono, tweak a bit, try to find the right spot, and if nothing improves (or if it gets worse) then I skip it and chalk it up to following those 3 other fixes I put up there, since really those should come before time-aligning anyway.
Oh and as far as personal style, I never delay by as much as 10ms, it's rarely ever more than 4-5ms at the most. That's because I don't typically use room mics, and if I do, I leave them late because that's their purpose - to add reverberation. But I do align the snare, kick and any tom mics to the OH mics. Yes, that might make them contradict each other, but I also cut the tom mics unless they are playing, and the tone of the kick in the snare mic and vice-versa is different enough that phase cancellation would be unlikely anyway.
It's also not something I applied to drums because of the theory. In fact I resisted it for a long time. But then I just kept seeing the same patterns and fighting with some of the same sound issues regularly, and I started considering the logic of my approach as I was fighting an uphill battle against the laws of physics.

But I'm rambling again, sorry... lateness.

I'll hit it in key points:
- Time-aligning takes some practice and understanding of the physics involved and should be a fine tuning step after you follow those other cures I listed.
- You don't necessarily have to delay the close mics to match the distant ones. In the old days, that's what was done because you could only delay, not anticipate. But in a DAW you could pull the OH mics early to match the snare. So you can time align the whole drumset relative to itself and also choose which direction it might go to align with the 'reality' of the drummer's intent with the band.
- Absolute phase is still an issue - you can't just match the first positive peak in the kick mic to the first positive peak in the OH mics as that's probably the 2nd half of one of the waves.
- Oh yeah, and you don't need to do any of this.
If you leave it as is and it sounds good, you win.
If you use phase reversals and it sounds good, you win.
If you drop in a few delays and it sounds good, you win.
But if you do any or none of those and it sounds bad, then well, you lose.

-Jeremy

[JdJ]

The thing I don't quite get in a DAW is how the image of a wave is supposed to represent the entire frequency range of a sound comprised of so many frequencies (unless it's a sine wave). Following this logic, if you align the waves on the screen, you are only aligning a visual representation of the sounds, not the sounds across their respective frequency spectrums. So you are either (depending how the DAW arrives at it's image) aligning one frequency or an average of a bunch of frequencies. Either way, you still will have a bunch of frequencies that are out of phase, right? If you are going to try to time align mics- it seems to me using the one speaker mono thing is the way to go.

There was a point in time when I would align all of the mics by site during mixing, which did make a real punchy transient-aligned sound. Of course, this also creates a bunch of other problems- if all of your drum transients are aligned- you lose any masking, and essentially reduce your headroom by quite a bit. And for the reasons I mentioned above, there are still phasing issues going on despite the big thwaak! That's how I have come to doing things as Drumsound and Chris have said- and I am much happier with my mixes. Occasionally, if I am still having issues after tracking, I still have to do some bumping of regions, but I only use the visual representations as a basis, bumping things either way until they sound right.

Cheers,

J

[MoreSpaceEcho]

there are supposed to be timing differences in close and ambient sounds.

+1 to this and everything else garges wrote. your overheads are supposed to be a couple ms behind your close mics. your room mics are supposed to be a few ms behind the overheads. i mean...obviously. i dunno why this is perceived to be a 'problem'. it's just the way it is. every single time i've tried the time aligning thing it sounds Completely Unnatural to me. which is fine if that's the effect you're after but i really think if you need to do this to get a 'punchy' drum sound then you kinda suck. i don't really even understand why anyone thinks time aligning is necessary...i mean every record everyone loves has drums that almost guaranteed WEREN'T time aligned...is it just cause now you can see the waveforms so 'oh we CAN do this so we SHOULD!'...

aimless ranting, my apologies.

[christiannokes]

What kind of phase reversals are best? What is the price range? I don't ever recall seeing them in music stores.

[MoreSpaceEcho]

just take a cable you already have and reverse pins 2 and 3. instant phase reverse cable. free.

[eeldip]

i really think if you need to do this to get a 'punchy' drum sound then you kinda suck.

tee hee. i've been time aligning and using drum replacer instead of EQ for all my recent mixes.

i am loving it. screw realism.

i punch YOU.