Double miking acoustic; placement and mix techniques

[Doug Young]

Originally Posted by upsidedown

I like to double mic. I think it adds a dimension to the sound.

Since updating to GarageBand 10, I haven't been able to find the button for recording a track in stereo anyway. Too bad; especially seeing as how all the advice you guys have been giving me deals with stereo!

But I have been recording in mono and as Doug said, creating a stereo effect with pan and volume on the multiple tracks.

You can always record to two different mono tracks, and with panning hard left and right you have stereo. Same thing, but many DAWs support a stereo track that just hides the details from you, and also makes it easier, since plugins can be applied to the stereo track all at once, etc. I think garage band supports stereo tracks, but I don't use it, so I'm not sure how to set it up. It may depend on your hardware interface.

The whole thing of combining mics gets really "interesting". I always record stereo, but I have been using at least 2 pairs of mics, and lately have been experimenting with 3, so 6 mics total. That means I end up with this same issue that mics get combined, even tho my result is stereo. It gets tricky, and phase relationships can be a huge problem. I have a plugin that helps me micro-align tracks to avoid some of that, tho part of what you get when you mix two mics is phase cancellation, which can be bad or good - just depends on how it sounds. I like combining different kinds of mics, I'm usually combining a condenser and a ribbon, for exactly what I think you mean. You get the different colors of each mic. Sometimes it works. But I'd think the jump to stereo and some standard stereo micing techniques, would be a far bigger improvement than mixing two mics in mono. Of course, in the end, it's what works, there's no truly right or wrong in recording, only what sounds good and what doesn't, and that's largely up to you as long as you're happy with the sound.

[rick-slo]

Providing a link to one or more of your recordings (preferably without reverb being added) would be informative.

[Wyllys]

Originally Posted by Doug Young

I have a plugin that helps me micro-align tracks to avoid some of that, tho part of what you get when you mix two mics is phase cancellation, which can be bad or good - just depends on how it sounds.

This is a very sticky area. What exactly is the applicable reference point for this micro-alignment? I'm assuming the alignment is in the time domain, but at what frequency? Or does the plug-in simply align the zero-point?

Yes, the bottom line is how it sounds. A good way to check up on things is to collapse any two of the microphones dedicated to picking up a single source to the center. Any gross phase anomalies will be instantly and glaringly apparent.

A coherent two mic/element array is arguably that with the fewest negative phase issues. The standard or classic arrays are well documented. In fact, photos of them are included in the link I posted earlier in this thread. Going to anything beyond a pair is getting into pretty deep water. Suffice to say that the old axiom "the right mic in the right place" is still the best guideline.

[Doug Young]

Originally Posted by Wyllys

This is a very sticky area. What exactly is the applicable reference point for this micro-alignment? I'm assuming the alignment is in the time domain, but at what frequency? Or does the plug-in simply align the zero-point?

Absolutely. Phase is totally frequency dependent, there's no free lunch. I don't know exactly how the plugin I've been using works, but it works well. I was skeptical, but I tried it, and it works. You can check it out yourself, free demo, I think: https://www.soundradix.com/products/auto-align/ I also work a lot with mic placement, so the plugin rarely adjusts by more than a few samples, tho I can hear the difference. My last CD was recorded with 4 mics (actually 2 in spaced pairs + a stereo ribbon), without the benefit of anything to help automatically align them, and I was happy with the sound.

There are all kinds of examples where people do mix multiple mics, even if it's just bleed from a vocal/guitar or drum mics. it's always a bit of an issue, but it can also be used advantageously. Here's an example of someone else in a studio session with more mics than I can count (and this resulted in a very nice sounding CD). I'd like to know how all these mics were used, no idea if they were actually mixed, or just used to choose between:



And here's an example that's closer to the setup I've been using:



I'm not really advocating the OP's goal here, tho. I wouldn't use 2 mics and blend to mono. Stereo would just be a better use of the 2 mics, for the sound I want to hear. In both these examples, the end result was stereo.

[Wyllys]

Doug...

Thanks for the reply.

I mostly do live recording, live broadcast and audio feeds for video, so there's really no post-. I have to get it right from the git-go or it's no-go.

[Doug Young]

I'm sure that's a whole different ballgame. You just use one mic? Or one mic per source?

[Wyllys]

It varies with the logistics, budget, time available and such. On some broadcast things I'm given splits from all house mics, usually no more than 32 channels.
Sometimes I'll bring my own splitters and take what I need from the house and add a few of my own "broadcast only" mics.

Other jobs are a main M/S array with spot mics or something similar. Some of the cable access TV stuff requires me to do house sound and internet feed simultaneously with camera feeds. I'll run a couple of back-up safety recordings as well.

Fun...and pretty good business.

[rick-slo]

Originally Posted by Doug Young

Phase is totally frequency dependent, there's no free lunch. I don't know exactly how the plugin I've been using works, but it works well. I was skeptical, but I tried it, and it works. You can check it out yourself, free demo, I think: https://www.soundradix.com/products/auto-align/ I also work a

I often just use a time delay on the left or right channel (typically less than 0.4 milliseconds), watch a correlation meter and ultimately go by ear via headphones. Perfectly in phase on all frequencies is going to be mono, so you make you choices which few frequencies you want to more closely align.

I have played around with a demo combination delay and phase shifter from Voxengo. It works well. Interface photo below:

[Doug Young]

Yeah, my impression is that this plugin just automates that process, finding the best correlation. It takes about 30 seconds of sample to scan and compute.

[rick-slo]

Originally Posted by Doug Young

Yeah, my impression is that this plugin just automates that process, finding the best correlation. It takes about 30 seconds of sample to scan and compute.

I just go by ear and don't use the automatic calculation. The phase shift plus the track delay is an interesting combo of things to play with. However sometimes I do imagine things - more than once I have made plug-in changes thinking I hear some subtle change only to look and find the plug-in was in bypass mode the whole time.

[Doug Young]

Originally Posted by rick-slo

more than once I have made plug-in changes thinking I hear some subtle change only to look and find the plug-in was in bypass mode the whole time.

Glad I'm not the only one :-)

[Wyllys]

Derek, et al...

Perfect or absolute phase correlation at all frequencies is an impossibility. The only scenario where phase coherency can be absolutely perfect is a single sine wave. The best you can do is to time align the zero point of all tracks. All your mics will then be at a virtually equal distance from the source.

What it SOUNDS like after this is where we all make the judgement call, yea or nay. Any adjustments made need to be un-made and re-made for comparison. I assume this is what you're doing when you discover you've been working in bypass mode.

Phase difference is a natural part of all sound. It's one of the factors our brain uses to locate the location of a sound source. Any excessive phase differences show up when collapsing two signals to mono. In broadcast it is standard to aim for mono compatability, most commonly done by creating a stereo matrix using a mid/side configuration of elements. Some of the posted pics of studio miking show mid/side arrangements.

But you probably know all this. I post this to clarify for the casual reader that when you stated "perfectly in phase is going to be mono" it is not really correct. As I said, perfect phase alignment can occur only for a single sine wave. I can see how you'd call it "mono", but the "mono" is "mono" only in that it would be a single wave. Add a second frequency to the sound and all you can have is a consistent time relationship (phase difference) between the two waves, not a perfect correlation.

Here's a video that's not too bad:

https://youtu.be/Wv_RMhLu_So

The first minute is just extraneous chatter. The rest is good and the visuals quite helpful.

W

[ukejon]

Originally Posted by Bob Womack

Instead, I've moved to coincident pairs or to a rather interesting system first proposed by George Massenberg back around the time he engineered the Trio album for Parton, Ronstadt, and Harris in the '90s:

Bob, I have a pair of Shure KSM 137s that would work for this (I've done ORTF but didn't really like it for solo guitar....your modification makes much more sense). Could it work also with a KSM 137 paired with my Audix SCX25A, if I can work out the mounting issues? I like the idea of two different mics but what do you think about this setup?

Also, is the 110 degree reverse ORTF orientation still essential in this configuration? 90 degrees would be a sort of X-Y, right? And then less then that would be moving towards a closely spaced pair setup if I'm thinking about this correctly.

[Wyllys]

Originally Posted by rick-slo

I often just use a time delay on the left or right channel (typically less than 0.4 milliseconds), watch a correlation meter and ultimately go by ear via headphones. Perfectly in phase on all frequencies is going to be mono, so you make you choices which few frequencies you want to more closely align.

I have played around with a demo combination delay and phase shifter from Voxengo. It works well. Interface photo below:

I think these programs work by doing a series of time offsets between the two tracks/mics, flipping the polarity to register the amount of cancellation, registering that and moving on, repeating the analysis as needed per the program.

The idea being to identify the phase/time relationship with the least overall destructive interference equating the highest amount of cancellation from reversed polarity with the most desirable sonic result. I think this has been mentioned previously, being done manually to align the tracks. The plug-in just does it automatically, takes more samples for comparison and runs an analytic algorithm that doesn't suffer listening fatigue.

Seems a likely useful tool. I'd bet that the more expensive ones do a bit better job, but even a simple one should at least give some insight.

Of course, all this does is virtually re-position the mics. It is theoretically possible to have the mics in exactly the right place to produce tracks which need no further time-alignment. This is where the standard array techniques come into play. The plug-in will probably give relatively more improvement in tracks recorded with more randomly spaced mics because there's more "wrong" at the start.

In re the question about using different, unmatched mics for an array:

I'd say matched response will "line up" better. Any "sound" resulting from using different mics will be from the initial sound being "heard" differently. Such differences in frequency content due to mic polar sensitivity will be less amenable to post-production time alignment. Whether this is desirable or undesirable is a judgement call subject to what each individual prefers.

Ain't physics fun?

[Bob Womack]

Originally Posted by ukejon

Bob, I have a pair of Shure KSM 137s that would work for this (I've done ORTF but didn't really like it for solo guitar....your modification makes much more sense). Could it work also with a KSM 137 paired with my Audix SCX25A, if I can work out the mounting issues? I like the idea of two different mics but what do you think about this setup?

I'm sure the pair of Shures would do fine. I have no idea what the combination of KSM137 and SCX25A would sound like. Part of the idea of using two of the same mic model is to match the spectra for the two channels so you don't get "frequency steering" that pulls your attention towards one speaker. Using mismatched mics will obviously negate that benefit.

Also, is the 110 degree reverse ORTF orientation still essential in this configuration? 90 degrees would be a sort of X-Y, right? And then less then that would be moving towards a closely spaced pair setup if I'm thinking about this correctly.

The 110' 7" orientation was worked out by the French Office of Radio and Television Diffusion as a spacing that produced euphonic phase relationships between the two mics while both spreading the sound across the soundstage and providing a strong center image. The ORTF array is typically used at a pretty good distance on an ensemble. Unfortunately when used as a close-micing technique it tends to bring in too much room sound. If you reverse the capsule orientation (inwards rather than outwards) you get the same phase relationship but a different cancellation pattern that works very nicely at short distances without too much room tone. Frankly, I used the ORTF measurements as a shorthand and starting point to maintain phase relationships. I invested the huge $1.50 in a protractor and use it on sessions. That raises eyebrows! I have altered the orientation using my ears as a guide without bad effects. The point of the combination of 7" and 110' is to spread the sound across the speakers better than a co-incident 90' array.

Like Wyllys, I do a lot of broadcast mixing where a phase screw-up will bite you in the butt. These days I monitor my mixes on a DK jellyfish meter and I love it. Unfortunately they are hideously expensive.



But I got my start back nearly forty years ago when men were men and consoles weighed hundreds of pounds. Back then you got in the habit of punching the mono monitor button continuously to check for mono compatibility. We had a simple phase meter but the mono test was better. If it doesn't survive mono, you don't want it. Somewhere along the line it will give you fits. The easiest example of this is Paul Simon's "Mother and Child Reunion" where the guitar was split to two channels and one was reversed in phase to spread things. If you sum to mono the guitar disappears leaving nothing but its phase incoherent reverb in the mix.

Bob