The Reverberant Field

Graph the strength of your singers versus the volume in the orchestra pit. For any shared acoustic space, there is a point where your show becomes unmixable. 

Which is to say, nothing you do with microphones and speakers will restore the proper balance. And it's worse than that. Sound reinforcement for music -- much less a musical -- is not about winning the volume wars, it is about achieving intelligibility. To follow the story, the audience must be able to understand the words. And that's much tougher to achieve.




A large part of the reason is the reverberant field. When sound enters a space, it reflects. When it has reflected several times, over a period of up to several seconds, it becomes what we call the reverb tail. But there's several very interesting things about this tail. Since it has bounced back and forth from all available surfaces, it is essentially anisotropic and omnidirectional. Since it is composed of reflections from the last second or several of material, it contains no identifiable transients. And since high frequencies are more easily absorbed, it is weighted towards the low end.

What all this means is the acoustic remnant left after the vocal energy of singers, musical instruments, and sound reinforcement speakers has entered the space is a muddy, boomy, mush. Referencing only the reverberant field, a string bass is just "Ooommooommooom"; no pluck, no notes, no rhythm. For vocals, you are lucky to make out the basic vowels; consonants are lost.

And it gets worse.

As you add volume, whether directly from louder singers or instruments or by turning up microphones, you "pump" the space. The reverberant field becomes stronger in a non-linear way, becoming closer to a continuous wall of white noise. And certain frequencies are pumped as if in a laser cavity; room nodes emerge, standing waves, tones that are captured and amplified and sustained.

(There's a few other side effects; fixtures and architectural elements rattle, adding more noise, and the audience becomes noisier as well.)

Since the reverberant field is practically by definition at equal volume throughout the acoustic space, the direct sound forms a ratio with it. Put a singer at one end of the room. Close by the singer, their voice is louder. By the time you get to the back of the room, the reverberant sound dominates.

And this has huge implications for sound reinforcement.




FOH mixing in any space smaller than a stadium is all about selective reinforcement. Take a small-scale example; a recorder and trumpet duet. In a small space, you might put a mic on the recorder and feed it into the house system to help it match the trumpet's volume.

Problem is, your house speakers aren't on stage. So if you stand in front of a speaker, you get just recorder. Stand in front of the stage, you get just trumpet. Only a little ways back in the hall do you get equal amounts of trumpet and recorder.

Take another typical example; bass player with their own amp. From the audience, all they hear is "oomoomoom." So you take a DI off the bass and emphasize the attack a little. This means two streams have entered the acoustic space:

"Ooomooomooom"
"Tac! Dac! Dac!"

And at the proper distance, they blend into the proper "Toom doom da doom doom da doom." But anywhere other than down the alley where direct projection from stage is roughly matched to direct tweeter radiation from speakers, you get something other than a good bass sound.

Go back to the first. One alternative is to mic both instruments. So from anywhere within range of the speakers, you are hearing both amplified recorder and amplified trumpet. Plus the reverberant field. And therein lies the catch. If the trumpet player, freed from the responsibility of blending with the recorder, increases in volume, the reverberant field increases non-linearly. And as you increase the speaker volume to compensate, the reinforcement is also fighting its own reverberant field.

Why is it non-linear? Well, air has mass. It takes a certain amount of energy to overcome the inertia of the air and set up standing waves. On the listener end, perception starts being non-linear; it takes ten times the power to be perceived as twice the volume, but this is also frequency dependent and the frequency response changes at different volumes. At higher volumes, masking effects in the cochlea become more dominant, up through the onset of hearing fatigue; short-term hearing loss.

Which all work out to progressive loss of perception of the high frequency content. As you crank higher and higher, the desired sound; the harmonics that define the character and timbre, the higher-frequency sounds that define the attack transients, and of course the very transients themselves get lost in the smear of reflected, bottom-heavy, indistinct noise that makes up your reverberant field.

But good luck getting it fixed. If you ask for the trumpet to perhaps play a little softer, all you will get is well-meaning suggestions to turn his mic down. Which, of course, devolves us back to the first set-up; where the people directly in front of the speaker hear the recorder and only a couple of seats get a decent balance.

(Only you've lost that, too, since the reinforced recorder is now masking itself; the transients are lost in the echoes of the reverberant field, and the characteristic timbre is masked in the thrumming low end you've filled the hall with. And of course the trumpet is also lost in the reverberant field now, meaning you might as well have two fretless basses up there for all the character and melodic content that comes through).




Consider what selective reinforcement means for design of a musical. From the back of the house, the reverberant field overwhelms anything coming directly from the stage. From the back of the house, then, it becomes clear you need to roll off the low end and push the presence in the vocal microphones.

Which if done right can provide a pleasing voice in which it is generally possible to make out the words.

But move closer to the main speakers. If you stare down the throat of the speakers, you are getting just the reinforcement; an overly bright, tinny, compressed voice with lots of artifacts. Not pleasant to listen to!

In many seats towards the front of the listening space, some direct bleed comes off the stage and mixes with what's coming off the speakers, plus of course you have the reverberant field. So low end is filled in by the field, mid-range and localization cues come directly from the singer, and presence and definition and consonants are coming off the speakers. And if the speakers are time-aligned properly and not too far away in the horizontal plane (the human ear gives more of a pass to misalignment in the vertical plane) you get a pleasing voice.

Of course, the reverberant field is delayed. That's what it is; reflections from the distant walls. So that smears the lower end. And, yes, it is impossible to align your mains perfectly for all seats; a little geometry will demonstrate that! But here it gets even trickier; the human perceptual apparatus will bring in all the frequency data more-or-less without question, but it will provide perceptual focus based on location, time-of-arrival, and frequency content.

Which boils down to -- the total mix may or may not be pleasing, but it is nearly impossible for untrained ears to sort out which elements are contributing what to the total mix. The reverberant field, particularly, is generally unperceived. The influence that low end leakage has on the volume sensitivity and resulting frequency sensitivity curves in the listening ears will always be underestimated and overlooked.

As that reverberant field rises in ratio (whether due to sheer level, or to increasing distance of the listener from stage and/or speakers) the ability to sort out the desired musical information becomes less. It is like the dark matter of sound at that point; making voices sound muffled and inspiring lots of orders to "turn it up, turn it up!" when the problem is the unseen drag of low-frequency white noise.

And this means the correct emphasis to blend with the acoustic material in the space is different for each seat. And different for each voice. And changes there, too; the singer who needs a lot of help will be emphasized more in speakers, the singer with the strong voice will be almost non-existent in the speakers (depending on where you mix to, of course; whether you mix for the person standing in front of the speaker, the person in the back of the room, or try -- as most of us do -- to achieve a compromise that will work for as much of the audience as possible).

This means if Ethel Merman singes a duet with Little Suzy, the people near the speakers will complain all they hear is a child's screechy voice being amplified way too much, and the people everywhere else will hear nothing but Ethel Merman. This of course evens out the flatter your speaker coverage is!

And it gets subtler than this. Take two singers of similar vocal timbre but different strengths of production. As the reinforcement is balanced for tonal deficiency, they will sound the same in the sweet spot, but one will have an odd, artificial sounding microphone voice right next to the speakers. And that same one will be dull and muffled compared to his partner when listening from distant in the hall.

Or take either of those singers, and change the total volume of the song. At low volumes of reinforcement, the natural voice dominates in the front of the hall and the reverberant field dominates in the rear. At higher levels of reinforcement, the microphone sound dominates over greater and greater parts of the hall. And if the microphone sound is tailored to emphasize needed frequencies and otherwise selectively reinforce the voice, the amplified sound will pass through insufficient, to nicely blended, to artificial as you increase volume.



This is why reinforcement -- working around and supporting the natural acoustic sound from the stage and the orchestra -- is the most difficult kind of sound. 

It is also why amplification -- powering over the direct sound with a fuller-range, more accurate picture -- only works within certain boundaries.

And in both cases, you remain at the mercy of backline leakage. No matter what the strategy you pick, if a band continues to turn up their instruments, there will come a point where there is no alternative left for mixing. The show is just going to have to suck.

In other news, I just finished the third weekend of mixing "Poppins."