Automatic Mic Mixer Settings

[Malabaristo]

I have a high school theatre system with a Biamp Tesira DSP that uses a few inputs set up with AGC to act as an automatic mic mixer. The idea is that a non-technical person can use 1-2 handheld wireless mics and/or a couple wired mics without much training. That sounds reasonable in theory, but the A/V contractor did a terrible job on almost everything, so I've been slowly going through the whole system to make it actually work as intended.

I found a combination of settings on the AGC that behaves nicely across a wide range of people speaking, and in general it works better than just setting a static level and hoping people know how to talk into a microphone (spoiler: they don't). However, when there are a bunch of kids making noise and someone tries to talk loudly into a mic to get their attention, the AGC just brings it back down to a good speaking level. This obviously frustrates the user, and then they complain to me about it not being loud enough.

So, I'm struggling with finding the right balance. If I set the AGC target high enough to be heard over a crowd, then it's way too loud for normal use. If I disable the AGC and set a static level, then either it's nearly always wrong, or an operator is required to keep it in a good range as the mic gets passed around--which defeats the goal of simplified automation.

If this were only a lectern mic, I'd be tempted to stick a momentary, "Make me really loud" button on it, but since the person may be wandering around with the handheld wireless, there isn't a good spot to put that button. Also, today I had a situation where the crowd just never quieted down through the whole 45m presentation, so a momentary override would not have solved the problem.

 

[Jay Ashworth]

The general solution to this, in lieu of Dugan automixing, is to put gates on each mic, so it generates no signal to the stereo main until the volume on it breaks a certain level; a slightly different affair than AGC.

Do you have gates available?

 

[MNicolai]

I don't play around much with Biamp these days but for set-it-and-forget-it wireless you may be better off applying some moderate compression on those channels, gates, and bringing the gain up a little bit. Still preserves the general dynamics of the presenter but softens people who talk loudly/directly into the mic.

Think there's also a gating automixer Biamp has programmed in that you can use for similar effect as well.

AGC is actually better suited for audio conferencing. The purpose isn't really to preserve dynamics but to expand quiet sounds, soften loud sounds, and make any/all signal approximately the same intensity. This is so in a conference room application the person on the far end of the call can't tell which voice is the person seated next to the mic versus the person talking from the back corner of the room.

Related: I do a lot of video recordings in the field of mechanical and generator noise. I try to never use my iPhone for this because the AGC makes everything sound the same whether I'm 100' from a loud piece of equipment or 10' away. The mic on my phone is doing exactly what it's supposed to, but that makes it near impossible to get a sense of how loud something actually is based on a phone recording.

At the end of the day, there's only so much you can do for people who do not know how to talk into a microphone. Garbage In = Garbage Out. They'll learn quickly enough so long as the sound system is doing things in a logical manner. If the compression/AGC is really aggressive, then they'll just throw their arms up and say "Whatever I do, it always sounds terrible" and give up. Try not to over-think your signal flow.

 

[Jay Ashworth]

FWIW, current Behringer x-series boards (some of which are very small, and rackable) have actual Dugan automixing in them; not sure if that's pertinent.

 

[RonHebbard]

FWIW, current Behringer x-series boards (some of which are very small, and rackable) have actual Dugan automixing in them; not sure if that's pertinent.

@Malabaristo @MNicolai @Jay Ashworth Is a "ducker" with a separate mic sensing the ambient level a possibility and of any value in your situation?
Toodleoo!
Ron Hebbard

 

[Malabaristo]

Try not to over-think your signal flow.

But I want a magic solution that replaces an operator!

Oh, right. Reality...

The problem I was trying to solve by leaving the AGCs in place was balancing the people who hold hand-held mics down by their navel or stand 5' back from the lectern with those that try to eat either style of mic. The traditional role of a mic mixer in handling multiple inputs has actually been a pretty rare use case so far: it's usually only one input being used at a time, but sometimes two. The DSPs do have that feature, but I hadn't settled on going down that path yet. After a fair amount of experimenting, I found the right AGC settings to achieve the goal of adjusting all of the quiet talkers and loud talkers to a consistent level without sounding bad. It's just that the consistent level is not necessarily the right one depending on what's going on in the house.

For the sake of being a contrarian, I'm tempted to try adding a layer of complexity: I have a room mic used to feed distributed monitors... might be interesting to try to adjust the output level of the "simple" inputs based on that mic level...

Edit, just saw Ron's reply as I was posting: Yes, like a ducker, but inverted... A stander? A raiser?

 

[RonHebbard]

But I want a magic solution that replaces an operator!

Oh, right. Reality...

The problem I was trying to solve by leaving the AGCs in place was balancing the people who hold hand-held mics down by their navel or stand 5' back from the lectern with those that try to eat either style of mic. The traditional role of a mic mixer in handling multiple inputs has actually been a pretty rare use case so far: it's usually only one input being used at a time, but sometimes two. The DSPs do have that feature, but I hadn't settled on going down that path yet. After a fair amount of experimenting, I found the right AGC settings to achieve the goal of adjusting all of the quiet talkers and loud talkers to a consistent level without sounding bad. It's just that the consistent level is not necessarily the right one depending on what's going on in the house.

For the sake of being a contrarian, I'm tempted to try adding a layer of complexity: I have a room mic used to feed distributed monitors... might be interesting to try to adjust the output level of the "simple" inputs based on that mic level...

Edit, just saw Ron's reply as I was posting: Yes, like a ducker, but inverted... A stander? A raiser?

@Malabaristo In 1999, the company I was with installed an Alcorn McBride automated AV and system-wide show control installation in a two story museum for Tussaud's on the upper two floors of a three story structure in front of the casino with the gondola's and moat / canal on the strip in Las Vegas. We were able to employ BSS's ducker to momentarily lower the level of the programmed background tracks in one of their display rooms to enable a live staff member to speak over the tracks with a hand-held Shure wireless if / when required.
Toodleoo!
Ron Hebbard. (Yes, they even allow Canadians to install systems in Las Vegas)

 

[Malabaristo]

@Malabaristo In 1999, the company I was with installed an Alcorn McBride automated AV and system-wide show control installation in a two story museum for Tussaud's on the upper two floors of a three story structure in front of the casino with the gondola's and moat on the strip in Las Vegas. We were able to employ BSS's ducker to momentarily lower the level of the programmed background tracks in one of their display rooms to enable a live staff member to speak over the tracks with a hand-held Shure wireless if / when required.
Toodleoo!
Ron Hebbard. (Yes, they even allow Canadians to install systems in Las Vegas)

Any recommendations on where I can find a ducker that would effectively lower the level of a room full of rowdy high school students when someone they don't want to listen to is talking?

 

[MNicolai]

Any recommendations on where I can find a ducker that would effectively lower the level of a room full of rowdy high school students when someone they don't want to listen to is talking?

Swing a sock full of nickels around on a rope and people will duck and listen to you telling them to be quiet real quick like.

 

[RonHebbard]

Any recommendations on where I can find a ducker that would effectively lower the level of a room full of rowdy high school students when someone they don't want to listen to is talking?

@Malabaristo Do you have any friends, female or male, who are REALLY good with bull whips??
Toodleoo!
Ron Hebbard

 

[FMEng]

If the Tesira has a gate with a sidechain input, you could put it on the output of the AGC. Drive the side-chain of the gate with a room mic. Set the gate to attenuate 10 dB most of the time, with 10 dB of make up gain following it. Adjust the threshold of the gate to open only when the signal from the room mic is high. Kids yell, and presto, 10 dB more level. Kids get quiet, and it goes back to normal level. Make sure you don't get feedback when the gate is open and the AGC is at max boost (or minimum compression).

 

[Malabaristo]

If the Tesira has a gate with a sidechain input, you could put it on the output of the AGC. Drive the side-chain of the gate with a room mic. Set the gate to attenuate 10 dB most of the time, with 10 dB of make up gain following it. Adjust the threshold of the gate to open only when the signal from the room mic is high. Kids yell, and presto, 10 dB more level. Kids get quiet, and it goes back to normal level. Make sure you don't get feedback when the gate is open and the AGC is at max boost (or minimum compression).

The gate block doesn't, but the ducker does... which is fine in this case: I just need something that gives a logic "true" when a threshold is exceeded. There's also a source selector block that takes logic inputs and lets me specify different levels for each input. One issue I'm seeing is that there's no hysteresis on the logic output of the ducker; the attack and release times only affect the audio passing through it. Adding a logic delay of 0.5s seems to help keep it from bouncing back and forth too much, but 10dB is still pretty big for an instant jump in either direction, so....

More complexity!

As an experiment I used three ducker blocks with different thresholds, plus a combination of AND & NOT logic gates to control a 4-way source selector (0db, +3dB, +6dB, +9dB). By playing crowd noise through some effects speakers, I could verify that it switches up and down as the thresholds are passed, so it's all in there correctly. The delays aren't quite as good as actual hysteresis, but they definitely help. Now I just need an event with a crowd I can experiment on in order to figure out what actual values to use for the thresholds.... and whether this all makes sense, or whether it's just some overcomplicated silliness. If nothing else, I'm learning more about how the software works!

 

[FMEng]

That's a very clever approach. It isn't ideal, but in the circumstances where you need it, it might be fine.

 

[Malabaristo]

So... it doesn't quite work as intended. All of the logic is correct, the premise is fine, and it basically does what I told it to do except for one thing: self-pollution. The room mic picks up too much of the output from the PA system, and it can't distinguish between crowd noise and the amplified sound of the presenter. If I had a way to filter that out, then it would probably work better, but I think this has been enough of an experiment to give up on the idea and go back to a more manual approach.

Also, I had a person today who seemed to make it his goal to prove that an automated system can never handle every scenario. He was given the option of either a hand-held wireless mic, or a lectern-mounted mic. ...and he decided to use both. The wireless allowed him to wander over and point at things in his Powerpoint, but when he happened to be near the lectern, he would sometimes set the handheld down and switch to that mic mid-sentence. Not consistently, of course: sometimes he would lean on the lectern and keep talking into the handheld while well within the pickup range of the other mic. There's just no way to handle that sort of thing without having a human operator watching and adjusting to match the whims of the presenter.

 

[rwhealey]

FWIW, current Behringer x-series boards (some of which are very small, and rackable) have actual Dugan automixing in them; not sure if that's pertinent.

Tesira has a gain sharing automixer but doesn't license the Dugan name like Yamaha does or say "Dugan-style" without permission like Behringer (https://www.audiomediainternational...-clarifies-use-of-name-by-other-manufacturers). The original Dugan patent for a gain-sharing automixer has long expired. Tesira is an extremely powerful audio DSP - I would expect it to far exceed the processing capabilities of an X-series product.

A gain sharing automixer is useful when you want to avoid feedback with a lot of open mics, like a panel discussion. It's also useful in musical theater. For an auditorium presentation system, I find that the best approach is to not give the user the option to have that many open mics. If they need to do a panel discussion, an operator can use an automixer built into the sound board. This is obviously user-dependent - for instance, if you are talking about a higher ed classroom you may not be able to use this approach but with a staffed auditorium at some point you need someone to run the sound board, even if they are just babying a panel discussion.

So... it doesn't quite work as intended. All of the logic is correct, the premise is fine, and it basically does what I told it to do except for one thing: self-pollution. The room mic picks up too much of the output from the PA system, and it can't distinguish between crowd noise and the amplified sound of the presenter. If I had a way to filter that out, then it would probably work better, but I think this has been enough of an experiment to give up on the idea and go back to a more manual approach.

Biamp has an ambient noise compensating feature in Vocia, their paging system. I believe this functions like Atlas/IED's solution (ariport paging) where the program signal is automatically filtered out of the ambient mics before being used to raise/lower the level of the page. I am not finding that this is built directly into Tesira, and I'm not sure how well it would work with in-room program microphones. I expect poorly. Biamp has excellent support in my experience - you could always ask!

At the end of the day, there's only so much you can do for people who do not know how to talk into a microphone. Garbage In = Garbage Out. They'll learn quickly enough so long as the sound system is doing things in a logical manner. If the compression/AGC is really aggressive, then they'll just throw their arms up and say "Whatever I do, it always sounds terrible" and give up. Try not to over-think your signal flow.

+1

The feedback reducer in Tesira is quite good, even when it is set aggressively. Much better than the older stuff that is universally panned. When I do "simple systems" in auditoriums, I typically only give the option for the user to have 2-3 open mics. I will give each mic a fader on a touchscreen at the lectern position with the maximum level of the fader set to be around the point of feedback with a normally-held microphone. If the user wants to be loud and pushes the volume to the top or eats the mic, they will hear they are on the threshold of feedback but will get rescued by the feedback reducer. The touchscreen control lets the user adjust for loud/soft talkers, or at least lets the auditorium manager show the complaining user that they have a volume control. There just isn't much that can be done about someone holding the mic at waist-level.

 

[Jay Ashworth]

Maybe I missed it, but I thought the x/Air boards actually licensed it, no?

 

[FMEng]

Also, I had a person today who seemed to make it his goal to prove that an automated system can never handle every scenario. He was given the option of either a hand-held wireless mic, or a lectern-mounted mic. ...and he decided to use both. The wireless allowed him to wander over and point at things in his Powerpoint, but when he happened to be near the lectern, he would sometimes set the handheld down and switch to that mic mid-sentence. Not consistently, of course: sometimes he would lean on the lectern and keep talking into the handheld while well within the pickup range of the other mic. There's just no way to handle that sort of thing without having a human operator watching and adjusting to match the whims of the presenter.

The presenter needed to be given a choice of either podium or hand-held, instead he was allowed the choice of both. If he chose to use the hand-held, the podium mic should have been removed. The best of both choices might be to provide a stand on the podium to hold the wireless when he wanted to free his hands.

 

[FMEng]

Maybe I missed it, but I thought the x/Air boards actually licensed it, no?

Behringer is quite upfront that they don't license Dugan and that the two methods work differently. They say the result is "Dugan-like." Besides, why would they spend money to license an expired patent.

 

[Jay Ashworth]

They might not. But they might license the trademark, which is a separate item.

Hmmm.

Well, I'm not the only person who missed it:

https://www.bhphotovideo.com/c/product/1095112-REG/behringer_x_air_xr18_18_channel.html