Video Transcript
Priority Ducker
15m 21s
00:06
Okay, welcome back.
00:08
So, what we have left to do is integrate this
00:10
paging system into our design.
00:13
While you were away,
00:14
I built out a couple of generic signal paths
00:18
for different types of rooms.
00:20
If you look at my design here, I’ve got my lobby
00:22
and my conference room built out with a couple of things.
00:25
The lobby has a couple of different music sources
00:28
that can be chosen by this router,
00:31
which gets summed down to a mono channel.
00:33
It’s going through a continuous ambient compensator,
00:35
which will constantly keep the levels in this room high
00:38
or low based on how much crowd noise there is.
00:42
We’re finding that crowd noise via an ambient
00:45
microphone that’s in the ceiling,
00:46
which is also coming in through my COR’s
00:48
microphone channels and going out to a
00:50
loudspeaker in the room, probably a bunch of them.
00:54
In the conference room,
00:55
I’ve got four microphones coming in on Dante,
00:58
going through an AEC system.
00:59
Oh, I didn’t put in a reference yet;
01:01
I’ll fix that in just a second. In fact, it would come from right here.
01:04
This would be my AEC reference that would go
01:07
back to the reference sent to the far-end caller.
01:11
Meanwhile, the far-end caller’s voice is coming in.
01:13
I’ve also got a couple of HDMI sources in this room
01:16
that are going through an HDMI router.
01:19
I’m stripping the audio off that HDMI, sending it,
01:22
and mixing it in with my foreign caller’s voice
01:24
to go to the loudspeakers in the room.
01:28
Again, I’m referencing that out right there and also sending it to a display.
01:32
Not a lot going on here,
01:33
just talking you through some things that
01:34
might be in these various rooms.
01:36
We want to integrate these zones into these rooms. We’ll just do the lobby and the conference room.
01:43
We want to duck down the rest of the audio in
01:46
those rooms when this happens.
01:50
There are a lot of ways that you can do this,
01:52
and if the way that you do it successfully makes
01:54
the audio duck down, well, all right, you did it right.
01:58
You could do it in a really manual way.
02:01
You remember that in the PA router’s control panel,
02:03
we get this active LED that lights up when you’re
02:07
sending something to that zone.
02:08
You could use that LED to wire to the mute pins
02:14
of these audio sources to turn off things that are
02:16
in that room when the zone is active.
02:19
That’s just kind of forcing your way through it and punching a lot.
02:22
We want to do it a little bit more nuanced,
02:24
a little bit more subtly than that. There are two real different ways you could
02:31
think about doing this.
02:33
Option one is to take the outputs of the zones and
02:37
then integrate them into the signal paths of these various rooms.
02:41
Option two is to take these various rooms and
02:44
funnel them to the PA router and back again.
02:48
I’ll show you the difference between the two of them.
02:51
We’ll do option one first, where we take the
02:53
zones and send them to these rooms.
02:56
We’re going to do that using a component called the priority ducker.
03:02
It’s under your audio components.
03:03
I’ll grab this in here.
03:05
The priority ducker is not specifically a paging component.
03:09
You could use the priority ducker for all kinds
03:11
of different things and situations,
03:14
which means that in order to use it for our system here,
03:16
we are going to have to finesse it just a little bit.
03:19
The priority ducker by default comes in with eight channels.
03:22
I’m going to reduce that down to a single channel
03:24
so we can talk about what it does.
03:26
Even though I reduce it to a single channel,
03:28
it still has two input pins. What’s going on here?
03:31
Well, what the priority ducker does is it allows
03:34
whatever is wired into its input pins to flow
03:37
directly out to its output pins.
03:39
Nothing happens; it just passes through unless
03:44
something is detected on its final input pin called the priority pin.
03:50
Here’s how it would work.
03:51
I would insert this, going to break this path right here.
03:55
I’m going to insert it at the end of my signal path here in the lobby.
03:59
I’m going to wire my lobby’s program material
04:02
through the priority ducker, and right now,
04:06
the same thing happens as normal as was previously in that lobby.
04:08
The audio just flows through.
04:10
Then I’m going to grab the lobby PA material and
04:13
paste that onto the priority pin.
04:15
Now, whenever signal presence is detected in that priority pin,
04:21
it will duck down whatever is coming in on its other pins,
04:24
replace it with the priority pin,
04:27
and then when that priority pin goes quiet again,
04:29
it returns back to the original program material.
04:31
So, you don’t need to mix this audio in with a mixer later on.
04:35
The priority ducker does both.
04:37
It both ducks down and replaces,
04:40
then sends it through to the output,
04:43
and that’s what we want in this scenario.
04:46
If you had multiple channels, that happens to every single channel.
04:50
A priority ducker does not mix its inputs together.
04:55
These are all separate.
04:56
What goes in one goes out one;
04:59
what comes in two goes out two.
05:02
Those one and two are not mixing together.
05:04
So, you might want to use a priority ducker to
05:07
apply a particular channel to duck down
05:10
multiple different signal paths that have no
05:12
association with each other whatsoever.
05:14
You could do that in our case.I’m going to use independent, uh,
05:17
priority duckers for each of our rooms,
05:20
and let’s actually see if we can get this to work.
05:24
But before I get it to work, I want to warn you about something.
05:27
Whenever you start integrating paging into a
05:32
particular signal flow, you want to think very
05:34
carefully about where it goes in that signal path
05:38
and what it might affect in that signal path as well. Be careful.
05:43
I chose to put this at the end of the signal path.
05:47
Now, that might not have been the best choice because
05:50
this room has a continuous ambient compensator.
05:53
The point of the compensator is that it’s going to
05:56
rise or lower the gain of the things that you’re
06:00
hearing in the room if the crowd is loud.
06:03
If the crowd is loud, I probably want that same
06:06
thing to occur to my announcements so that
06:09
the announcements are also loud and louder than the crowd.
06:14
Right now, I’ve placed the priority ducker after
06:16
the continuous ambient compensator,
06:18
which would mean that my announcements are
06:20
played at a fixed level no matter what
06:22
—no matter how loud the crowd is,
06:24
no matter how quiet the crowd is.
06:25
That might not be the right choice for this particular room.
06:27
I might want to put that priority ducker earlier in the room,
06:31
before the compensator,
06:32
so that it’s part of the audio that is changed by the compensator.
06:36
That’s probably a good idea.
06:38
Another thing that may go wrong with the
06:40
way I’ve done it right here is that the compensator,
06:45
if you know what the compensator does,
06:47
requires a reference of what’s being sent out of the loudspeakers.
06:51
The compensator is listening to the room via a microphone,
06:54
and it also hears all the noise that you’re putting
06:56
into the room via that microphone,
06:57
and it has to ignore that noise so it can just hear the crowd itself.
07:01
Right now, I am not referencing out the announcements
07:05
that I’m playing into the room, and if I don’t do that,
07:08
then the compensator would think that these
07:10
announcements are the result of someone
07:12
in the crowd yelling really loudly.
07:15
Then the compensator is going to actually increase
07:18
the volume of the program material,
07:20
trying to compensate for the crowd it thinks has gotten louder,
07:24
fighting against the priority ducker,
07:26
which is trying to duck down the program material.
07:29
Lots of complications could arise.
07:31
So, if I really wanted my priority ducker after the compensator,
07:34
I would need to adjust my reference so that
07:36
I’m referencing the last thing that goes to my loudspeaker.
07:40
But more realistically, like I said,
07:41
I would probably actually put the priority ducker
07:45
before the compensator at all,
07:47
so that it is part of what is ducked or it’s part of
07:51
what is compensated to get it louder as the crowd gets louder.
07:55
However, that could also potentially be problematic
07:58
because what if you have something after the priority ducker
08:03
that’s got a mute button on it or a gain control on it?
08:09
What if, after this point,
08:11
I’ve given the user the ability to turn that volume down?
08:15
That’s now being applied to all of your announcements.
08:18
What if you’re sending an emergency announcement
08:20
to the priority ducker and the users have
08:23
muted that audio because they didn’t want to listen to it?
08:25
Now they’re muting your emergency announcement
08:28
and they’re not hearing the call to evacuate.
08:30
So, be careful about the way that you place this.
08:36
Generally speaking, I would recommend that the
08:38
priority ducker would almost always go at the very
08:41
end of your signal path because you want to make
08:44
sure that it is going to be audible regardless of what is
08:51
happening in the room.
08:52
In an instance like this, where I’ve got a continuous
08:54
ambient compensator that is also changing the volume,
08:57
what I’m going to do is I would take the applied
09:00
gain of this continuous ambient compensator,
09:02
which tells you how much it’s changing the gain,
09:04
and I would apply that to the gain that’s coming out of this room.
09:09
So, I can go to my lobby and I can say Zone One Gain.
09:15
I would apply the same gain of the CAC to that
09:19
zone so that no matter what,
09:21
my announcements will also be going up and
09:23
down at the same level as the continuous
09:24
ambient compensator while not being subject to
09:28
any muting or gain change that would otherwise
09:30
happen after the signal flow.
09:31
It gets tricky, right?
09:33
So many things to think about.
09:36
Just be careful where you put that placement
09:39
and what else in your flow might be affected by it.
09:44
Let’s make sure that it works.
09:45
I’m going to save the core and run.
09:48
All right, we’re connected.
09:49
I’m going to pin open once again the hover
09:51
monitor on my output here.
09:54
I’m going to play some music from my audio player.
09:58
I can see that’s coming through.
10:01
And now, I’ll play an announcement.
10:04
An announcement from Fake Hotel:
10:07
“Ladies and gentlemen, Fake Hotel would like to
10:09
remind you that our elevators are also fake.
10:12
Please stop pressing the elevator call button and
10:14
start climbing those stairs. Thank you.”
10:18
Sounded a little weird, right?
10:20
Because—let me close that—
10:22
we needed to adjust the priority ducker controls to
10:26
make sure that it’s actually going to react the way we
10:29
wanted to in an announcement.
10:31
Uh, there might be a little bit of quietness
10:34
as someone is talking.
10:35
Uh, as someone speaks, they might take a breath.
10:38
A priority ducker does not know that they’re just taking a breath.
10:42
They don’t know that the person isn’t going to continue talking.
10:45
The priority ducker just knows that there’s less audio here,
10:48
so it’s returning us to the music that was already playing.
10:51
So, that’s why it kind of sounded like you’re hearing the
10:53
announcements and you’re also hearing the music at the same time.
10:56
The ducker needs to be finessed.
10:58
So, let’s take a look at what this ducker’s got for its control panel.
11:02
Uh, you saw in here this response graph that was
11:04
bouncing around based on the audio as it was coming in.
11:07
It’s trying to figure out, “Hey, how loud is the thing on the priority pin,
11:12
and does that loudness constitute something that we should react to?”
11:16
That’s based on the threshold level.
11:18
If you lower your threshold significantly down to the very bottom,
11:21
then pretty much any noise that comes in on that
11:24
channel is going to be treated as something
11:27
that is worthy of ducking the other channels.
11:31
And then you choose how much you want to duck it down.
11:34
Uh, we’re ducking it down significantly.
11:36
You can see this red line at 100 dB.
11:39
Maybe you only want it to lower a little bit.
11:40
Maybe you want to hear the audio music just a
11:43
little bit in the background. That’s up to you.
11:46
Uh, you can also choose how much you are
11:48
changing the level of the priority channel.
11:52
And then some time constants down here.
11:54
The attack time is how quickly we duck down
11:57
that material when we get the presence of the announcement.
12:01
Ten milliseconds is pretty quick.
12:02
Maybe it’s too quick.
12:03
Maybe you want to give a slower fade on that so
12:06
that the audio or the music gradually drops
12:09
down while the preamble is playing.
12:12
Hold time and release time are very important.
12:14
I would increase your hold time to at least a second.
12:17
In general, that’s going to help compensate for
12:20
people who take pauses in their sentences.
12:25
You don’t want it to come back in while they’re paused.
12:28
A one-second hold time means that after the signal
12:31
presence has gone away, it will wait a full second
12:35
before it starts to release the program material
12:38
back to its normal level.
12:39
And if that signal presence reignites within that one second,
12:43
then it’s going to keep on going,
12:44
and then that hold time will start again the next time they stop talking.
12:47
The same thing might be true with a nice slow release time.
12:50
A couple of seconds of release time, three or four,
12:53
will allow the background music to slowly and
12:55
gently come back up to where it was.
12:58
Uh, and that way we’re not getting too much interruption.
13:02
So, let’s try that. I’m going to open up my
13:04
PA router as well and make sure we see everything coming through.
13:07
I will once again pin open my monitor.
13:13
Let’s play some music—the best music money can buy.
13:23
An announcement from Fake Hotel:
13:26
“Ladies and gentlemen,
13:26
Fake Hotel would like to remind you that our elevators are also fake.
13:30
Please stop pressing the elevator call button and
13:33
start climbing those stairs. Thank you.”
13:36
Now it’s done.
13:41
It’s returned back to its level.
13:42
Makes sense? You can adjust all of those at your discretion.
13:48
Make sure it makes sound,
13:49
make sure it makes the right kind of sound,
13:51
and reacts the right way that is appropriate for your venue.
13:53
All the controls are here. That’s what you want to play with.
13:56
So, that was all option one, the priority ducker.
14:00
That’s when we are taking each of these to the rooms.
14:03
If I wanted to do that for my conference room,
14:05
same thing. I would take the conference zone outputs.
14:09
We would add another priority ducker down here
14:12
before the output of my conference room,
14:15
and this time I would use the conference channel and send it here.
14:24
I would once again need to update other things in the room.
14:28
In this case, I have an AEC that is referencing all of the audio in the room.
14:32
I’m going to readjust my reference so that it is
14:35
including the announcements in the room
14:37
as the things that it wants to ignore from your microphone
14:40
channels as part of your teleconferencing system.
14:43
Uh, and that way, that is included,
14:46
and we’re going to remove that audio from the
14:48
audio that is sent to the far-end caller.
14:50
If you want the far-end caller in this case to
14:52
hear your announcements,
14:54
then you would just send those announcements
14:56
to the far-end caller as well. In this case,
14:57
I’m going to remove them from the room.
14:59
So, that’s our priority ducker sending this out to
15:03
each of the signal paths, and that might be the right way for you.
15:06
What I like about the priority ducker is that my
15:08
signal flows all stay linear. I can see them on one path.
15:12
They stay straight. I’m not breaking them and sending
15:14
them somewhere else and coming back again.
15:16
That’s what we’re going to do with our second option.