AEC & Q-SYS Conferencing System (Part 2)

Video Transcript

00:07
Welcome back. Now that we’ve looked at the process in a little detail,
00:11
let’s look at some standard conferencing system terminology:
00:15
Let’s say that I’m in a room using a Q-SYS system for conferencing.
00:19
I am what’s called the ‘near end’.
00:21
This means my AEC algorithm will cancel echo for the callers into my room.
00:26
Note that the AEC used in my system doesn’t help me, it’s there only for the remote participants.
00:33
The ‘far end’ is the remote party in my conference.
00:36
Again, the AEC in my system as the ‘near end’ is to benefit them.
00:41
As noted before, the AEC algorithm doesn’t know what signal it should be looking for and attempting to cancel.
00:48
We do this by connecting that signal to the AEC reference pin of the processing block.
00:53
We’ll discuss best practice for this later in the training module.
00:57
The process of ‘convergence’ is what the adaptive filter is doing
01:01
to achieve the best echo cancellation possible given the circumstances.
01:06
When the far end is hearing no echo, we say the AEC algorithm is converged.
01:12
As we’ve discussed, proper AEC processing is an intensive operation.
01:17
In general as DSP processes get more intensive the audio latency,
01:22
or the time it takes a signal to travel through a signal path increases.
01:26
Some of this latency is purposeful in order to give the algorithm more time between the arrival
01:33
of the reference signal and the microphone signal that needs to be processed.
01:36
Note that the input to ouptut audio latency through Q-SYS is 3.17ms without AEC in the signal path.
01:45
The AEC algorithm introduces 21.4ms into the
01:50
signal path of the 110f and 10.7ms for all other Q-SYS cores.
01:55
That’s a total latency of 24.57ms in the core 110f and 13.87ms in all other cores.
02:03
That may seem like a long time,
02:06
but remember conferencing systems aren’t typically latency sensitive applications.
02:11
It usually takes far longer for the signal to travel to the far end,
02:15
so this is just one component of the overall conferencing latency.
02:20
A larger room will usually have reflections that return to the microphone later than a small one.
02:26
The tail length of the AEC algorithm determines how long it ‘looks’ for reflections to be cancelled.
02:32
The Q-SYS AEC algorithm has adjustable tail lengths of 100, 200, 300 and 400ms.
02:41
Remember the number of AEC algorithms a given core will support is specified for the 200ms tail length.
02:48
300 and 400ms algorithms require more processing resources, so fewer microphones can be processed.
02:56
For example, the core 110f can support 16 200ms algorithms
03:01
while the number decreases to 8 for 400ms tail lengths.
03:06
The Q-SYS AEC algorithm employs a concept known as ‘Reference to Microphone Level Ratio’,
03:13
or RMLR to help us properly calibrate the system.
03:17
This is analogous to another term common to AEC systems – ‘Echo Return Loss’ or ERL.
03:23
It represents the natural loss through the output gain structure, through the room and back to the microphone.
03:30
The RMLR represents the difference between the level at the reference
03:34
and the resulting signal heard at the microphone.
03:37
In this example, the level of the reference is -20db while the resultant level at the microphone is -30db.
03:45
The difference, or RMLR in this case is 10dB.
03:50
To get the best performance from the algorithm, our goal is to calibrate the block to achieve an RMLR of 0.
03:58
Later we’ll see how to make this adjustment.
04:00
The Echo Return Loss Enhancement, or ERLE, represents
04:05
the relative success of the algorithm in cancelling the echo.
04:08
Mathematically, it is the difference between the level of the signal at the input of the block
04:13
and the resultant level.
04:14
In this case, we see a level of -30dB into the algorithm with a level of -50dB post processing.
04:22
The computed ERLE would then be 20dB.
04:26
We can think of is this way: The RMLR represents
04:30
the natural relationship between the level in the room vs. the level of the echo.
04:34
The ERLE represents the work done by the algorithm itself.
04:38
We saw before that the AEC algorithm isn’t always completely successful. ‘Residual Echo Suppression’,
04:45
or RES, is the non linear processing that removes that left over signal.
04:51
It’s called ‘non-linear’ because this part of the algorithm breaks the residual signal
04:55
into many small frequency bands and adjusts the level of each independently.
05:01
For example, there may be -40dB of signal at 500Hz while there’s -50dB of signal at 2kHz.
05:08
RES will attenuate the 500Hz band more than 2k.
05:14
As we continue to explore AEC applications in this module,
05:17
we are considering rooms small enough that participants
05:20
can hear presenters without the need for local reinforcement.
05:24
There are generally two conferencing system applications that require in-room reinforcement:
05:29
Voice lift is the term often used when we’re referring to
05:33
local sound reinforcement of a presenter in a given room.
05:36
This is not exclusive to conferencing systems,
05:40
but is still quite common in rooms with a large presenter’s area
05:43
that’s far from the conferencing participants.
05:46
In very large conference rooms or other applications such as legislature spaces,
05:51
there are enough participants that not all of them can hear each other without some reinforcement.
05:56
In most cases a scheme known as ‘mix minus’ is employed so that a given talker
06:02
is only reinforced in parts of the room where they aren’t heard directly.
06:06
This requires a sophisticated mix and many discrete speaker circuits to get the best results.
06:12
Best practice and troubleshooting for these two scenarios are covered in the Quantum level 2 curriculum.
06:17
So, stay tuned for news on that.
06:19
For now let's just take a quick break, and come back whenever you're ready.