Professional Documents
Culture Documents
The Complete Guide To Measurement Micropho - Nathan Lively
The Complete Guide To Measurement Micropho - Nathan Lively
Introduction
Terms
Height
Head-height or ground plane?
Multi-tiered seating
Focus on the Problem
System 1: Coupled
System 2: Uncoupled—Vertical Displacement
System 3: Uncoupled—Horizontal Displacement
System 4: Uncoupled—V+H Displacement
What about the right side?
Find a Solution
System 2: Subs on the ground—Vertical displacement
Petal Aim
Max offset
System 3: Center Flown Sub—Horizontal displacement
System 4: Center Ground Sub—V+H displacement
System 5: Center Ground Sub Arc —😬 displacement
FOH
Average Phase
Further Questions
If it can't be aligned everywhere, why align anywhere?
Why did you use 10dB for the crossover region and not 12dB or
20dB?
Florian Spille: Where do I have the best chance to make a good
decision for most of the audience vs. where do I have a big
chance to make a real bad decision?
Marc Zeebregts: If I have a subwoofer array, let's say it's an array
of 8 subs evenly spread between L&R mains:
Where do I measure?
Do I measure with all subs on?
Is there a difference in approach when I arc these subs
(either physically or with delay)?
Sajag Gupta: How do I align flown subs in an LCR system?
Liam Devlin: How is the process affected by multiple sub
positions? For example, if you had subs flown behind mains and
you also had ground stacked subs that were further from your
measurement position?
Τάκης Λόντρας: Mic placement in closed-space venues and small
venues.
What do manufacturers recommend?
Final thoughts
Acknowledgments
Bibliography
T C G
M M
P
S A
2nd Edition
by Nathan Lively
Everything is chaos.
Learn how to live with the chaos.
-Mauricio Ramirez
I
Where is the ideal microphone placement for subwoofer alignment?
If we can only have one alignment point, but there are 800 seats in the
theater, does it even matter?
It does, because the differences over location aren’t random. They follow
patterns. Mic positions sort out the patterns and connect them together.
Each mic position has specific roles to play with purposeful (not random)
placement. (McCarthy 446)
When it comes to subwoofer crossover alignment, we can define mic
placement with some precision. The ideal placement will result in the least
error across the audience. It will put the maximum number of audience
members in the coupling zone, where there is positive summation; and the
minimum in the cancellation zone, where there is negative summation.
While it is possible to calculate the magnitude and phase relationship of
direct sound at any specific point, all we really need to know is how much
those relationships have changed across the audience and whether or not the
original alignment position has given us the best shot at minimum variance.
Towards this goal, we will consider the ramifications of source
displacement with a detour through measurement height.
Terms
Some dry information that you'll probably skip over and then come back
to later.
1. ONAX : On-axis
2. VBOT : Vertical bottom
3. VTOP : Vertical top
9. Phase : The relationship in time between cycles of a wave. The radial
component of the audio waveform expressed in degrees.
10. Summation zones (McCarthy 135)
1. Level-based
2. Phase-based
H
Head-Height Or Ground Plane?
Head-height is preferred since it most closely matches the listener's
experience, but should be abandoned for ground plane if actionable data is
heavily compromised by floor reflections. The lower the microphone
position, the shorter the path length difference due to floor bounce, the
higher the comb filter frequency. Keep in mind that changing height will
also require a change in depth.
In the example below, if you decide that you want to measure at ¾ depth,
consider a line from your main that intersects with your mic and then the
floor and move back to this position. Mauricio Ramirez says, "If this is your
first time ever, do it this way."
If you are measuring at ground plane and getting low quality high-
frequency (HF) data because the floor is covered with carpet or full of
chairs, consider using a pane of glass, glazed ceramic tile, or rigid plastic,
with an area of 30-60cm 2 (1-2ft 2 ) and 2-5mm (0.08-0.2in) thick. Make
sure this object does not resonate or introduce (self)noise created by
mechanical vibration or rattling. Or, simply ignore the HF response during
your crossover alignment, then move back up to head height for other
calibration processes that focus on the HF.
If you are using multiple microphones (see Average Phase below),
consider lowering the height over depth in order to further randomize the
reflections and improve the average. For example, a 3-mic average might
use the following heights: VBOT 1.75m (5.7ft), ONAX 1.6m (5.3ft), VTOP
1.45m (4.8ft).
Here you can see the cancellation pattern created by the floor bounce at
143Hz . I have moved all of the measurement mics into the null to create a
clear demonstration of what they might look like if they ended up there by
accident.
If the microphones all end up in the null, your average will have a null.
I'm showing you the magnitude response here instead of phase because it's
easier to see the results on top of each other.
You can counteract this behavior by making a small adjustment in the
microphone height against the cancellation behavior.
Then the average between the microphones will not favor the null as
much.
Multi-Tiered Seating
If you have more than one audience plane, it can help to create a new
theoretical audience plane that represents the average connecting VBOT of
the ground floor with VTOP of the balcony. More on how to use this in
your calculations below .
F P
In the beginning, there was a single source and the innocent audience
knew not of relative phase misalignment. Then, in 1964, the first subwoofer
was patented, the operating range was extended, and complexity grew. The
end user became responsible for a successful marriage between sources.
With great power comes great something-something.
Coupling is a magic trick. We take two independent sources with
different locations and response characteristics and make them sound like a
single source to the listener. When there is minimal change between
distance offsets, phase relationships are consistent and microphone position
is less critical (and you probably don't need this guide).
If you walk around the audience from left to right and front to back and
the difference in distance from you to the main and sub (Δdist) stays the
same, so will the phase alignment (or misalignment). Meanwhile, if there
are large changes in Δdist then there will be a large phase span (aka phase
problem). Identifying these areas of significant change will help us focus
our efforts.
System 1: Coupled
Are these subwoofers coupled or uncoupled?
Observing the phase graph for measurements of each sub where Δdist is
greatest (side location), we can see that they are easily within 120º and
therefore we can expect them to be coupled at any other point in the
audience.
In the horizontal plane Δdist changes the same as when the subs were in
the air. The change in distance is 1.38m.
This leaves us with the question of depth for mic placement, which we'll
consider in the next section.
System 3: Uncoupled—Horizontal
Displacement
In another variation on this theme, we can look at the case of a center-
placed sub with uncoupled mains. Now we have distances that are matched
over depth and unmatched over width.
Over depth, the difference in distance offset is 1.31m ; over width the
offset is 10.23m.
System 4: Uncoupled—V+H Displacement
A center sub on the ground with uncoupled mains in the air is a complex
design that introduces displacement in both vertical and horizontal. Every
point in the audience will experience different arrival times.
Let's look at how to deal with these various forms of asymmetry.
Looking at a measurement taken from on-axis with the left side, it looks
like the left and right main arrays are close enough in level for combing .
If we combine the left main array with the sub array then their combined
level takes us to the edge of the combing zone.
Observing the phase graph confirms that measuring the left and right
side together does add ripple (arriving 50ms late), but does not significantly
change the phase.
If I switch to ⅙ octave smoothing, they look even more similar.
F S
System 2: Subs On The Ground—Vertical
Displacement
Petal Aim
Let's dip our toes into the problem by looking at a section view of our
design. Here's a prediction at 63Hz with no processing added. On the right
you can see magnitude and phase measurements of main and sub in red and
blue.
Some things to notice:
I like to use Merlijn van Veen's Sub Align calculator . After inserting the
speaker positions and audience plane you'll need to choose a crossover
frequency. In Merlijn's instructions, he recommends choosing a frequency
that is ⅓ octave above the crossover frequency of your system. If your
crossover frequency is 80Hz, then 80 * 1.26 = 101Hz.
For further investigation you could use the highest frequency in the
crossover region, which may provide greater accuracy with wider crossover
regions. In this example, the highest frequency where the magnitude
relationship is within 10dB is 92Hz.
Now you can adjust the phase offset until you discover your preferred
balance of coupling and tonal uniformity.
If you don't have information about the crossover region, then use
Merlijn's suggestion and focus on balancing the phase span.
Here are the basic steps:
1. Zero offset line: This method will focus on discovering a line through
the audience where the offset between and main and sub does not
change. This is where we have the greatest opportunity for summation
or cancellation since magnitude levels will be matched.
2. Audience split: This method will focus on balancing alignment errors
evenly across the audience.
To find the zero offset line in this design, simply find the halfway point
between on-axis with the main and on-axis with the sub (see isosceles
triangle).
Looking at a plan view , you can see that there is comb filtering with its
center on the zero offset line. Here you can see a combined prediction on
the right.
For errors to be spread evenly across the audience, though, we'll need to
synchronize our arrivals at the center of one side of the audience, which just
happens to be on-axis with the main. If you count from the center petal of
summation you can see that we are almost three cycles out by the time we
get to the center of the audience, but you'll see that it is the same in the
other direction as well, and therefore, balanced.
A combined prediction reveals the added benefit of less summation and
power alley down the middle of the audience.
We can also observe this phenomenon in Sub Align.
The audience split is totally dependent on the audience dimensions. The
speaker positions may change, but the mic stays the same all day.
System 4: Center Ground Sub—V+H
Displacement
Where is the zero offset line after you throw vertical asymmetry into the
mix?
Merlijn van Veen's article recommends a head height intersection with
the median plane . We'll use the depth from System 1 using the Sub Align
calculator.
Here's a prediction using the zero offset line location, keeping with the
convention of showing left main and sub together at 80Hz, then left, right,
and sub.
And here's the prediction for an alignment to split the audience.
FOH
Aligning at FOH is, of course, monarchy. But, as long as the FOH depth
is at 50-100% of the audience, it is a sensible option that will avoid putting
the mixer in a null on accident. This is important because the decision of the
mixer affects everyone. This is especially relevant to the system designs
with more asymmetry.
Imagine if you went to all of the work we have discussed so far to
choose the best alignment position for the audience, but then FOH ended up
in a big area of cancellation? (see :( below)
In this case, you may prefer to move the alignment position to FOH. (see
:) below)
A P
Creating an average between multiple measurement locations can be a
lifesaver in the field. When done correctly, it can remove many local
reflections and noise for more actionable data. The trick is to keep your
measurement locations in the coupling zone and balanced around the
alignment position to get a valid average.
Why 50Hz? Because that's the highest frequency at which the sub array
is still coupled. A conservative estimate I often use is 2/3λ.
f = 345 / distance * ⅔ = 57.5Hz
highest coupling frequency = speed of sound / distance * ⅔
No one asked me, but I can't help but comment on the system design
here. A wide spread of speakers evenly across the proscenium is pleasing to
the eyes, but lowers the operating range at which the subs can couple
effectively and narrows the coverage. In this case that may be exactly what
the designer in tended , but it gives me an opportunity to make a point.
One particular conflict of interest with respect to horizontal arrays is the
desire to extend the line length to the width of the stage or venue. Narrow
venues require long lines whereas wide venues require short lines
regardless of the width of venue or stage. -Merlijn van Veen, Low
Frequency Control
Floor and ceiling reflections are the ones that lead to more risk of
error. (L-Acoustics 8)
The proposed measurement area is limited to [1/4 – 3/4] of the
audience depth, with no measurement locations closer than 2 meters
from the side walls. (L-Acoustics 6)
First-order reflections are usually energetic and cannot be dissociated
from the direct sound in contrast to higher-order reflections and
reverberation. This makes any time windowing method inefficient to
remove them from the measurements. (L-Acoustics 2)
F
My intention with this document is not to make anyone wrong. Instead,
it is a celebration of the complexity of acoustics and loudspeakers. If it's not
fun, why do it? :)
Acknowledgments
A big thank you to the people who helped me edit this document and
provide feedback.
Terry Nelson
Dwayne Roberts
Adam Hill
Paolo Tommasini
Elis Bradshaw
Harerton Dourado
Vincent S. Du Beau, Ed.D.
Tucker Burnes
Ivonne Martínez
Samantha K. Cristobal
B
L-Acoustics. Optimum measurement locations for large-scale
loudspeaker system tuning based on first-order reflections analysis . Audio
Engineering Society, 2019.
McCarthy, Bob. Sound Systems: Design and Optimization: Modern
Techniques and Tools for Sound System Design and Alignment . Taylor and
Francis, 2016.
van Veen, Merlijn. Know Your Time Records . Merlijn van Veen, 2020.
Merlijn van Veen , https://www.merlijnvanveen.nl/en/study-hall/178-know-
your-time-records.
van Veen, Merlijn. “Subwoofer Alignment: The Foolproof Relative /
Absolute Method.” Merlijn van Veen ,
https://www.merlijnvanveen.nl/en/study-hall/166-subwoofer-alignment-the-
foolproof-relative-absolute-method. Accessed 5 October 2020.