To Splay Or Not To Splay

My introduction to combining speakers properly came from reading Bob McCarthy’s book “Sound Systems – Design & Optimization” but it was during his SIM3 class when he demonstrated the approach that I clearly understood.

On paper the concept seems rather simple.

“-6dB + -6dB = 0dB”

Combine speakers at their -6dB point and they sum back to 0dB.

A key piece of information is that manufacturers state speaker coverage based on their -6db point. If a speaker has a stated coverage of 90 horizontal x 60 vertical, that means that traveling from 0 degrees on axis (ONAX) in an arc to being more and more off axis (OFFAX), when you get to 45 degrees in either direction horizontally or 30 degrees vertically, the mid / high frequencies will have dropped by 6db. If you need more horizontal or vertical coverage than that, you will need another speaker.  The not so obvious question to ask is “at what frequency am I trying to combine them?” Speakers are not inherently linear in directionality.  In general, speakers are omni directional at lower frequencies and more and more directional at higher frequencies so you have to consider which frequency you’re paying attention to.  In order to clear up my own doubts, I recently wrote Bob to get his take on choosing the right frequency to combine speakers and this was his reply:

“6 db – the flat part of the beam width – typically 1k above but with small boxes could be higher. E.g UPA above 1k, UPJ 2k, junior 3k , UPM 4K”

I’ll add that speaker directionality is a moving target. Here is an example of what I am talking about. A directional plot for a QSC AD-S12.


Pay attention to the red part which represents odB and where the yellow starts which represents -6dB. Between 2k and 15k, this speaker is rather consistent in directionality but it is by no means linear. At the two extremes, the cabinet is omni directional below 100 (not shown) and coverage begins to narrows above 8k. Where is -6db. I would take Bob’s suggestion and choose around 2k for this cabinet.

Let’s combine some generic speakers and see what happens. One of the mistakes we can make when assembling a sound system is to put two or more full range speakers side by side in a horizontal array and not to splay them at all. Like this (0 degrees):
The result of a 0 degree splay is broad band comb filtering, something to methodically avoid. You’d be better off unplugging one of them.


If you don’t already know what comb filtering is, please read this article: – What Exactly is Comb Filtering?


The opposite scenario would be to aim the speakers in opposite directions (180 degree splay) which would be silly. Like this:


Believer it or not, 180 degree splay between two cabinets would sound better than a 0 degree splay. In this case you would have no high frequency combing and extra in the low end. If it were me, I’d still unplug the one facing the wrong way.


Options beside 0 and 180? Some might logically assume that a trapezoidal cabinet’s sides indicate the proper splay angle and simply butt them together like this:
In some special cases the sides of a speaker cabinet do indicate the correct splay angle but in general the side angles of a trapezoidal box do not indicate the proper splay angle so it’s not an assumption we want to make.


As a starting point for discussing proper splay, I chose the Meyer UPA as a point of reference because it was the first trapezoidal cabinet.

Here is an article that explains what made the UPA special: – 1980 Meyer Sound Labs UPA 1 Arrayable Trapezoidal Speaker

“From the outside, the UPA-1 was decidedly different: It was the first trapezoidal speaker (U.S. patent D271,967), now a common practice within the industry. Its sloped sides allowed the creation of tightly packed, wide-coverage horizontal arrays to minimize the comb-filtering effects that occur when spaced drivers reproduce the same frequencies.” – patents


How does one splay Meyer UPA speakers for the right amount of overlap? Meyer Sound has done all the research for us and published the results. The proper splay angle for combining UPA-1P cabinets is 70 degrees. Like this:

On the other hand, the proper splay angle for combining two UPA-2P cabinets is 30 degrees which just happens to be the angle of the cabinet side. Like this:


The UPA-1P & UPA-2P share the same cabinet dimensions but use a different horn, have different coverage patterns and require different splay angles for proper combining.

The UPA-1P has a coverage patter of 100 degrees horizontal x 50 degrees vertical (-6dB). – 2 UPA-1P spec sheet PDF
The UPA-2P has a coverage pattern of 50 degree horizontal x 50 degree vertical (-6dB). – 2 UPA-2P spec sheet PDF

Because the sides of a UPA-2P indicate the proper splay angle, you can place them with their sides touching in what Meyer calls a “tight pack”. – 2 UPA-2P tight pack – 3 UPA-2P tight pack

With a UPA-1P, there has to be a gap at the front of the cabinets.
This PDF article explains how to properly combine a pair of UPA-1P. – 2 UPA-1P 70 degree splay

Seems simple enough but in the following PDF, Meyer states that the same cabinets can be splayed 85 degrees apart. What is the catch? You sacrifice a linear frequency response for more coverage.  This suggests that while there is a correct splay angle for proper combining, there is some wiggle room. – 2 UPA-1P 85 degree splay
In this Q&A PDF, Meyer provides the following details: – UPA-1P Q&A


In the owners instructions for a UPA-1P, Meyer Sound has the following to say regarding designing arrays using UPA-1P cabinets : – UPA-1P operating instructions
“Array Design
Creating an effective array with the UPA-P requires a precise understanding of how to combine the coverage
area and SPL of the individual speaker with those of adjacent speakers. Array design is a trade-off between
increasing on-axis power and creating smooth transitions between the coverage areas of adjacent speakers.
As the splay angle (the angle between adjacent cabinet faces) decreases below the coverage angle of the indi-
vidual speaker, the power at the center of the array increases, but the coverage overlap between adjacent speakers
causes comb filtering and other frequency response variations. As the splay angle increases toward the coverage angle, the power at the center of the array decreases, but the variations in frequency response diminish. As the splay angle increases beyond the coverage angle, noticeable gaps begin to form in the array’s coverage area.”


Before we move on, it’s important to recognize that some cabinets have a rotatable horn. In those cases, obviously we need to verify the orientation of each horns in order to properly combine them and if necessary we may need to rotate them for our purposes (more vertical coverage or more horizontal coverage). A Meyer UPJunior has a rotatable 80 x 50 horn. – UPJunior spec sheet PDF – UPJunior operating instructions PDF

This is what Meyer has to say about splaying UPJunior’s properly for horizontal and vertical arrays:


As you can see by now, Meyer Sound goes out of their way to make sure their users have the information they need to make good decisions. This is not true of most speaker manufacturers. In fact I can think of only one other company that might provide those details as part of their product literature. What this means is that for most speakers, you will have to figure it out for yourself.  The first thing to do is to find the spec sheet for the speakers you will be combining so that you know what the stated horizontal and vertical coverage pattern is. From there you can either make your decision based on rules of thumb or use your measurement rig to find the proper splay angles.

For measurement demonstration, I have access to some QSC KW122 cabinets to illustrate the concept of splay & how proper splay affects your end result (seen via the measurement results). Note that unlike the Meyer UPA, the KW122 cabinet is not a trapezoid so the there is no method for achieving a “tight pack”. This means that the horns of KW122 array can only be so close together which limits your ability to combine them optimally: – KW series spec sheet PDF
A KW122 has a 75×75 degree horn. What QSC called Asymetric 75 degree. If you want to combine two KW122, a good starting point is 1/2 of the stated coverage pattern. In this case each KW122 is worth 75 degrees of coverage and combing two cabinets side by side would suggest a splay of 37.5 degrees apart.

INSERT 0 degree splay trace

INSERT 37.5 degree splay trace

INSERT 75 degree splay trace

What if the cabinets you have to work with don’t have a make and model shown (possibly home made) and your time is running out? This is where having a measurement rig is helpful. You would establish a base level and frequency response on axis (ONAX) of one of the speakers and then move the mic in an arc or better yet, leave the mic in place and rotate the speaker until you find the -6db point @ 2k. This experiment would indicate an approximate splay angle. If the -6db point was 25 degrees off axis @ 2k, you would need roughly 50 degrees of splay between the center of two of those speakers to combine them properly.

To illustrate the point that there is no “right” way to splay speakers but instead many opinions and many variables, you may enjoy reading the following thread on the forum – Setting the proper splay angle

Note that some of the posts on that thread come from widely respected individuals in the industry. “It depends” could be the stated consensus. Distilling what I read, let’s go with this.

1. Ask the manufacturer for guidance. If anyone has a vested interest in making sure their speakers are used together properly it will be the manufacturer.
2. Use the stated coverage pattern of the cabinets as a starting point and adjust as necessary.
3. Measure.

If the manufacturer states how to properly combine their speakers, that is where I would start. I would verify that their stated method approximates the process previously described. If the manufacture doesn’t provide any information regarding combining their speakers (typical), I would ask them for guidance. If they can offer none, do the measurements and figure it out for yourself. Then share your findings publicly.

Ignorance is NOT bliss and with a measurement rig, you can stop being ignorant and stop relying on the opinions of others. Once you have learned for yourself the proper splay angle for a given speaker cabinet combination, you will be on your way to better sound and the confidence it takes to make good decisions about combining speakers in general.


I just stumbled across SpectraPLUS by Pioneer Hill Software. The PC based application seems to be in use by various industries.
spectrasc_main_small – screenshots

Note that there are (3) versions of the software and each one is available as a 30 day unlimited functional demo. – downloads

The websites technical articles reveals some of what the software is being used for, – technical articles

Typical Applications

Audio equipment testing
Vibration testing
Distortion measurements
Frequency Response tests
Room Acoustic measurements
Noise measurements and monitoring
Precision signal generation
Sound Power measurements
Musical instrument manufacturing and testing
Production testing
Underwater acoustics, pile driving noise levels
Rotating machinery analysis
Impact testing

Key Features

Real-time spectrum analysis of live input
Record, Playback and Post Process WAV files
Displays: Time Series, Spectrum, Spectrogram, 3-D Surface, Phase
Full Featured Dual Channel Signal Generator
High Resolution FFT Analysis up to 1,048,576 pts
Octave Analysis from 1/1 to 1/96 octave
Up to 24 bit precisionShenand
Digital Filtering, Distortion Analysis, Transfer Functions
Acoustic Tools: RT60, Equivalent Noise Level (Leq)
THD+N versus Frequency
Acceleration, Velocity and Displacement
Order Analysis
Data Logging
Advanced Programming API

minidsp 2×4 balanced dsp unit – in use

This past week I was able to install a miniDSP 2×4 balanced dsp unit in a rack that powers a 2 speaker rehearsal sound system for the ballet company I work with.  The device worked out perfectly. Right price, right size.

In order to use the device you have to choose and purchase a plugin. – 2 x 4 plugin options

In my case, I needed basic routing and EQ so I went with 2×4 Advanced option: – 2 x 4 Advanced plugin

While the 2×4 Advanced plugin did what I needed to do, the user interface could use some refinement.

Here is a Smaart 8 trace showing the response of the left speaker before and after processing with the minidsp device. I eqed each zone for a relatively flat response via the input EQ and then after combining them made a small low mid cut to both.  I’ve used 1/3 octave smoothing on the first trace to make it clear what is what. Brown is pre eq, green is post eq. Note the coherence issue (red) between 1k and 2k.  rehearsal sound system pre and post eq

This is the same trace but with 1/48 octave smoothing instead of 1/3:

rehearsal sound system pre and post eq no smoothing

This following image indicates the EQ settings used in the miniDSP 2×4 advanced plugin app. Note that there are 5 parametric eq bands available on the inputs and I used 3 filters:

minidsp rehearsal sound system eq 1

Once I measured the system with both speakers eq-ed as show above, I added one more filter in the output section of the miniDSP plugin to balance out the LF.

minidsp rehearsal sound system eq 2

After I configured the single speaker EQ, I measured at different places on axis of that speaker. The following trace indicates those results. First 1/3 octave smoothing and then 1/48:

rehearsal sound system post eq 4 different locations smoothing


I may replace the power amp in the rack with an amp that has built in DSP soon and if so, I won’t need the miniDSP 2×4 balanced device for that purpose any more and can use it for something else like tuning my studio monitors. Neat device!

RF Venue – RF Spotlight antenna

A local venue that I support has recently suffered some serious RF (radio frequency) interference for the first time. Not surprising since here in the DFW area, there is less and less RF spectrum for wireless mic purposes. I contacted RF Venue and explained the RF circumstances. RF Venue staff recommended using (2) RF Venue RF Spotlight antennas with their Distro 4 antenna distribution device as a starting point to avoid further RF interference at the venue. This is what an RF Spotlight looks like:


The next link below will take you to a webpage where you can watch a demonstration comparing a RF Venue CP Beam directional antenna with an RF Spotlight. Very impressive.

RF Venue demonstrating the RF Spotlight antenna in Central Park

My initial tests are very promising. The following images were created using a RF Explorer hand held RF scanner with RF Venue’s VANTAGE scanning software for Mac. The first image is of the immediate RF environment using a whip antenna (like those provided with most wireless mics and in ear monitors). The second image is the substituted RF Spotlight antenna. Notice the drop in almost all levels and a significant drop in general noise floor.

This scan was taken with the whip antenna

rf venue rf explorer rack pro whip antenna

This scan was taken using the RF Spotlight

rf venue rf spotlight antennaTomorrow I will do a scan at the venue with their wireless mics on as a reference point and see what sort of reduction there is in hostile RF conditions. More soon…

minidsp 2×4 balanced dsp unit

I have been looking for a small dsp unit to use on some projects where space is a consideration and I finally decided on minidsp’s balanced 2×4 device. – balanced 2×4
I recently got to play with the non balanced version of the same device (RCA inputs and outputs) but wanted the balanced version to interface with my other balanced equipment. Out of the box, the unit doesn’t actually do anything. You have to purchase a “plugin” to give it a purpose. In my case, I selected minidsp’s most straight forward plugin, “2×4 Advanced” for $10. There are more elaborate plugins that can perform 4 way crossovers and such but I just needed some basic EQ out of the box. For a complete list of available plugins, this link explains them all. – 2×4 plug-ins
Once you purchase a plugin and install it, the plugin becomes your interface for the device. Once you’ve connected to the device, you can change parameters in almost realtime via the USB connection. Once you’ve finished making your adjustments, you can store the configuration to the device and unplug it from the host computer. Then the device will stand on it’s own. This coming week I will be installing the 2×4 on a small portable sound system I use on my ballet gig to EQ the speakers. Here is a video that explains the minidsp concept:

Application notes for the 2×4 series include: – stereo 2 way xover – auto eq with REW software – subwoofer integration with minidsp

Booker T Washington HS – theater sound system 090616

Today I visited the Montgomery Arts Theater at Booker T Washington High School in Dallas Texas to do a check up on their sound system before their upcoming production of “Oklahoma” begins the tech process.
Booker T Washington Montgomery Arts Theater

The main sound system consists of:
(2) Renkus Heinz PM-3(ST4/44R) Side L / R
(2) Renkus Heinz PM-3(ST4/64R) Main L / R
(1) Renkus Heinz PM-1L(ST4LW) dual 12″ sub woofer
(1) dbx 4800

A few things that slowed the process down. First, the dbx 4800 dsp is located in the amp rack on the 3rd floor of the theater and without a way of controlling it remotely, a lot of time was spent climbing stairs and ladders. Next time I would be extend the dbx 4800 network cable to FOH measurement position. Secondly, during the measurement process, it was discovered that the Main L / R speakers were out of polarity. How do you troubleshoot a connection that runs all over the building? Easy! You use Soundtools XLR Sniffer / Sender tester. What is known in the industry as a “Rat Sniffer”. Unlike a normal cable tester box, the two parts of a Rat Sniffer / Sender kit can be used at any location all over the building. This is one of the 2 most important audio tools I use on a daily basis. Get one or more if you don’t have one already. If you have more than one kit, you can troubleshoot multiple lines at the same time.

soundtools rat sniffer
Using a RAT SNIFFER we narrowed it down to a bad XLR cable between the catwalk patch point and the self powered speakers. Resolving that issue made the system much more usable:)

measurement results to follow: – rePhase FIR program – rePhase FIR tool
Quote: “rePhase is a Windows-based freeware program written by Thomas (aka “pos”), a long time miniDSP community member. rePhase generates finite impulse response (FIR) filters that “reverse” the phase shifts introduced by a loudspeaker crossover. rePhase can also generate linear-phase crossovers. With the aid of a real-time FIR filtering engine or “convolver” such as miniDSP’s OpenDRC or miniSHARC, the result is a linear-phase loudspeaker system.”

Acourate – the High-End Audio Toolbox – Acourate


“The powerful software enables you
to measure your audio system
to display, interpret and process measurement data
to carry out multiple mathematical calculations
to generate crossovers and other FIR filters
to identify reverberation times
to establish correction filters for speaker drivers and the listening room
to identify harmonic distortions
to filter WAV tracks
to play music tracks for test purposes”

FIR-ward Thinking: Examining Finite Impulse Response Filtering In Sound Reinforcement Systems by Pat Brown – FIR-ward Thinking: Examining Finite Impulse Response Filtering In Sound Reinforcement Systems page 1 – FIR-ward Thinking: Examining Finite Impulse Response Filtering In Sound Reinforcement Systems page 2 – FIR-ward Thinking: Examining Finite Impulse Response Filtering In Sound Reinforcement Systems page 3