Speaker placement

[scott2000]

Maybe it's an optical illusion but it appears to me that speaker placement in some mastering and control rooms doesn't adhere perfectly to the triangle thing. Looking at some of the giant PMC for instance, it doesn't seem possible that the tweeters would be at ear height..... The distance from left to right is wider than you'd expect in some rooms... obviously with multiple speaker setups, something has to give here......
Hope that makes sense....

In the back of my room where there's a couch, it sits where your ears are probably a foot lower than the tweeter of my larger speakers but the imaging is really good... it does make the image appear about a foot up I'd guess.. Fine for movies since the screen is at the same height ...
The distance between speakers and to seating position is equal here though.....just height is off....

Is there more to it than the triangle? Any extra thoughts on this? Was thinking about raising the couch 12" with some blocks but idk ....

 

[pucho812]

I bet you they eyeballed it and said good enough. Two words that ruined our industry good enough.

 

[john12ax7]

Many things in acoustics get misapplied.

The equilateral triangle is one of those things. For a 2 speaker setup you are better off optimizing the speaker and listening position to give the best modal response instead of forcing equal distance.

And for the height you actually want your ears to be at the acoustic axis of the particular speaker, which will depend on the design. The speaker manufacturer should tell you where this is on the speaker. If you don't have this information the tweeter can be used as a proxy.

 

[JohnRoberts]

and then there is the room.

JR

 

[PermO]

Have you tried tilting the speakers slightly ?

 

[scott2000]

Have you tried tilting the speakers slightly ?

no I haven't.
Never felt the need to but I'm making some small changes and was thinking maybe it's a good time to decide if I need to adjust some things.

I guess my main question is if I'm looking at some of these mastering or control rooms wrong by seeing speakers placed where it deviates from the triangle method.
Some of these rooms it's obviously not some half baked willy nilly decision. I guess john12ax7 is hinting at some of this ...

 

[scott2000]

ahh ..
my computer is offline and this phone use is as clunky as I can stand... 5G isn't fast....what is 1kb/s....oh wait ..zero kb/s

so I'm guessing I'm off vertical axis by my one set of speakers being 12" higher?

How would I determine the degree that I'm off axis if it's 104" to seating position with the tweeters being 12" above my ears?
or am I not understanding this right?

 

[kags]

The 60 degree triangle is a good starting point - but some speakers and some music sound better with a different angle. I’ve heard great setups with 80-85 degree angles and some slightly less than 60.

Similarly, you can predict that the modal response will be most even around 38% from the front wall… or the rear wall. So if you decide the most even bass response is 62% from the front wall, your speakers may wind up in the front corners.

Now we get into space loading in the corners, which increases the LF output- good or bad- up to you.

If you love the speakers away from the walls, the space-loading goes away but the speaker-boundary-interference becomes unmanageable.

You can see the balances of angle vs. listening position vs. speaker placement.

No good rule goes unbroken…

I say start with equilateral triangle (with the point being just behind the listener’s head and adjust from there.

 

[k brown]

I wouldn't be surprised to see unequal L/R distances in some control/tracking rooms; there they are much more concerned with balances than the fine points of stereo imaging; but in a mastering room, that would be bordering on malpractice.

True that tweeter height is irrelevant; your ears need to be at the point where the drivers' sounds arrive at the same time - this can't be determined visually (and with poor speaker designs, never) - only by careful listening, usually to mono pink noise. One listens for the cabinet vertical tilt and/or height where the pink noise sounds the smallest, like a point source, indicating minimum time smear.

The main thing that may require deviation from the equilateral triangle (except maybe individual taste) is the influence of room side wall reflections, which can alter the impression of the stereo width.

In a well designed live-end / dead-end room, the equilatral triangle results in a well balanced left-center-right spread for most people. A great test is sending a mono sound to the speakers; for a centered listener exactly equidistant from the speakers, it should sound as well-focused as though there were a third speaker in the center. Flipping phase on one channel should result in a completely unfocused sound that 'comes from nowhere', or even from behind.

 

[scott2000]

One listens for the cabinet vertical tilt and/or height where the pink noise sounds the smallest, like a point source, indicating minimum time smear.

Is this as simple as it sounds?
I could actually find that the speaker's tweeters being 12" higher than my ears almost 9' away would possibly yield this.?
I guess the off axis response and everything else plays into this?..

 

[k brown]

Is this as simple as it sounds?
I could actually find that the speaker's tweeters being 12" higher than my ears almost 9' away would possibly yield this.?
I guess the off axis response and everything else plays into this?..

Off-axis response doesn't affect time aligning the drivers; it will affect the tonal balance of the room reflections. To find the right height/tilt, you just have to conduct the test I described; not as simple as just measuring that the different drivers are the same distance from you - crossovers can introduce phase shifts/delays. You just have to do the listening tests.

 

[Disco Volante]

Making a number of FR measurements at varying heights will give you an idea of the phase characteristics/coherency of your specific speakers (and your room).
It's quite likely that the elevated position of your speakers serves to alleviate a phase mismatch, especially if they are built on a flat baffle.
It is possible to determinie the "acoustic centre" (the centre of the radius of the radiated wave) of each driver with a series of measurements and then compensate either in the XO or by making a stepped baffle to achieve a better phase response/time alignment.

I learned from the Swedish acoustics guru Ingvar Öhman that the optimal horizontal offset is 23 degrees. Something to do with how we hear and the placement of the ears. I'd have to look up my notes to explain this properly;-/

Speaker building/placement/optimization is a rabbit hole...

Happy tinkering!

 

[scott2000]

I learned from the Swedish acoustics guru Ingvar Öhman that the optimal horizontal offset is 23 degrees. Something to do with how we hear and the placement of the ears. I'd have to look up my notes to explain this properly;-/

Speaker building/placement/optimization is a rabbit hole...

Happy tinkering!

That would be great to see....
Yes....I'm in the rabbit hole now.
Headphones are looking pretty good ...lol

 

[CalavoBob]

Here's a trick that can help lining up those speakers.
There's a box that I made years ago that I think I got from the publication The BAS Speaker from the Boston Audio Society. This box produces a positive only pulse that allows you to adjust the repetition rate of the pulses and the narrowness or width of the pulse (which corresponds to the frequency of the pulse). It is used to determine the proper polarity of drivers in a multi-driver speaker box by observing the pulses over a microphone. Of course you need to know that the mic, and its signal path are in correct polarity too. That's easy to determine in practice if you start with a woofer and a 9V battery, the polarity of which you can see easily. Then check the rest of your signal path with an oscilloscope. So now you can check the polarity of your test mic.
Now on to the subject of this thread. When using this system on multi-driver speakers, you can place the mic at the listening position and actually see the arrival times of the tweeter versus the mid. Of course it's easier to see what's going on if you're in a control room with a reflection free zone, but it can help you adjust the positions of the multi-drivers. You can also see how well the driver-amp combination is damped because the drivers keep going back and forth after the pulse! The initial spike is what you're looking for though.
We used to use it to slide tweeter boxes forward and back to align large PA systems for concert sound. It works pretty well. At one gig we noticed that the pulse from one of our horn-driver units looked malformed, so we opened it up and found a cracked horn driver.
I've used it to go through all my mics to make sure they all have the same polarity. I'll look around and see if I can find the schematic for this thing. It uses a single 15V supply, has a knob for pulse width, a knob for repetition rate, a pulse output, and a sync output for setting up the oscilloscope. A simple diy project.

 

[Disco Volante]

Sounds like a great project. I've been doing something (crudely) similar with lots of measurements in REW, comparing FR and phase graphs, moving the tweeter back and forth. Moving the mike up and down works on finished spkrs...

 

[btyreman]

as far as I am aware the equilateral triangle presumes that you're basically in an anechoeic chamber environment, which 99.9% of studios are NOT like. Even abbey road studios does not follow the EQ triangle approach, if it was perfect then all pro studios would comply to it, for me personally I found that having my speakers parallel to the back wall gives me a less fatiguing sound that is more 'real world' and balanced because the off axis response is very good of my studio monitors, the entire spectrum sounds more even like this.

 

[btyreman]

it's also worth noting that flush mounting speakers into the back wall will always be better than free standing because of diffraction.

 

[abbey road d enfer]

I learned from the Swedish acoustics guru Ingvar Öhman that the optimal horizontal offset is 23 degrees.

You should long know not to believe gurus.
This 23° BS may be valid for a certain speaker, but not for all.
Basing positioning on coincident arrival is good if you have isophase speakers, which are not very common.

 

[abbey road d enfer]

There's a box that I made years ago that I think I got from the publication The BAS Speaker from the Boston Audio Society.

Probably inspired by that
https://reverb.com/item/2448118-scv-electronics-pc80r-mkii-white

 

[CalavoBob]

I couldn’t find the original article on this pulse generator, so I took it apart last night and drew the schematic for it. The U1A op amp creates the raw sync pulse and VR2 adjusts its repetition rate. The U1B op amp finds the center of the single sided power supply and references signal ground to it. The sync signal joins the feedback loop of the U2 and U3 op amps at the summing junction on the input of U2A. The 1N4001 diode slices off the bottom half of the signal in this loop at the output of U2B and the dual VR1 potentiometer selects the pulse width in the loop which also contains two integrators.
Note: Signal ground is not the same thing as -V, which is the negative terminal of the single sided power source.

The first o’scope photo shows the sync pulse on Channel 1(which is normally connected as an external trigger during operation), and on channel 2 is a narrow (high frequency) pulse as selected with the range control. The second o’scope photo shows the same sync pulse, but on channel 2 is a wide (low frequency) pulse as selected with the range control. Sometimes I like to get 2 pulses across the screen to see clearly when the pattern repeats. If you are doing measurements in a noisy environment or have a speaker system with a lot of undamped ringing, its nice to see the known second pulse on the right side of the o’scope traces.

Now I’ve put the sync pulse onto the o’scope’s external trigger input and the original pulse output is Y-branched to Ch 1 of the o’scope and to a power amp feeding an NS-10 speaker. I have a KM-84i placed 16 inches in front of the NS-10 midway between the tweeter and the woofer, and I’m feeding the preamp output of the KM-84i into channel 2 of the o’scope.

In the third o’scope photo you see the narrow hi frequency pulse from the generator on Ch 1, and on Ch 2 is the slightly delayed pulse picked up by the microphone. Notice the first part of the received pulse is a positive spike, but it is followed by the natural ringing of the driver. The downward pulse after the initial spike can often be as big as or even larger than the original spike! The fact that the first spike is positive however, tells us that the system is in correct polarity through the mic, the mic pre, the console out, the power amp, the crossover network, and the tweeter connection. I have gone through each of those junctions with this generator and have proven to myself that the polarity is correct through this system.

Now in the fourth o’scope photo I have cranked the range control over to produce a low frequency pulse. The crossover network will send this pulse primarily out of the woofer. Again, the initial pulse is positive, and again there is a lot of undamped energy both negative and positive that you see after the initial pulse. I have done this test on servo-controlled subwoofers and was amazed at how much control it had in reducing the ringing after the first pulse. A good servo-controlled subwoofer is impressive.

I tried moving the listening position up and down and adjusted the pulse width up and down to see if I could use this to help find a good angle to line up the tweeter-woofer triangulation and it was somewhat difficult to see. You can tell looking at the previous o’scope photos that there is a time delay of just a few milliseconds as the sound leaves the electronic path and goes through the air to the mic. It looked to me as if the woofer has an extra delay of about 1 millisecond. This can easily be caused by the acoustic center difference between the tweeter and woofer when mounted on the same baffle though, and since the voice-coil-diaphram combination of the tweeter is much lighter than the woofer, its going to be able to react more quickly. That’s why some systems use slanted baffles, and others even use digital delays to pull the tweeter back. I’ve experimented with all-pass phase shift networks in active crossovers to correct this with some good success.

Apart from its usefulness in checking the polarity and timing of speaker systems, this pulse generator has been great going through my recording setup to discover if I have any pieces that are still pin 3 hot. It’s a good idea to check through your mic collection to see if any are wired backwards. I’ve found that there are many consoles that invert the signal at the insert points or even at line outs.

To use this thing intelligently when doing tests through the air, you need to first verify that you have a speaker that’s in correct polarity. A 9-volt battery touched with + to the red cabinet jack pin of a woofer should produce a forward movement in that woofer. Now you have a starting reference. Put the pulser into your console and then measure the power amp terminals into the o’scope and prove that the output is in correct polarity. Now when you connect to that woofer you will have a positive pulse that hits a microphone. Any stock Shure SM-57 or SM-58 I’ve ever measured has correct polarity. Use this as a starting reference to then go through your mics and prove they’re all correct.

Here's a BOM for the generator:

(3) TL072 or NE5532, 4558, AD712, etc.
(1) 1N4001 diode
(2) 1 kΩ ½ W resistor
(1) 4.7 kΩ ½ W resistor
(4) 10 kΩ ½ W resistor
(1) 27 kΩ ½ W resistor
(1) 68 kΩ ½ W resistor
(4) 100 kΩ ½ W resistor
(2) ¼” phone jacks
(1) dual 500 kΩ linear potentiometer
(1) single 50 kΩ linear potentiometer
(1) 1 µF/25V electrolytic capacitor
(2) 0.01 µF/25V film capacitor
(2) banana jacks or a 9V battery snap

I made this thing about 35 years ago and actually used NE5535 opamps which probably don’t even exist now, but any good generic dual 8-pin op amp will work.

The last photos show the component layout and the copper side of the board. It looks like I hand etched this one.
Enjoy.