• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Frequency affected when speakers near boundaries.

Cote Dazur

Addicted to Fun and Learning
Joined
Feb 25, 2022
Messages
651
Likes
833
Location
Canada
When placing a speaker near a boundary, say just the front wall, the gain is said to be 3 db, but at what frequency does the gain start to occur and what does it looks like? Just a flat 3 DB gain from that point?
 
Google SBIR (Speaker Boundary Interference Response)
 
Is this what you're looking for? (Source: KEF KC62 White paper).

1657548547102.png
 
Google SBIR (Speaker Boundary Interference Response)
You are so smart, thank you for your insight.
Is this what you're looking for?
It is an interesting graph. Would “average room gain” be 3db gain from typical one boundary gain when speaker place at the back wall? Most of the gain seems to start at 150 hz and flat from there. Interesting that it also does not seem to be listening level dependent.
Also KC 62 is a woofer/sub woofer, not a full speaker.
 
Last edited:
You are so smart, thank you for your insight.

It is an interesting graph. Would “average room gain” be 3db gain from typical one boundary gain when speaker place at the back wall? Most of the gain seems to start at 150 hz and flat from there. Interesting that it also does not seem to be listening level dependant.
Doesn't "average" like that. Boundry gain increases as frequency drops. "Room boom" happens when you stick any speaker in a corner. The nature of that boundry gain also depends on the shape of the room. Klipschorn's use boundry gain to maximum effect:

R.jpg
 
When placing a speaker near a boundary, say just the front wall, the gain is said to be 3 db, but at what frequency does the gain start to occur and what does it looks like? Just a flat 3 DB gain from that point?
https://www.genelec.com/monitor-placement (primarily below 200 Hz or so, but there is somewhat of a hinge effect starting at 500-1000 Hz. Can see here https://www.audiosciencereview.com/...-wall-can-this-be-settled.29550/#post-1034140, keeping in mind that free space is pretty flat frequency response
 
When placing a speaker near a boundary, say just the front wall, the gain is said to be 3 db, but at what frequency does the gain start to occur and what does it looks like? Just a flat 3 DB gain from that point?

See Figure 8, here, from Allison’s paper on the subject.

For ratios of distance from boundary over wavelength <= 0.1, the gain is approximately 6dB SPL. This gain disappears at ratios > 0.2.

eg: wavelength of 170Hz is ~2m, so full boundary gain is achieved if the source is <= 0.2m (8 inches).

May not be totally achievable, practically though - leaky walls, etc.
 
(primarily below 200 Hz or so, but there is somewhat of a hinge effect starting at 500-1000 Hz.
Interesting, but the graph (at 6db)
does not show the frequency value on the lower axis but it shows that it seems to raise quickly and stay flat.
monitorplacement-radiation-space.jpg

On this site where, supposedly, all serious audiophile measure everything, I would hope, many (some) people have measured what I have asked in OP and could share their measurements. Keeping in mind that every room will be somehow different, but the principle should remain the same.
 
eg: wavelength of 170Hz is ~2m, so full boundary gain is achieved if the source is <= 0.2m (8 inches).
and would remain constant at 6db gain, flat, for all subsequent frequency. if the speaker is flat itself?
 
The system graphs include the woofer's roll-off so do not accurately reflect SBIR.

Allison's paper referenced by @dc655321 includes plots showing the "native" effects of boundaries on response as well as numerous plots of various test systems.

As the speaker is moved away from boundaries, direct coupling is reduced, but reflections still occur from surfaces on the way to the listener, making analysis more complicated.
 
making analysis more complicated.
Yes, absolutely, the analysis is extremely complicated and room dependent.
I would just like to look at what some measurements from forum user look like.
So far it seems they will look like a sharp 6 db rise around 150Hz and flat after that, if the speaker is flat itself.
 
Interesting, but the graph (at 6db)
does not show the frequency value on the lower axis but it shows that it seems to raise quickly and stay flat.
monitorplacement-radiation-space.jpg

On this site where, supposedly, all serious audiophile measure everything, I would hope, many (some) people have measured what I have asked in OP and could share their measurements. Keeping in mind that every room will be somehow different, but the principle should remain the same.
I'm a recovering audiophile, and like most "Serious" audiophiles I didn't pay the close attention to scales and graphs the way the E.E.s would. Didn't stop me from making recordings. Some people at ASR will give you the data you seek. This one will tell you that acoustics does not boil down to one single, all-purpose equation. Rooms are very different: an acoustically awful domestic office can destroy the sound quality of some mighty fine audio gear, I've heard it. The best domestic replay I've heard had the room designed around the audio gear: in the early days of digital editing, Jack Vad built a house in Marin County with an audio editing room, the whirling hard drives in a padded, ventilated closet. That system, with big NHT floorstanding speakers and [probably] Krell amps had the scale of orchestra recordings thanks mostly to the design of the room. Somewhat like a concert hall, a horn, with the "stage" expanding outward.

In any case, these "graphs" look more like representative illustrations that actual frequency responses. The nature of acoustics in most domestic environments involves lowish ceilings and smallish proportions that help produce ripples and bumps in the upper bass/lower midrange.
 
and would remain constant at 6db gain, flat, for all subsequent frequency. if the speaker is flat itself?

Theoretically, yes.
Practically, the gain is a function of the number of boundaries, their material, and the room modes in play.
As @DonH56 said, it's complicated.
 
Interesting, but the graph (at 6db)
does not show the frequency value on the lower axis but it shows that it seems to raise quickly and stay flat.

On this site where, supposedly, all serious audiophile measure everything, I would hope, many (some) people have measured what I have asked in OP and could share their measurements. Keeping in mind that every room will be somehow different, but the principle should remain the same.
You didn't ask in the OP for people to share measurements ., but Toole's book Sound Reproduction does include the ones I linked. Your "flat 3 dB gain" was not accurate, so I provided links otherwise, which I thought were fairly self-explanatory. The amount of gain would depend on proximity, and it iwould not be flatt below a specific frequency but rather highly dependent on wavelength and distance, hence @Kal Rubinson's mention of SBIR.
 
I would just like to look at what some measurements from forum user look like.
So far it seems they will look like a sharp 6 db rise around 150Hz and flat after that, if the speaker is flat itself.
Ok, I will oblige your desire for measurements!

Be warned, I think this is going to be a lot more complicated and nasty looking than you expect...
(My setup is very much still a work in progress).

I have several identical satellite speakers (B&W M-1, Mk2, not popular here) each mounted differently, which gives the possibility of comparing different wall proximity situations.

The measurements are taken during Audyssey setup. They are the average of 8 mic positions (all of which are quite close to the rear wall, meaning that LBIR comes into play).

Least bass is generated by the front wide speakers (near a side wall)...
20220711_201713.jpg

Screenshot_20220711-200331_MultEQ.jpg

(X-over=90Hz, Level=+6.0dB)


Next highest bass is from the rear surrounds (near a corner)...
20220711_201728.jpg

Screenshot_20220711-200318_MultEQ.jpg

(X-over=60Hz, Level=+4.0dB)


Then the front height speakers (front wall mounted, near side wall and fairly near ceiling)...
20220711_201756.jpg

Screenshot_20220711-200345_MultEQ.jpg
(X-over=60Hz, Level=+4.5dB)


Finally, rear heights (right in a corner and close to the ceiling)...
20220711_201735.jpg

Screenshot_20220711-200402_MultEQ.jpg

(X-over=40Hz, Level=+4.5dB)


As you can see, the inherent speaker response isn't very flat anyway (Audyssey is working very hard). However, there are clear differences in the amount of bass, and therefore the auto detected frequency for crossover to the sub, depending upon wall proximity.

(Ignore any differences right of 1kHz, only the lower frequency differences are likely to be due to the proximity of the walls).

Edit: Sorry I compressed the photos which has caused aliasing artifacts on the grilles.
 
Last edited:
When placing a speaker near a boundary, say just the front wall, the gain is said to be 3 db, but at what frequency does the gain start to occur and what does it looks like? Just a flat 3 DB gain from that point?
Thought experiment with one boundary only, the front wall: when two coherent sounds are spaced apart there is constructive and destructive interference towards an observation angle because distance to each source is different, a path length difference.

Destructive interference happens when the two sounds are in opposite phase at any particular frequency, either is lagging 1/2wl and multiples of, and constructive interference happens when the two are in phase, roughly one wl apart or multiples of. As wavelength changes with frequency this relationship changes and we have an interference pattern in frequency response where dips and peaks alter, a comb filter. Onset of the comb filter starts with first null, when the two sound sources are 1/2 wavelength apart. Below this frequency and especially below about 1/4wl when the two sources are less than 90 degree apart in phase only constructive interference happens. This is your boundary gain. I'm ignoring room modes and room gain and such, considering only the one boundary for simplicity.

Now, your speaker is one sound source and reflection from the front wall is the other. Lets imagine reflection from the front wall is identical to direct sound, only delayed by path length from the speaker baffle to the front wall and back towards listener. The reflected sound has to travel two times the distance the baffle is from the wall. This is roughly the path length difference and constructive interference happens on frequencies whose wavelength is 4 times or more, or 8 times the distance from the baffle to wall.

More numbers: 200Hz is roughly 160cm long (for easy math) and to couple (to get the gain) with the wall the speaker baffle needs to be roughly within 20cm from the wall. First null of combfilter would happen at 80cm long wavelength (~400Hz) and repeat every 400Hz so 800Hz, 1200hz etc. would have a dip.

If your speaker enclosure is 40cm deep, not too uncommon, and butt against the wall, its 80cm roundtrip which is half wavelength of 200Hz and a nice dip right there. Gain happens from roughly 100Hz down. Small 20cm difference to the previous example made our 200Hz from boost to null, there is huge difference due to the reflection making a roundtrip.

Due to roundtrip of the reflection from front wall might be the longest reflection path of all first reflections making the lowest interference combilter dip, ceiling reflection might be longer, or back wall, or even a sidewall reflection depending on the listening setup and ironically floor reflection has about always the shortest path length difference to direct sound, highest frequency onset of combfilter from all the first reflections. Typically many of these first reflection combilters would pile up and make one massive dip to suck out all kick bass and the a rollercoaster above that up until high enough frequencies speaker is directive enough and less sound heads to the first reflections. Basically the boundary reflections in a room boost the lows and "destroys" the midrange. Getting enough directivity at midrange to reduce the interference, say from 200Hz or even from 1000Hz up needs big speaker, waveguide or array. Or perhaps hearing system just ignores it all as long as the reflections are similar to direct sound and less attenuation, smaller speaker is fine. What ever the speaker is positioning it and the listener changes the interference a lot.
 
Last edited:
Back
Top Bottom