• 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!

How much delayed sound comes through the cone from the inside of the loudspeaker in a closed box ?

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
One of the reasons Linkwitz didnt like the sound of boxed full range speakers was because of the delayed sound coming through the cone and giving coloration to the sound .
Do you have a link to exactly this statement from SL?

One of the reasons Linkwitz didnt like the sound of boxed full range speakers was because of the delayed sound coming through the cone and giving coloration to the sound .
...
"...some in the process of flexing the cabinet walls. Much of it reappears outside the box, because the thin cone presents a weak sound barrier. Just how much is difficult to measure, but it is a contributor to the frequency response. I am of the opinion that the effect is most notable in the low hundreds of Hz region, where stuffing materials are ineffective and the internal dimensions not small enough for the internal air volume to act as a pure compliance. Consequently, enclosures should be either very small (less than 1/16th of a wavelength) or extremely large, both of which are not very practical for different reasons. ”

Okay so I understand everything correctly, let's go through two closed box examples. A midrange with 0.2m cabinet depth and a woofer with cabinet depth 0.5m.

The phase inverted sound of the cone back side is radiated and passes through the damping material, is reflected and inverted at the rear wall of the cabinet, passes again through the damping material and penetrates through the woofer cone with a time delay to the outside - so far simple to understand and logical.

The time delay for the two examples is therefore about 1.2ms and 2.9ms respectively. If we assume that the damping material "slows down" the speed of sound a bit, we might add another 10-20% delay (1.4ms / 3.5ms). The delays are of course correspondingly lower with lower cabinet depths.
The assumption or estimation is now that this time-delayed sound is radiated through the driver cone with only -20dB attenuation and supposedly "smears" the reproduction precision.

Now let's compare this with a concept that is supposed to be clearly superior (has already been mentioned here and was also the choice of S. Linkwitz) - open baffle speaker.
There are no reflections that are reflected back to the driver cone and radiate, with a time delay, through the cone.
But of course it is with open baffle speaker so that the backward radiated, phase-inverted pressure fluctuations, take the detour around the baffle and are radiated to the front with a time delay.

In this case, we can use simulation to determine fairly accurately how high the SPL of the sound directly from the woofer cone is compared to the SPL of the backwards radiated sound on the reference axis of the speaker. The time delay (or phase shift) of the rear radiated sound depends of course on the dimensions of the baffle used. Should be roughly comparable to those of the CB speaker or slightly less.

I recently simulated an open baffle speaker for the ""Analysis" of cardioid speaker radiation via lateral slots - like D&D 8c" thread and reuse the results here. The speaker baffle looks a little unusual, but for our purposes it will do. The time delay would be in this case about 0.3m --> 0.9ms.
Here are front and rear views of the OB speaker and crucially the simulation of front and rear radiation and the resulting overall frequency response all measured on axis:
1665953473995.png 1665951032445.png 1665954717399.png
We now have our first data points. At 400Hz the rear sound is attenuated by -7dB compared to the sound from the front of the woofer, at 700Hz by -15dB.

So if the time-delayed sound coming through the cone of a sealed speaker with about -20dB of attenuation has an audible effect on the main signal, then an OB speaker would have to sound downright awful, since the attenuation of the time-delayed sound in the midrange is just -7 to -15dB.
Unless you argue that the time delay of an OB speaker compared to a CB speaker is on average smaller and therefore less important (1ms delay okay, 2ms delay very bad) - but I don't know of any study on this off the top of my head.
 

Head_Unit

Major Contributor
Forum Donor
Joined
Aug 27, 2018
Messages
1,340
Likes
688
After i stuffed the enclosure and added felt to the walls witch reduced significantly the output of the housing and the sound was more clear.
I've posted elsewhere about modifying my best friend's husband's old speakers while he decided what he wanted to buy new. I added internal bracing, auto body undercoated the inside (the only vaguely antivibration stuff I could get right away), and stuffed plumply with fiberglass. As a loudspeaker design engineer I intellectually expected improvement yet speakers were so transformed it was astonishing. He still won't sell them, 18 or so years later.

As for the leakage through the cone, I did once see a measurement. I think it was done by putting a pulse into the speaker, then looking at the time response. I also vaguely think MLSSA could do a measurement like that? Well, you could get a step response and...I forget.

Thin cones in unstuffed, thin-walled boxes (like my friend's husband's) would be the worst offenders. But open baffle comes with a host of other problems-whether cone or electrostatic these really seemed to need to be out in a large room to sound best. And the reflected sound may be nice, but not really "accurate" to the original which 99.99999999999999% of the time is mixed on conventional speakers.* I think it is better if you have the room to make a very stuffed dead large etc enclosure . The Beveridge electrostatic actually had a back enclosure.

*Steely Dan being an exception. In an interview Donald Fagen said they used I think Martin Logans (Magnepans?) to mix one album. Sorry I forget which. He said later they decided that was not a good idea and didn't do that for the next album.
 

Head_Unit

Major Contributor
Forum Donor
Joined
Aug 27, 2018
Messages
1,340
Likes
688
Do you have a link to exactly this statement from SL?
I don't, but I do recall reading that. I think it was in Speaker Builder. He doesn't say anything on his own site?
 

fluid

Addicted to Fun and Learning
Joined
Apr 19, 2021
Messages
691
Likes
1,196
I think that -20 dB is a very bad attenuation of the internal box-sound. If it is this bad , its for sure a problem. Here we often demands a SINAD of 110 with dacs , but why should we accept a SINAD of only 20 in a closed box loudspeaker ?

I think its strange that noone have real measurements from investigations of this.:(.
Please link or post If you have some interesting measurement.
I wouldn't get too worried based on the test of placing another speaker inside the box. Don has the right idea.

Earl Geddes
"The sound coming through the cone test is completely invalid as in a real situation because the speaker would be playing and cone movement that is not in the signal would be strongly opposed by the amplifier. This is what the impedance, i.e. back emf is all about."

https://www.diyaudio.com/community/...rom-speaker-cabinet-walls.314837/post-5246035


By the time the frequency gets low enough to worry about the room is already mucking things up nicely for most people.

An impedance response is a good way at looking at cabinet and other extraneous radiation that is not from the driver. Compare free air to inbox. With a decent amount of the right kind of stuffing there isn't much to see.

David Smith, worked for JBL, Mcintosh, Snell, Bose

"You still assume that you can let the rear radiation out into the room and treat it to a level that is lower than what good absorption in the box would achieve. You need to rethink that."

https://www.diyaudio.com/community/threads/box-colourations-really.215133/post-3098693
 

fluid

Addicted to Fun and Learning
Joined
Apr 19, 2021
Messages
691
Likes
1,196
Hi @fluid Do you have any practical recommendation on what is best to use?
Thank you
Sure, it does depend on the type of cabinet but for a sealed one felt (automotive sound insulation type) Fiberglass and felt is a good combination. There are other alternatives to the fibreglass that work well but in a sealed cabinet it is usually cheap and practical. Denim is good but sheds, poly batting material varies from having almost no effect to being as good as fibreglass it all depends on the product. The sort of material that is used for wall insulation usually works well and is cheap because it is not an "acoustic" product.

I made some comparisons of different materials here, you can see how the different schemes affect the resonance peak in different ways too
https://www.diyaudio.com/forums/full-range/303417-range-tc9-line-array-cnc-cabinet-post5053920.html
 

Birdy

Member
Joined
Jan 11, 2022
Messages
70
Likes
31
Location
Brazil
I made some comparisons of different materials here, you can see how the different schemes affect the resonance peak in different ways too
https://www.diyaudio.com/forums/full-range/303417-range-tc9-line-array-cnc-cabinet-post5053920.html
Great comparisons, very useful. I noticed subjectively (hearing) that felt on the walls reduced the out going sound passing the wall. Resonance in the wall of the enclosure was apparently much reduced. You could say, how less energy passes through the wall, the better. Some old British designs are based on the shift of the internal resonances to higher frequency's by using thin walls and few or no damping. The higher resonances from the enclosure should be less disturbing, apparently with good results. Looks like the returning sound passing the speaker cone is not seen as that huge isseu! Considering the small surface area of the cone in relation to the large area of the enclosure and the more even dispersion by the cone, maybe the irregular dispersion of the box is more of a problem?
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
Sure, it does depend on the type of cabinet but for a sealed one felt (automotive sound insulation type) Fiberglass and felt is a good combination. There are other alternatives to the fibreglass that work well but in a sealed cabinet it is usually cheap and practical. Denim is good but sheds, poly batting material varies from having almost no effect to being as good as fibreglass it all depends on the product. The sort of material that is used for wall insulation usually works well and is cheap because it is not an "acoustic" product.

I made some comparisons of different materials here, you can see how the different schemes affect the resonance peak in different ways too
https://www.diyaudio.com/forums/full-range/303417-range-tc9-line-array-cnc-cabinet-post5053920.html
Thank you @fluid appreciate it, very nice comparison!

I am currently using 400gr/m2 polyester wad both applied on the walls + fully stuffing the cabinet
Wondering if replacing it with felt would be better or not?
 

No. 5

Active Member
Joined
Nov 1, 2019
Messages
144
Likes
121
I think that -20 dB is a very bad attenuation of the internal box-sound. If it is this bad , its for sure a problem. Here we often demands a SINAD of 110 with dacs , but why should we accept a SINAD of only 20 in a closed box loudspeaker ?

I think its strange that noone have real measurements from investigations of this.:(.
Please link or post If you have some interesting measurement.
A couple observations: SINAD is a measure of signal + distortion components and noise, but what you are describing is just a repetition of the original signal reduced in level with no distortion components or additional noise. Audibility thresholds will be different. to that point, in Olive and Toole's paper "The Detection of Reflections in Typical Rooms" showed detection thresholds for delayed speech to be above -20dB for delays less than 20 milliseconds... mind you, that's for one kind of signal, and the reflection was angularly separated from the direct sound. But it's something to think about.
Capture1-e1441666415941.png


Secondly, a (roughly) identical signal delayed by 2 milliseconds or so and attenuated by 20dB would potentially put about 2dB of ripple in the frequency response, so if those numbers are correct, it should be visible in existing speaker measurements.
 

Birdy

Member
Joined
Jan 11, 2022
Messages
70
Likes
31
Location
Brazil
Thank you @fluid appreciate it, very nice comparison!

I am currently using 400gr/m2 polyester wad both applied on the walls + fully stuffing the cabinet
Wondering if replacing it with felt would be better or not?
I used polyester as well, looks like it does not damp the mid band so much.. after i covered the walls with felt, their was a significant reduction of sound passing the walls (pvc pipe) and less secondary sound sources from the enclosure reflecting trough the room (subjectively). I use both, felt for the walls and polyester to fill up. Bye the way, i use ceramic tiles (5 mm) glued to the front baffle (inside) to create a barrier with different resonance frequency's. Probably the damping helps to transfer the internal sound into heat and i suppose that sound coming back trough the bass cone is very limited. The (poly) cone is not acoustic transparant and will pass mainly it's resonance frequency's i think.
 

SoundGuy

Active Member
Joined
Aug 13, 2022
Messages
115
Likes
120
Sound waves are rather weak - speaker efficiency is a terrible 3% or less. Not much mid range comes out of a totally open tuned port - except in bad designs. Magnet and motor strength are another critical filter. Flea powered speakers with high sensitivity will be the worst as vibrations will not be tightly controlled by the amp. High power designs will tightly control cone movement and filter any back wave. Finally damped fabric inert cones will not ‘ring’ as they do not store energy but dissipate it. Not so for ultra rigid cones which tend to vibrate - hence the rubber dampers on ceramic drivers.

ATC designed a system for the UK Navy that noise cancelled the engine sound from smokestacks on destroyers. The amplifier was able to control the woofers and reduce audible noise. The woofer cone can easily overcome ambient induced vibrations.

Sound waves are weak - this isn’t a big issue or designers would be worried about it. Nautilus are great sounding speakers but the design is mostly marketing BS overkill as most well designed drivers control and suppress the back wave. The only thing that dominates and contributes very heavily to the speaker sound is the acoustic suspension which operates at the desired tuned LF (to be complementary to the driver resonance) and is just like an elastic spring.
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
From the Swedish forum Faktiskt.io, there are some people stating that the sound from the inside of the box trough the membrane is at about -20 dB in a normal speaker. No measurements was shown.
If this is true, - I think -20 dB is a high number and should be clearly audible ?

Experiment anyone can perform:

Import a file into Audacity.

Copy and Paste a new copy into the same window.

Delay the copy by say, 3ms, as a simulation of the reflection from a relatively deep box.

Hit play, and mute the copy track at will. The doubling is easily heard.

Attenuate the copy by 20dB.

Play again, mute and un-mute the copy.

Now what do you hear?

I don't hear the copy any more.

You might, or might not.
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
Secondly, a (roughly) identical signal delayed by 2 milliseconds or so and attenuated by 20dB would potentially put about 2dB of ripple in the frequency response, so if those numbers are correct, it should be visible in existing speaker measurements.

Thus, at -20dB attenuation, one would expect a sound pressure fluctuation of +-0.8dB in the worst case.
According to Linkwitz, special attention should be paid to the 100-500Hz range:
SL: "I am of the opinion that the effect is most notable in the low hundreds of Hz region, where stuffing materials are ineffective and the internal dimensions not small enough for the internal air volume to act as a pure compliance."

Fittingly, I have two near-field measurements of CB speaker on the hard disk.
One is a 15'' driver in a 0.5m deep cabinet and the other is a 4'' driver in a 0.2m deep cabinet.
For each frequency the path length of the reflection remains the same:
- CB15: 1m (2x0.5m) path length for the CB speaker with 15'' driver
- CB4: 0.4m (2x0.4m) path length for the CB speaker with 4'' driver

The sound coming through the driver cone is "in-phase" with the main sound again (the reflecting back wall inverts the inverted sound of the back cone again), but delayed in time (phase shifted).
This means that at 2pi/360° phase shift the main and delayed signal add up optimally, at pi/180° the delayed signal reduces the main signal most.

Since the path length of the reflection never changes, we need to find the frequencies that satisfy the 2pi and pi conditions.
CB15: 2pi "maximum" 344/1 -> about 340Hz, pi "minimum" 170Hz
CB4: 2pi "maximum" 344/0.4 -> about 860Hz, pi "minimum" 430HzHz
If the damping material slightly reduces the speed of sound inside the enclosure, then max/min will be shifted towards low frequencies.

I have drawn the max and min frequencies into the frequency response measurements of the two CB speakers.
With a little imagination, you can see something like a tiny wave motion, but they do not quite match the theoretical max/min.
CB15: 1666119641690.png CB4: 1666119593676.png

Assuming the effect would be due to reflections inside the cabinet and the transmission through the driver cone, this would be negligible in contrast to the influence of the time-delayed sound on OB speaker.

For comparison, here is the influence of the time-delayed sound radiated from the rear of an OB speaker, normalized to the frequency response of the sound radiated directly from the woofer front:
1666120463881.png
The sound pressure level is increased by up to +3.4dB (see example from post#21) by the time-delayed sound of the rear panel.



Experiment anyone can perform:
...
You might, or might not.
A brilliant idea!
The audibility is even made a bit more difficult, since the effect only occurs over a certain frequency range, as air compliance to low frequencies and the damping material to high frequencies act like high and low pass filters.
 

fluid

Addicted to Fun and Learning
Joined
Apr 19, 2021
Messages
691
Likes
1,196
I am currently using 400gr/m2 polyester wad both applied on the walls + fully stuffing the cabinet
Wondering if replacing it with felt would be better or not?
That kind of fill looks like it is just changing the effective box volume. For that purpose it works. Felt on the walls or in front of the wadding stuck to the walls would likely give a different result. The best measurement comparison for this is an impedance response, all it takes is REW a known resistor value and some spare parts, if you don't have a separate interface, even the headphone output and line in on a laptop will work well enough to give you an idea.
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
That kind of fill looks like it is just changing the effective box volume. For that purpose it works. Felt on the walls or in front of the wadding stuck to the walls would likely give a different result. The best measurement comparison for this is an impedance response, all it takes is REW a known resistor value and some spare parts, if you don't have a separate interface, even the headphone output and line in on a laptop will work well enough to give you an idea.
Thanks @fluid
What do you think about using something like this: http://www.ctk.eu.com/assets/files/CTK DOMINATOR 370*500.pdf
I have used these in my rear truncated line arrays
 

fineMen

Major Contributor
Joined
Oct 31, 2021
Messages
1,504
Likes
679
One of the reasons ... coloring the sound?

Is it in the magnitude of -20 dB below the direct sound from the driver , or is it less ?
Can it be heard ?

Do You hear IT? What is Your comparison when identifying decremental influences? The original? It it was so, by how much is Your enjoyment of music--the very reason often forgotten, spoiled by such imaginably avoidable imperfections?

Yes, a KEF BD139 passive radiator leaks about -20..30dB of midrange. That's a good starting point. But a 'timely delayed' frequency component doesn't make any sense! Frequency does not include delay as a concept! You cannot delay a frequency, a physical state cannot transmit information at 'one frequency'.

What You might speak of legally is a signal, which carries information to somewhere in space and time. That signal has to, mathematically and hence logically, comprise a plethora of different frequencies to. literally, make sense.

In case You speak of sound components that reach the sancuary listener, for her sole pleasure of course, You have to consult the group delay. That is phase over frequency, basically. And suddenly the picure changes colors. That leaked sound components alter the group delay of the reproduction by a tiny bit. Very much like the subdued components that leak through a branch of a cross over.

I really hope that You and maybe others could eventually give up ideas, that didn't understand the problem to solve to begin with, and which neither are fully understood by the audience secondly. Why is it, that many only ingest the presumably bad news, but never report on idependently found reliefs?
 

fluid

Addicted to Fun and Learning
Joined
Apr 19, 2021
Messages
691
Likes
1,196
Thanks @fluid
What do you think about using something like this: http://www.ctk.eu.com/assets/files/CTK DOMINATOR 370*500.pdf
I have used these in my rear truncated line arrays
That sort of car sound deadening sheet can be quite good but also very variable. I glued 3mm textured rubber sheet onto the sides of my enclosure with a silicone type adhesive for a bit of flex., mainly because I had a roll sitting there taking up space :)

That isn't an alternative to felt though, more something to use as well. Search ebay for automotive felt or find a removal company as they often use it for moving blankets.
It is basically chopped up recycled fabric all smooshed back together. Usually pretty cheap and worth a try, nothing else really works quite the same.
 
OP
Tangband

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
Thus, at -20dB attenuation, one would expect a sound pressure fluctuation of +-0.8dB in the worst case.
According to Linkwitz, special attention should be paid to the 100-500Hz range:


Fittingly, I have two near-field measurements of CB speaker on the hard disk.
One is a 15'' driver in a 0.5m deep cabinet and the other is a 4'' driver in a 0.2m deep cabinet.
For each frequency the path length of the reflection remains the same:
- CB15: 1m (2x0.5m) path length for the CB speaker with 15'' driver
- CB4: 0.4m (2x0.4m) path length for the CB speaker with 4'' driver

The sound coming through the driver cone is "in-phase" with the main sound again (the reflecting back wall inverts the inverted sound of the back cone again), but delayed in time (phase shifted).
This means that at 2pi/360° phase shift the main and delayed signal add up optimally, at pi/180° the delayed signal reduces the main signal most.

Since the path length of the reflection never changes, we need to find the frequencies that satisfy the 2pi and pi conditions.
CB15: 2pi "maximum" 344/1 -> about 340Hz, pi "minimum" 170Hz
CB4: 2pi "maximum" 344/0.4 -> about 860Hz, pi "minimum" 430HzHz
If the damping material slightly reduces the speed of sound inside the enclosure, then max/min will be shifted towards low frequencies.

I have drawn the max and min frequencies into the frequency response measurements of the two CB speakers.
With a little imagination, you can see something like a tiny wave motion, but they do not quite match the theoretical max/min.
CB15: View attachment 237937 CB4: View attachment 237936

Assuming the effect would be due to reflections inside the cabinet and the transmission through the driver cone, this would be negligible in contrast to the influence of the time-delayed sound on OB speaker.

For comparison, here is the influence of the time-delayed sound radiated from the rear of an OB speaker, normalized to the frequency response of the sound radiated directly from the woofer front:
View attachment 237938
The sound pressure level is increased by up to +3.4dB (see example from post#21) by the time-delayed sound of the rear panel.




A brilliant idea!
The audibility is even made a bit more difficult, since the effect only occurs over a certain frequency range, as air compliance to low frequencies and the damping material to high frequencies act like high and low pass filters.
Thanks for this , ctrl:)
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
That sort of car sound deadening sheet can be quite good but also very variable. I glued 3mm textured rubber sheet onto the sides of my enclosure with a silicone type adhesive for a bit of flex., mainly because I had a roll sitting there taking up space :)

That isn't an alternative to felt though, more something to use as well. Search ebay for automotive felt or find a removal company as they often use it for moving blankets.
It is basically chopped up recycled fabric all smooshed back together. Usually pretty cheap and worth a try, nothing else really works quite the same.
Thank you @fluid as always :)
Last question, I promise: would felt add to Vb or deduct from Vb? Meaning will it increase the cabinet size or decrease it? I am asking since I know that wad increases it but not sure if felt behaves in the same fashion when it comes to that

In the meantime I had this idea: why not combine all these? Like maybe use a thinner plywood, let's say 15mm + apply 4mm of vibrodamping material (like the one I linked above) that would act like an equivalent of a wall thickness of 19mm, right? And then add felt to the walls (10 or even 16mm) and then fill the box with wad
 

fineMen

Major Contributor
Joined
Oct 31, 2021
Messages
1,504
Likes
679
Thank you @fluid as always :)
Last question, I promise: would felt add to Vb or deduct from Vb? Meaning will it increase the cabinet size or decrease it? I am asking since I know that wad increases it but not sure if felt behaves in the same fashion when it comes to that

In the meantime I had this idea: why not combine all these? Like maybe use a thinner plywood, let's say 15mm + apply 4mm of vibrodamping material (like the one I linked above) that would act like an equivalent of a wall thickness of 19mm, right? And then add felt to the walls (10 or even 16mm) and then fill the box with wad

Why do You ask; maybe it is because You cannot do the measurement part on Your own? If so, what are Your chances to conclude on an optimal solution in the near future? I assume it would be less laborsome and even quicker to first learn about measurement as a sequence of logical steps ( to put it briefly: analysis/modelling, generating model-specific expectations, discriminating observations, evaluation ), and just start over with actions targeting the posed problem. It is true, that such an attempt would include a lot of failed trials. Actually, thats the discriminating part: to, for sure, know what doesn't work often presents the most valuable part of the total endevour.

Summary: it doesn't make sense to ask for warranted recipes.
 
Top Bottom