Title pretty self-explanatory 
... spin-off question: given a far-enough listening distance, is the comb-filtering effect will stay audible ?
... spin-off question: given a far-enough listening distance, is the comb-filtering effect will stay audible ?
If I may piggyback. For two Mid or Full range drivers would angling them toward each other (a Vee arrangement) help reduce the effects (centers closer together) or would it just exacerbate the problem?
With 2 like cone drivers mounted side-by-side, you'll get noticeable comb filtering up close & beaming at any distance. Far enough away, any loudspeaker arrangement will sound like a point source, but you'll still get beaming. Standing the enclosure up on end, you'll get better horizontal coverage while moving the comb filtering into the vertical, where the ears are less sensitive to it.
If you don't have the option for vertical mounting, then cross-firing the drivers is the next best option (the Vee arrangement). Try angles from 90 degrees to 120 degrees max.
If you don't have the option for vertical mounting, then cross-firing the drivers is the next best option (the Vee arrangement). Try angles from 90 degrees to 120 degrees max.
Last edited:
An externally hosted image should be here but it was not working when we last tested it.
If i understand correctly, the audible effect of comb filtering will be only if the listener moves his head (or in the unlikely event that the speaker moves itself) ?
And that will translate into a change in SPL is some frequencies (deep dips) that might be more or less audible (if the ''comb'' is coarser than 1/3-1/6 octave and/or if more than 1db) ?
A loudspeaker's off-axis response has much to do with how it "sounds". Comb filtering is a phenomenon where, above some frequency, interference creates nulls in the off-axis response from the loudspeaker. So comb filtering can and will change how the loudspeaker sounds. This is not easy to quantify because it has to do with how the room interacts with the loudspeaker.
If you and the loudspeaker could magically be suspended in the middle of empty space, far away from any boundary, you would only hear the direct sound that travels in a straight line from the loudspeaker to your ears. In a room the sound radiates outward from the loudspeaker and interacts with walls, floor, ceiling, furniture, etc. before reflecting back into the room and interacting with the direct sound. A loudspeaker that has one or more nulls off-axis will sound different than another that has a different distribution of nulls or has none at all because the sound that is "returned" to the listener by the room will change. The power of room reflected sound can be as high as -6dB below the power of the direct sound so it can have a large influence on the overall character that you "hear". When the boundaries are far away, e.g. in a "large" room, the effect will be lessened if you are in the normal listening location (e.g. 6-10 feet away from the loudspeakers).
As you move far away from the loudspeaker, the direct and room contributed sound levels will approach each other so nulls will have a greater effect at larger distances on the overall sound character of the loudspeaker in a room. If you listen very close to the loudspeaker, the level of the direct sound will be significantly more than the room reflected sound, and nulls or comb filtering will have a reduced influence on the sound character.
If you and the loudspeaker could magically be suspended in the middle of empty space, far away from any boundary, you would only hear the direct sound that travels in a straight line from the loudspeaker to your ears. In a room the sound radiates outward from the loudspeaker and interacts with walls, floor, ceiling, furniture, etc. before reflecting back into the room and interacting with the direct sound. A loudspeaker that has one or more nulls off-axis will sound different than another that has a different distribution of nulls or has none at all because the sound that is "returned" to the listener by the room will change. The power of room reflected sound can be as high as -6dB below the power of the direct sound so it can have a large influence on the overall character that you "hear". When the boundaries are far away, e.g. in a "large" room, the effect will be lessened if you are in the normal listening location (e.g. 6-10 feet away from the loudspeakers).
As you move far away from the loudspeaker, the direct and room contributed sound levels will approach each other so nulls will have a greater effect at larger distances on the overall sound character of the loudspeaker in a room. If you listen very close to the loudspeaker, the level of the direct sound will be significantly more than the room reflected sound, and nulls or comb filtering will have a reduced influence on the sound character.
https://en.wikipedia.org/wiki/Comb_filter#Applications
In any enclosed space, listeners hear a mixture of direct sound and reflected sound. Because the reflected sound takes a longer path, it constitutes a delayed version of the direct sound and a comb filter is created where the two combine at the listener
Also:
Why be so overly concerned about comb filtering in column speakers when you probably get this all the time with the speakers y
After hearing an initial signal, the brain will suppress any later signal, such as an echo, for a time up to about 30 or 40 milliseconds. This inhibition is called time or temporal masking. In effect, you do not hear the sound from the higher or lower drivers that would otherwise interfere with locating the nearest source of the sound.
If the arrival time of any echo is longer than this, then two distinct sounds are heard, even if the second arrival is as much as10 dB higher. On the other hand, if there is no arrival time difference or only a very small time difference, the source will be heard to be coming from straight ahead. Even in a room having several reflections, we can reliably localize the real source.
In any enclosed space, listeners hear a mixture of direct sound and reflected sound. Because the reflected sound takes a longer path, it constitutes a delayed version of the direct sound and a comb filter is created where the two combine at the listener
Also:
Why be so overly concerned about comb filtering in column speakers when you probably get this all the time with the speakers y
After hearing an initial signal, the brain will suppress any later signal, such as an echo, for a time up to about 30 or 40 milliseconds. This inhibition is called time or temporal masking. In effect, you do not hear the sound from the higher or lower drivers that would otherwise interfere with locating the nearest source of the sound.
If the arrival time of any echo is longer than this, then two distinct sounds are heard, even if the second arrival is as much as10 dB higher. On the other hand, if there is no arrival time difference or only a very small time difference, the source will be heard to be coming from straight ahead. Even in a room having several reflections, we can reliably localize the real source.
As far as I recall, the brain does not "suppress" the sounds up until 40 or so milliseconds but rather it combines these sounds from this time window together. This is where the perception of the "character" of a sound and your immediate environment is generated.
While it is true that the room reflections are delayed and filtered copies of the direct sound radiated in many directions, the brain has adapted to use the clues from it to determine the true character of the source. Nulls in the radiation pattern will influence this perception.
Last edited:
So basically, i shouldn't even think about making a FAST with an array of 3fe22, is that right ? 😀
I built an OB line array using 16 of them and it works very well for me. Do not feel I have any issues at listening distance with comb filtering at all.
EV3
In a textbook diagram, comb filtering looks terrible.
But as per CharlieLaub, if you give a moments thought to non-point sources, how a stereo cello recording comes out of two non-point-source speakers and their accidental phases, where your two ears are and how stationary they are, and just how you could "hear" comb filtering (it isn't like hearing the presence of 2nd harmonic distortion or the absence of sound band 4kHz to 5kHz).... i
For comb filtering to occur, the phases need to connect precisely and stably from the person playing the cello to your ears, otherwise the "filtering" takes place all over the physical location and frequency maps and is constantly shifting about. Which is to say it isn't perceivable. So it is really inapplicable theoretically and irrelevant practically.
Stereo speakers do interact on pure tones when you do frequency traces with a stationary point-source mic, as anybody who has run as many frequency curves as I have can attest. It's a stretch, but I suppose you could call that "filtering". But that isn't the Halloween comb filtering defect.
B.
But as per CharlieLaub, if you give a moments thought to non-point sources, how a stereo cello recording comes out of two non-point-source speakers and their accidental phases, where your two ears are and how stationary they are, and just how you could "hear" comb filtering (it isn't like hearing the presence of 2nd harmonic distortion or the absence of sound band 4kHz to 5kHz).... i
For comb filtering to occur, the phases need to connect precisely and stably from the person playing the cello to your ears, otherwise the "filtering" takes place all over the physical location and frequency maps and is constantly shifting about. Which is to say it isn't perceivable. So it is really inapplicable theoretically and irrelevant practically.
Stereo speakers do interact on pure tones when you do frequency traces with a stationary point-source mic, as anybody who has run as many frequency curves as I have can attest. It's a stretch, but I suppose you could call that "filtering". But that isn't the Halloween comb filtering defect.
B.
Last edited:
Thanks to all for the pictures and explanations, it's very interesting.
Will surely start to make some test of my own, with different c-c distances and different drivers.
Will surely start to make some test of my own, with different c-c distances and different drivers.
What are you expecting to hear? Or not hear? What does comb filtering sound like when a singer is singing - the horrific comb-like picture in the textbook?
B.
Last edited:
I have not much expectations since i have limited experience with the multiple drivers concept (line array and such) so far.
As far as i know, i'd expect a change in SPL when the listener's position changes (ex: seated v.s. standing), also frequency-dependant... so basically pretty much chaotic for the ears, theoratically. In practice, if the changes are only 0.5-1.5db its not that bad.
Last edited:
My Comb filtering experiment
Even though most common commercial speakers have traditional tweeter/woofer combinations paying no heed to this reputed issue, I wanted to try it on for size, using parts I already have at hand.
I did wind up buying a pair of 2 inch Seas fabric domes, for the sake of closer spacing.
This uses a mix of active 4th order Linkwitz Riley active, and passive 1st order Solen Split passive filters.
I made my best guesses for power tapering and filter knees, since I don't use a computer to model designs for me.
So far, the tweeters are two pair of 3" metal domes bandpassed actively at 600 hertz 4th order active, and 2400 first order passive Solen Split. The Seas is highpassed Solen Split at 6000 hertz.
Individual drivers are similar in efficiency by the paperwork, and online calculators were used and them adjusted with test tones and well known to me recordings for reference. The online passive components were far off in fact values from predicted values.
The tweeters as a whole are actively crossed 4th order Linkwitz Riley at 600 hertz to a pair of woofers. In keeping with the theme of comb filtering determining driver spacing and crossover points, these are Solen Split passively, only one woofer having an inductor on it to lowpass its output at about 115hz Butterworth first order. This does double duty besides addressing comb filtering- it gives a low end kick up, tuning these drivers to the simple U Frame they're in.
This reduces a midrange hump greatly, and makes things sound much more natural than high passing the woofer above, or running both without passive filtering of any kind in series or parallel.
I've settled on the woofers in parallel, one of each pair with an inductor on it.
It's not a revelation in night and day sound, but in a single speaker system for all I listen to, I've noticed good recordings mostly sound better, poor recordings sound worse, and in my non remote control 2 channel system, movie dialogue does not have me straining to follow it, or have to turn the playback level down in action sequences.
Living in an apartment building, as an aside that may have nothing to do with.the speakers, I do notice street noises now during music, as well as the sounds of people next door moving about or their muffled voices.
Even the postie delivering mail through the door slots of neighbours.
To wrap it up, these don't sound all that similar to other speakers I've been building over the last 30 years or so, but I'm leaning towards liking them best.
Even sound in my room, little to no floor or ceiling reflection intrusion, wide wide sweet spot, not quite the depth of stage as the best full rangers I've heard, but very close and consistent all across the room.
A bit of a pain to build since I got hurt in October, but well worth it.
One inexpensive stereo amp per L/R side, and my old crossover actively rolls off bass at 32hz second order to help protect my woofers from overexcursion.
Even though most common commercial speakers have traditional tweeter/woofer combinations paying no heed to this reputed issue, I wanted to try it on for size, using parts I already have at hand.
I did wind up buying a pair of 2 inch Seas fabric domes, for the sake of closer spacing.
This uses a mix of active 4th order Linkwitz Riley active, and passive 1st order Solen Split passive filters.
I made my best guesses for power tapering and filter knees, since I don't use a computer to model designs for me.
So far, the tweeters are two pair of 3" metal domes bandpassed actively at 600 hertz 4th order active, and 2400 first order passive Solen Split. The Seas is highpassed Solen Split at 6000 hertz.
Individual drivers are similar in efficiency by the paperwork, and online calculators were used and them adjusted with test tones and well known to me recordings for reference. The online passive components were far off in fact values from predicted values.
The tweeters as a whole are actively crossed 4th order Linkwitz Riley at 600 hertz to a pair of woofers. In keeping with the theme of comb filtering determining driver spacing and crossover points, these are Solen Split passively, only one woofer having an inductor on it to lowpass its output at about 115hz Butterworth first order. This does double duty besides addressing comb filtering- it gives a low end kick up, tuning these drivers to the simple U Frame they're in.
This reduces a midrange hump greatly, and makes things sound much more natural than high passing the woofer above, or running both without passive filtering of any kind in series or parallel.
I've settled on the woofers in parallel, one of each pair with an inductor on it.
It's not a revelation in night and day sound, but in a single speaker system for all I listen to, I've noticed good recordings mostly sound better, poor recordings sound worse, and in my non remote control 2 channel system, movie dialogue does not have me straining to follow it, or have to turn the playback level down in action sequences.
Living in an apartment building, as an aside that may have nothing to do with.the speakers, I do notice street noises now during music, as well as the sounds of people next door moving about or their muffled voices.
Even the postie delivering mail through the door slots of neighbours.
To wrap it up, these don't sound all that similar to other speakers I've been building over the last 30 years or so, but I'm leaning towards liking them best.
Even sound in my room, little to no floor or ceiling reflection intrusion, wide wide sweet spot, not quite the depth of stage as the best full rangers I've heard, but very close and consistent all across the room.
A bit of a pain to build since I got hurt in October, but well worth it.
One inexpensive stereo amp per L/R side, and my old crossover actively rolls off bass at 32hz second order to help protect my woofers from overexcursion.
Attachments
Just a generalist weighing in: In any real world listening room (e.g. not an anechoic chamber), comb filtering is inevitable. The human ear adjusts remarkably well to it. Listening to stereo? Move your head a few inches to one side or the other. It's almost certain you've substantially changed the "response" at each ear.
Sometimes, this is used as a special effect in pop music, accidental or deliberate. The best example I can think of is on The Moody Blues album "In Search of the Lost Chord." There opening song, "Departure," has the usual mystical poetry with a background, a slowly ascending sine wave. When it reaches into the midrange, you'll get a curious impression (at moderate volume): the tone swings from one ear to the other and back, as the pitch rises.
I suspect this works equally well even if listened to via a single speaker; having the two or four of a normal stereo speaker pair would only make it more complex.
Sometimes, this is used as a special effect in pop music, accidental or deliberate. The best example I can think of is on The Moody Blues album "In Search of the Lost Chord." There opening song, "Departure," has the usual mystical poetry with a background, a slowly ascending sine wave. When it reaches into the midrange, you'll get a curious impression (at moderate volume): the tone swings from one ear to the other and back, as the pitch rises.
I suspect this works equally well even if listened to via a single speaker; having the two or four of a normal stereo speaker pair would only make it more complex.
Last edited:
I have a big advantage over you. I can hear a pair of speakers that starts out by nearly eliminating comb filtering at its source, before room reflections compound what does get produced by playback in the speaker.
Nice dissertation dissing the results of what I built, minimizing comb filtering from driver spacing, crossover points, and floor and ceiling reflections though.
If I didn't mention it, or if you missed it, it's a wide listening window, in my room I can walk side to side about 10 feet without changing the tone.
Not the few inches you mention.
Nice dissertation dissing the results of what I built, minimizing comb filtering from driver spacing, crossover points, and floor and ceiling reflections though.
If I didn't mention it, or if you missed it, it's a wide listening window, in my room I can walk side to side about 10 feet without changing the tone.
Not the few inches you mention.