pooge is correct. The term "toe-in" was not invented by speaker testers. In was borrowed from anatomical descriptions of things like "pigeon toed" or "in-toeing". Someone who is pigeon toed has a toed-in stance -- their feet are not aligned parallel to their body's fore-and-aft axis, but rather point inward toward each other. The same sense carried over into mechanical engineering to describe things like wheels that are not aligned perfectly parallel to each other and to a vehicle's straight-line axis of motion, but rather angled toward each other. A pair of speakers that is not toed-in or toed-out will be aligned normal to a line through the centers of the speakers -- and that line is frequently parallel to the front wall of the listening room.
Toe-in refers to speakers pointing at your feet
### Off axis angle refers to listening position ###
Zero is on axis (reference angle/speaker response)
Normal offaxis is "outside" the listening position
Toe-in offaxis is "inside" the listening position
Doesnt matter, its just words
Its what you do that matters
### Off axis angle refers to listening position ###
Zero is on axis (reference angle/speaker response)
Normal offaxis is "outside" the listening position
Toe-in offaxis is "inside" the listening position
Doesnt matter, its just words
Its what you do that matters
I could find no reference to the term Toe in speaker setup as coming from the claimed sources of pointing at toes or having anything to do with toes. My best guess is that the term was taken from it's use in alignment terminology, where I'm sure it existed long before speakers did, as a term for the alignment of two parallel lines such as the tires of a car. Toe-in referring to the front end of the tires being closer together than the rear. While the term toe in terms of automotive engineering may have started as a reference to how we point our toes, as I said, I couldn't find reliable evidence of that being the case. It makes sense, and thats about it. As far as what it means in Audio, I've heard 100's of people argue over the meaning, which seems silly. We should just pick one, agree with it for now, and the discussion can go from there. Dr. Geddes seems to prefer to discuss toe-in as any angle of the speakers greater than on-axis angle. If we were to think of it in terms of automotive alignment terms, it would mean any angle greater than parallel, to which the front of the cabinets are closer together than the rear. It seems that some have heard one, and some have heard the other. Again, if I were to venture a historical guess as to the origin of the term, it would be the latter mentioned defention, as it makes more intuitive sense (historically), but I'm fine with Dr. Geddes preferred definition, and it has created no confusion for me in setting up my speakers. However, I do wonder if his and Markus' point has been lost on some with regard to toe-in and speaker design. Markus and Dr. Geddes have mentioned before that Toe-in where the listener is listening on the outer off-axis angle of the speaker (aggressive toe-in?) is not sensible with normal speakers. Toe-in of speakers so that the listener is on-axis makes sense for any speaker which was designed to be flat on axis, as it won't be flat off-axis. This would mean that people arguing the car alignment definition, in which on-axis is actually some toe-in, as being wrong, are arguing that they prefer not to listen on-axis, at the angle which the speaker is likely the flattest. This would seem silly to me, unless I have misunderstood the argument.
Now as an alternative issue of consideration, I'm curious about different ways to control vertical directivity issues. I have no desire to get into an argument over the flatness of the response or the vertical nulls that may exist. I'm purely thinking in terms of the reflections off the ceiling and floor, and ways of minimizing these. Obviously one method is adding greater vertical control of the response, which we have some here, but that is only a partial solution I would think. Another is to treat the reflective surface in order to absorb or scatter the reflections. This is probably the best solution, especially scattering, given that complete absorption is not easy and not attractive. However, Dr. Geddes suggested setup, as well as reading the old JBL Everest papers gave me another thought.
What if we place the speakers slightly above the listening axis in height, but point the speakers down so that they are directed to the listener on their vertical axis? It would seem that doing this would increase the amount of sound being directed toward the floor, which has carpet and a carpet pad for some absorption, and less toward the ceiling, which can not be as easily treated. I know that Dr. Geddes had mentioned that it would require very thick carpet to eliminate all reflections, but it still seems a more aesthetically pleasing area to treat, and something that can be done more cheaply. In my own place, which I rent, I have fairly thick Burbur carpet with a thick foam rubber underpad. On top of that carpet, I have a large wool area rug which nearly doubles the carpet thickness. The only negative I have here is that under the pad is a cement slab. I think if building a dedicated room, a floating floor would make better sense, and if designed correctly, could actually serve as a very thick acoustic absorber as well.
So what do you think, does this make any sense? I currently have my speakers slightly too high just due to the speaker stand height (24") and the screen placement. I due have them pointed down at a slight angle, maybe 5 degrees or so.
What if we place the speakers slightly above the listening axis in height, but point the speakers down so that they are directed to the listener on their vertical axis? It would seem that doing this would increase the amount of sound being directed toward the floor, which has carpet and a carpet pad for some absorption, and less toward the ceiling, which can not be as easily treated. I know that Dr. Geddes had mentioned that it would require very thick carpet to eliminate all reflections, but it still seems a more aesthetically pleasing area to treat, and something that can be done more cheaply. In my own place, which I rent, I have fairly thick Burbur carpet with a thick foam rubber underpad. On top of that carpet, I have a large wool area rug which nearly doubles the carpet thickness. The only negative I have here is that under the pad is a cement slab. I think if building a dedicated room, a floating floor would make better sense, and if designed correctly, could actually serve as a very thick acoustic absorber as well.
So what do you think, does this make any sense? I currently have my speakers slightly too high just due to the speaker stand height (24") and the screen placement. I due have them pointed down at a slight angle, maybe 5 degrees or so.
Matt
You understand perfectly and you see it the same way that I do. But I hate to go on and on in these forums about simple stuff like that. As I said, I don't care what the deffinition is as long as we all understand it and stick to it. I can see the arguments both ways and so its just a mater of picking one and going with it. I offered up one, but people objected. I am not strongly committed either way.
You understand perfectly and you see it the same way that I do. But I hate to go on and on in these forums about simple stuff like that. As I said, I don't care what the deffinition is as long as we all understand it and stick to it. I can see the arguments both ways and so its just a mater of picking one and going with it. I offered up one, but people objected. I am not strongly committed either way.
pjpoes said:
Matt, the answer is highly dependnet on the specific situation. In my room the ceiling reflection is handled better than the floor, so your proposal doesn't work for me. But clearly is you can't do anything about the ceiling then it makes sense and is worth a try.
Someday I'll finish the elliptical waveguide and give that a try, but I just don't seem to be able to get the time. I remain unconvinced about its effectiveness since, as the discussion with Tom Danley has pointed out, its just not as simple as some would like it to be. The devil is in the details.
Understood, and not to be overly redundant, but when you gave your setup preference, you made it clear what it meant, and so I udnerstood. When others have done the same, I just adjust my thinking.
The bigger point seems to be listening at the listening axis which gives the flattest response. If the speakers were parallel, pointing straight ahead, then the listener would be at the inside off-axis point. This places the listener 30 degree's off axis per each speaker, if I'm not mistaken, using the equilateral triangle setup. This is quite a large amount to be off-axis for a speaker not designed with controlled directivity. Most speakers are not only not flat at that position, they have nulls around the crossover point beginning to appear. While not universally true, it's more often the case than not in my experience.
For normal speakers which develop these nulls as well as having a response that get's pretty nasty off-axis, listening with the speakers toed so that the listener is on-axis makes the most sense.
Take this image of the PSB response. By 30 degree's the speaker's response looks terrible, with a large roll off toward 10khz, followed by a hump, and some ragidness. Not something anyone would accept as the measured response of the speaker. That speaker needs to be listened to on-axis.
The CD speakers such as the Gedlee designs are smoothest off-axis. So much so that you can get away with even more aggressive toe, and have a response which not only is smoother, but avoids early reflections off the sidewall. Something, according to what I've been told by Dr. Geddes and since read up on my own, was actually a key design consideration in the original JBL Everest. Pretty cool in my opinion, and too bad it's been so lost to time.
The bigger point seems to be listening at the listening axis which gives the flattest response. If the speakers were parallel, pointing straight ahead, then the listener would be at the inside off-axis point. This places the listener 30 degree's off axis per each speaker, if I'm not mistaken, using the equilateral triangle setup. This is quite a large amount to be off-axis for a speaker not designed with controlled directivity. Most speakers are not only not flat at that position, they have nulls around the crossover point beginning to appear. While not universally true, it's more often the case than not in my experience.
For normal speakers which develop these nulls as well as having a response that get's pretty nasty off-axis, listening with the speakers toed so that the listener is on-axis makes the most sense.
Take this image of the PSB response. By 30 degree's the speaker's response looks terrible, with a large roll off toward 10khz, followed by a hump, and some ragidness. Not something anyone would accept as the measured response of the speaker. That speaker needs to be listened to on-axis.
The CD speakers such as the Gedlee designs are smoothest off-axis. So much so that you can get away with even more aggressive toe, and have a response which not only is smoother, but avoids early reflections off the sidewall. Something, according to what I've been told by Dr. Geddes and since read up on my own, was actually a key design consideration in the original JBL Everest. Pretty cool in my opinion, and too bad it's been so lost to time.
gedlee said:
I think Tom's arguement makes plenty of sense, and was always how I looked at it. For whatever reason, some seem to look at it as a sensible fix with no trade-offs, or at least, the trade-offs are smaller than those of a symmetric waveguide. I don't see this as being the case, and without having either your or Tom's knowledge or experience, I saw this problem.
The other issue with the arguement, center to center spacing, doesn't make sense to me either, while I understand it's importance in crossover design, it was my understanding this distance was a function of the radiator size and frequency at the crossover point. The wavelength at your crossover point is more than large enough to allow a center spacing of that size. It's acoustically a small distance with regard to that frequency range, and by a relatively large margin. I did the math myself when I first got the speakers, as I was curious, and have never seen it as a problem. None the less, that's a different argument all together.
I've also wondered about using a foam with variable density in which the foam was densest toward the top and bottom of the waveguide. This would not alter directivity in the same way, if I understand it's function correctly, but would have a similar effect in that it would reduce output in the vertical axis extremes to a greater degree. I would think the foam plugs would have to be custom cast to these specs and I would imagine a linear change in density would be best. Probably way too expensive for the small benefit, and I'm guessing you still may have the same problem that an elliptical waveguide has.
A cheaper and less attractive solution could be foam of higher density than the foam plug placed around the speakers, with a lip that extends around the edges to a greater degree on the top, and runs through the center of the speaker as well. Again, to look and work right, probably very very expensive, and without measurements, I'm not sure how effective it would be. This would also have the benefit of absorbing sound diffracting around the edges though, and in an installed theater where the speakers are hidden, it's probably less of an issue. In fact, I bet this could be built into the false wall usually placed behind the acoustically transparent screen seen in some installed theaters of this sort.
This is the response of the Escelante speakers. A not overly well designed speaker in my opinion, with pretty serious physical and electrical design problems. None the less, these types of problems aren't so uncommon in poorly executed DIY designs, nor are they unheard of in high end audio either. I've never seen anything this bad in a speaker from a big name like Polk, Infinity, PSB, etc., but it is seen in some high end esoteric designs. This is an extreme case of a speaker that should not be listened to 0ff-axis. Of course, this speaker shouldn't be listened to probably at all.
pjpoes said:
The definitiion of "toe-in" has NOTHING to do with personal preference. It is simply a reference to the geometrical relationship between a pair of speakers. Nothing more. A parallel pair is not toed in or out. Parallel is simply the geometrical reference zero for the terms, and doesn't rely on yet another variable of the listening position.
I think we have all seen people place loudspeakers in the faced-forward position. No one has stated this was a preferred position.
Toe-in was historically used in telling people how to stop using the faced-forward position of speakers to eliminate the ping-pong effect between the loudspeakers with lateral movement of the listening position. Whether or not the loudspeakers are flat off- axis, toeing the loudspeakers in, to the degree that they cross in front of the listening position, greatly helps to stabilize the central image with lateral movement of the listening position via time-intensity tradeoffs. Whether or not a preferred frequency response results therefrom is a totally separate issue. The stabilization of the central image is still better with most any speaker pair. It works whether or not the speakers are CD.
Like anything else, experimentation may be in order to get a preferred balance between the image stabilization and the timbral response.
So, you want toe-in to use the wall behind the speakers as reference?
But we still have on-axis/off-axis referring to listening position
Hmm, complicated
Lets see
If 0gr on-axis listening is say 60gr toe-in
To define 10gr off-axis
then 10gr off-axis is either 50gr toe-in, or 70gr toe-in
You need to know a fixed listening position to get the numbers right
Wall behind speakers is reference point
Those are strange numbers I have never seen used before
Or more easily, you could just name it 10gr off-axis or 10gr toe-in
This does not depend so much on listening position
Listening position is reference point
no more to say about it 
But we still have on-axis/off-axis referring to listening position
Hmm, complicated
Lets see
If 0gr on-axis listening is say 60gr toe-in
To define 10gr off-axis
then 10gr off-axis is either 50gr toe-in, or 70gr toe-in
You need to know a fixed listening position to get the numbers right
Wall behind speakers is reference point
Those are strange numbers I have never seen used before
Or more easily, you could just name it 10gr off-axis or 10gr toe-in
This does not depend so much on listening position
Listening position is reference point
no more to say about it pjpoes said:
That description of using foam reminds me of the m&k speakers
http://www.stereophile.com/standloudspeakers/137/
Yes, a lot of companies, like M&K, have used foam or felt to effect directivity of the speakers or reduce diffraction. It's nothing new in and of itself. M&K used it, I believe, more to reduce interference between the multiple tweeters, rather than to improve vertical directivity characteristics. I think a larger amount of foam is necessary to function in this capacity and has to be designed into the enclosure somewhat. The amount of reduction is proportional to the amount of time the sound can travel through the foam, so if you are trying to reduce the amount of sound produced within a certain vertical range, with greater reduction of sound at grater angles, then you need to have the foam shaped accordingly. On top of that, little 1" pieces just won't do much down to very low frequencies, and could reflective at very high frequencies.
I do think a better solution would be a sort of Foam condom for the speakers with variable foam density and specific shaping to best control directivity where we want the foam controlling it. In other words, not extending past the edge on the sides, but sticking out over the top part to control vertical directivity. This is no small feat I don't think. I would imagine this quantity of foam, not even considering the cutting part, could be a few hundred dollars per speaker. Figure even more for variable density. Add another 100 per speaker easy for custom cutting to shape, especially if there was any design to the shape. It has a lot of advantages that a waveguide alone cant deal with, like reducing edge diffraction further than the roundover can. However, I would think that controlling vertical directivity with the waveguide, foam insert, and aiming is the most simple and least expensive solution.
I do think a better solution would be a sort of Foam condom for the speakers with variable foam density and specific shaping to best control directivity where we want the foam controlling it. In other words, not extending past the edge on the sides, but sticking out over the top part to control vertical directivity. This is no small feat I don't think. I would imagine this quantity of foam, not even considering the cutting part, could be a few hundred dollars per speaker. Figure even more for variable density. Add another 100 per speaker easy for custom cutting to shape, especially if there was any design to the shape. It has a lot of advantages that a waveguide alone cant deal with, like reducing edge diffraction further than the roundover can. However, I would think that controlling vertical directivity with the waveguide, foam insert, and aiming is the most simple and least expensive solution.
Matt
I started with the foam to act as an aid in directivity, but it didn't work as well as I had hoped. The Index of Refraction is just not great enough until there is a lot of absorption. However, I did note how much better the speakers sounded even though the directivity is not that highly affected. Thats when I started to try and figure out what it was that was making such a positive result sound wise when it wasn't working directivity wise. I'm still not 100% convinced that I understand it.
I started with the foam to act as an aid in directivity, but it didn't work as well as I had hoped. The Index of Refraction is just not great enough until there is a lot of absorption. However, I did note how much better the speakers sounded even though the directivity is not that highly affected. Thats when I started to try and figure out what it was that was making such a positive result sound wise when it wasn't working directivity wise. I'm still not 100% convinced that I understand it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.