If there is a discontinuity between the waveguide and the wall, there will be diffraction moving through it and a reflection from the wall. Nothing can perfectly fix the transition (at least the reflection part) other than moving away from the wall. However...
While the reflection for the same waveguide out in the room will come later, the mouth transition will need to diffract more. Furthermore it is more practical to create a reasonable transition to a nearby wall than in free space.
What I don't like about the idea of a free space to room transition is it may affect DI, change the reflective conditions in that band, and perhaps make GD worse.
While the reflection for the same waveguide out in the room will come later, the mouth transition will need to diffract more. Furthermore it is more practical to create a reasonable transition to a nearby wall than in free space.
What I don't like about the idea of a free space to room transition is it may affect DI, change the reflective conditions in that band, and perhaps make GD worse.
Yea, about 50% of the time. Just what one would expect from random. hardly a stable means of judgment.
Both have reflections as I suggested. With directivity and proper setup the speaker in the room can have a reflection that is sufficiently delayed to be less of a problem. For a given amplitude the sooner the reflection is the more of a problem it is (unless it is > 10 ms then it is an asset.)
PS You sim lacks directivity and it will be criticized for that, but it makes a good point.
I was intrigued by Wayne's comments on the corner horn the last time the subject came up and ran some room simulations.
Note: These are only dirac impulse simulations in a bare room. But it's interesting to see how the reflections might play out. A speaker out in the room and a speaker in the corner. The green reflections are about 10-12dB down from the direct wave.
Comments?
The simulation from corner placement looks nice, doesn't it? When the source lies on the boundary, there essentially is no reflection, because it is coherent with the source. I did some animations like that too, done the source at various distances from the corners:
Did you see the links in my last post? They are AES papers on speaker placement and boundaries.
What I've learned in over three decades of working with constant directivity cornerhorns is that, above the Schroeder frequency, low frequencies radiate smoothly, almost like a pure ground plane. They are free from self-interference, it's just not there. And it's really easy to measure the improvement - Measure with the cornerhorn in the corner, and there are no self-interference notches. Pull it out away from the wall a couple feet and deep 10-15dB notches form in the lower midrange, around 100-150Hz, depending on the distances. This is the easy part, because in acoustic scale, a few inches just don't matter. You get the benefit just by being close.
Of course the closer the source is to the walls, the higher frequency this holds true. Only a flush mounted source is able to maintain this benefit all the way through the audio band. If there is a gap, then at some point, depending on the distance to the walls, it no longer acts this way. As frequency rises, the walls become increasingly further in scale. So by the top-octave, the walls aren't your friend anymore. Fortunately, the 45° off-axis output of a ~90° tweeter horn is down about 6dB, and so the grazing reflection is reduced. But it is still there. However it is really easy to deal with that using just about anything for an absorber. Even curtains work great. The sound traveling along a grazing path has to go through a lot of curtain material, so attenuation is very effective.
The only real problem as I see it is many rooms don't have suitable corners because of spacing or entryways or some other such thing. But it is pretty cool when you can take advantage of this configuration. It's just like soffit mounted speakers in recording studios, but even better, because not just the front wall reflection is removed, but also the ipsilateral wall. So if you have a room with the right corners, don't hesitate to try this configuration. I think it's really special.
I think we may have to quantify what constitutes an early reflection, and contrast that with coherent summing. I mean, a reflection that has zero delay might be called an "early reflection", but is it really? Now how 'bout one with 1uS delay? That's super early. Is that a problem? Of course not. So when does it become a problem? Would it be where delay is 100uS? Is it 1mS? Or maybe it is better to talk in terms of wavelength. Would the problem start at one wavelength? Quarter wavelength? A tenth wavelength?
Even in a DI-matched two-way (or any other loudspeaker with a crossover), you have a delay between the woofer and tweeter. That delay changes with respect to frequency. In the best designs, the two are really close to being in-phase in front of the speaker, at some point in the crossover overlap region. But even in this case, there is some shift through the overlap band.
I've seen guys trying to get phase-perfect crossovers, yet even the most aggresively focused on this aspect wouldn't be able to keep their phase shifts anywhere close to the coherency of a source and it's reflection off its own baffle. So I think if you're going to talk about early reflections, you have to concede that a reflection that is perfectly in-phase is not the same thing as one that is a few milliseconds delayed.
Agreed. A lot more controls are needed for good judgment. We don't always have them, so rely on what we can or can't hear.
I'd agree that it can, where 'almost' means that if I can hear it, I don't yet know about it.
I also like to think of surfaces as potential ground planes. Antenna construction seems to have limited appeal.
Definitely. One thing that separates a purpose-designed constant directivity cornerhorn from other speakers is it is designed to get the woofer physically close to the apex of the corner.
A constant directivity cornerhorn must also provide a seamless transition between bass bin (where walls are acoustically close) to the subsystems above it, which are growing increasingly acoustically distant. This usually means a large format midhorn blended with the bass bin, and snuggled back into the corner far enough it is acoustically close.
As frequency rises, the distance to the walls becomes less and less acoustically close so by the highest frequencies, the walls are acoustically distant. So the tweeter should be capable of directional control that limits the beamwidth, and it is nice to also have curtains or some other absorbent material to limit grazing reflections at HF.
A constant directivity cornerhorn must also provide a seamless transition between bass bin (where walls are acoustically close) to the subsystems above it, which are growing increasingly acoustically distant. This usually means a large format midhorn blended with the bass bin, and snuggled back into the corner far enough it is acoustically close.
As frequency rises, the distance to the walls becomes less and less acoustically close so by the highest frequencies, the walls are acoustically distant. So the tweeter should be capable of directional control that limits the beamwidth, and it is nice to also have curtains or some other absorbent material to limit grazing reflections at HF.
I does have some, tho I don't know how much with the Dirac pulse. It certainly shows some directivity at the start of the pulse. And dipoles show a pattern too. What CARA can't do is horns.
But of course we would hope that we aren't listening in a room with bare walls, and that things in the room would help break up the reflections a lot more than shown.
I just think that the lack of controls and the potential for unreliable results should lead one in the direction of a reliance more on objective data than subjective.
All too often people feel that subjective results are infallible and data is meaningless, usually because it does not support their subjective opinions. This is clearly the road to disaster where purely subjective evaluations are used to guide decisions. This is not going to lead to an improvement in the state-of-the-art.
I just re-read Sean Olive's Audio Musings, and "liking" has been studied quite a lot by Harman Ltd among others, but not with "audiophile" listeners (too small market??)
Please everybody, study his blog, lots of info and science about sound perception. Since this quoted article (year 2009) even Harman and Stereophile's measurement crew have understood the importance of controlled dispersion. Finnish and Canadian designers and manufacturers are pioneers of adopting this to home speakers starting in the 1980's.
(quote)
The panelists included myself, Steve Temme (Listen Inc.), Dr. Earl Geddes (GedLee), Laurie Fincham (THX), Mike Klasco (Menlo Scientific), and Dr. Floyd Toole (former VP Acoustic Engineering at Harman), who served as the panel moderator. After about 30 minutes, a consensus was quickly reached on the following points:
Please everybody, study his blog, lots of info and science about sound perception. Since this quoted article (year 2009) even Harman and Stereophile's measurement crew have understood the importance of controlled dispersion. Finnish and Canadian designers and manufacturers are pioneers of adopting this to home speakers starting in the 1980's.
(quote)
The panelists included myself, Steve Temme (Listen Inc.), Dr. Earl Geddes (GedLee), Laurie Fincham (THX), Mike Klasco (Menlo Scientific), and Dr. Floyd Toole (former VP Acoustic Engineering at Harman), who served as the panel moderator. After about 30 minutes, a consensus was quickly reached on the following points:
- The perception of loudspeaker sound quality is dominated by linear distortions, which can be accurately quantified and predicted using a set of comprehensive anechoic frequency response measurements (see my previous posting here)
- Both trained and untrained listeners tend to prefer the most accurate loudspeakers when measured under controlled double-blind listening conditions (see this article here).
- The relationship between perception and measurement of nonlinear distortions is less well understood and needs further research. Popular specifications like Total Harmonic Distortion (THD) and Intermodulation Distortion (IM) do not accurately reflect the distortion’s audibility and effect on the perceived sound quality of the loudspeaker.
- Current industry loudspeaker specifications are woefully inadequate in characterizing the sound quality of the loudspeaker. The commonly quoted “20 Hz - 20 kHz , +- 3 dB” single-curve specification is a good example. Floyd Toole made the observation that there is more useful performance information on the side of a tire (see tire below) compared to what’s currently found on most loudspeaker spec sheets (see Floyd's new book "Sound Reproduction").
Last edited:
Thats all nice as far as it goes. It certainly doesn't answer the question inherent in this thread.
If linear distortions (frequency response) are important then which ones and what weighting? If your speaker has perfect axial response and poor off axis response, while mine is so-so on axis but pretty good off axis, which is better?
This has been my primary complaint with a lot of the articles of Toole. He frequently comes to a conclusion that if a speaker measures well in all regards then it will be preferred in well controlled blind tests. This is hardly an earth shattering conclusion.
What I want to know (as a lazy speaker designer) is which parameters can I totally ignore and still have a well regarded speaker in the end. Which parameters are unimportant in subjective impression, or are there at least thresholds of "good enough" for some parameters?
This is whats wrong with these kind of discussions. Everyone comes in with an "everything is important" "optimize everything" view and we can all conclude that constant directivity or smooth directivity is a great thing. ("It can't hurt.")
Well controlled subjective tests are another matter.
Much more interesting are Olive's study where he came up with the weighting factors needed to create a rank ordering of a group of speakers that matched the blind listening ranking. He was able to do so with a high degree of correlation. (See Floyd's book.) The trouble is that he did the study twice with two different groups and came up with different weightings for different factors.
I still maintain that if you dig down through Floyd's earlier works, especially the early 2 part paper, you will conclude that power response (and therefore directivity index) has a poor or even negative correlation with subjective ranking.
David S.
If linear distortions (frequency response) are important then which ones and what weighting? If your speaker has perfect axial response and poor off axis response, while mine is so-so on axis but pretty good off axis, which is better?
This has been my primary complaint with a lot of the articles of Toole. He frequently comes to a conclusion that if a speaker measures well in all regards then it will be preferred in well controlled blind tests. This is hardly an earth shattering conclusion.
What I want to know (as a lazy speaker designer) is which parameters can I totally ignore and still have a well regarded speaker in the end. Which parameters are unimportant in subjective impression, or are there at least thresholds of "good enough" for some parameters?
This is whats wrong with these kind of discussions. Everyone comes in with an "everything is important" "optimize everything" view and we can all conclude that constant directivity or smooth directivity is a great thing. ("It can't hurt.")
Well controlled subjective tests are another matter.
Much more interesting are Olive's study where he came up with the weighting factors needed to create a rank ordering of a group of speakers that matched the blind listening ranking. He was able to do so with a high degree of correlation. (See Floyd's book.) The trouble is that he did the study twice with two different groups and came up with different weightings for different factors.
I still maintain that if you dig down through Floyd's earlier works, especially the early 2 part paper, you will conclude that power response (and therefore directivity index) has a poor or even negative correlation with subjective ranking.
David S.
I wonder... Say one were doing a very well designed "preference" type of test of high directivity vs. low directivity, with matched power response / subjective response in the room. These are fuzzy details, I'm no expert in setting up such tests, but you get the idea: the speakers sound the same as much as possible and both have pretty uniform off-axis response, but one is very broad dispersion, and one tightly controlled, and the room is fairly reflective so you can hear the difference in levels of reflected sound between the two speakers.
Now, what if you start applying a high quality realistic reverb effect to the more controlled directivity speaker? People like reverb, and I feel like that has to muck up these sorts studies a bit. No?
Hi Dave
I would completely agree that the discussion was pretty shallow in many ways. We stayed in areas where we all agreed and avoided areas where we don't. I don't agree with Floyd on precisely this topic - directivity. Floyd wants spaciousness in his playback in a room with basically fixed acoustics. The kind of approach that manufacturers look at where by the room is basically a given.
I look to modify the room so that a wider set of options comes into play. Basically narrow directivity to enhance image and a very reverberant room to enhance spaciousness. This is not a situation that Floyd ever looked at.
I also believe that there are things that go on at differing SPL levels. We could call this nonlinear distortion, but we have to remember that this does not mean that the system is nonlinear. It could just as well be that the perception is nonlinear. This is again not something that Toole and/or Olive has ever looked at. Their work is done mostly at a fixed level and usually a fairly low level at that.
I think that it is critical to put aside those things that we know are factors and discuss those that are not so well known. The above talk gets at about 80-90% of factors of importance, but it is the remaining factors that separate the really good systems from the Boss wave radios.
I would completely agree that the discussion was pretty shallow in many ways. We stayed in areas where we all agreed and avoided areas where we don't. I don't agree with Floyd on precisely this topic - directivity. Floyd wants spaciousness in his playback in a room with basically fixed acoustics. The kind of approach that manufacturers look at where by the room is basically a given.
I look to modify the room so that a wider set of options comes into play. Basically narrow directivity to enhance image and a very reverberant room to enhance spaciousness. This is not a situation that Floyd ever looked at.
I also believe that there are things that go on at differing SPL levels. We could call this nonlinear distortion, but we have to remember that this does not mean that the system is nonlinear. It could just as well be that the perception is nonlinear. This is again not something that Toole and/or Olive has ever looked at. Their work is done mostly at a fixed level and usually a fairly low level at that.
I think that it is critical to put aside those things that we know are factors and discuss those that are not so well known. The above talk gets at about 80-90% of factors of importance, but it is the remaining factors that separate the really good systems from the Boss wave radios.
This where you will get a disagreement because most rooms are NOT like this and so the results will be biased towards the more directive speaker (I would
expect.)
I once talked with Floyd about how one could do such a test and he was not optimistic. He found all kinds of complications etc. It was clear that he would not have been a fan of such a study.
I would optimize both the room and the speaker together and compare those two results. Using a fixed room acoustic will always tend to benefit one style more than the other.
I would also be careful with the definition of "preference" as this can lead to results that are so variant as to be pointless. This was the problem with the Orion versus Behringer study. I was asked to enter my speakers in this test, but I declined because the user was allowed to use whatever criteria of "preference" that they wanted. In the end the results did not really converge on any answer and so it looked like the Behringers were comparable to the Orions. Maybe they were and maybe not. Sometimes the question that you ask defines the result and sometimes it just doesn't define anything.
In all of the subjective tests that I have done there was a reference unmodified piece and a known and controlled modification. We could always find correlations, if they existed, and if they did not exist then I was confident that it wasn't because the question was ill-defined.
Last edited:
Most rooms are not fairly reflective? In any case, point taken, but the concept of the "test" was just hypothetical. I was just trying to say that I think extremely often, people will prefer a given recording with significantly more reverb than it has, and it might be hard to account for that in any sort of subjective test related to this topic. Or, hard to decide HOW to account for that, I guess you might say.
Last edited:
Some of us may have been telling ourselves we'll do that, for x number of decades and been looking forward to an all out assault some day.
What I am finding harder to process is things such as the 700Hz I often see. It's not that I disbelieve it. There are practical reasons for crossing these lines even before the arguments that it is a very broad line, and everyone is a little different.
"It can't hurt", is not something that bothers many a DIYer. I recall a particular 1940's Australian car that later became popular with restorers. The manufacturer wasn't certain with regards to maintaining the bodywork and used enough coats of paint that when these cars were eventually left to rust, they almost didn't.
I can't say I particularly like electronic reverb. Nor does conventional electronic reverb sound like the ambience of a recording room, my room or any room. It's just a one dimensional distraction.
Regarding low rating of the meaning of dispersion, gedlee made a good point - Harman's dedicated evaluating room is far from a domestic living or listening room. There on-axis response dominates and directivity effects are eliminated. I don't know if all experiments are done here. And in the picture, there is only one speaker playing.
The quality of whatever "conventional electronic reverb" you've experienced or been aware of is kind of irrelevant to the point I was making. You may substitute real reverb being magically added to original recording in my previous posts if you're skeptical about how realistic dsp reverb can be.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.