It is different but the question is "is 18-25ms a requirement or can the same perceptual effect be achieved with shorter delays"?
Markus
I would have thought that you would know the answer. If the side wall reflections begin to be in the < 10 ms range then the image will be strongly affected even if the side wall reflections do add some spaciousness. You need to keep the strong early reflections down while creating as much lateral later reflections as possible. I have said this so many times before.
I would have thought that you would know the answer. If the side wall reflections begin to be in the < 10 ms range then the image will be strongly affected even if the side wall reflections do add some spaciousness. You need to keep the strong early reflections down while creating as much lateral later reflections as possible. I have said this so many times before.
Early frontal lateral reflections are needed to generate plausible ASW. Sound eminating only from the speakers without any early frontal reflections will sound artificial.
- Elias
- Elias
Earl, I was wondering if you know what counts as "reflection"? In the work that you've done and that you've read, are these reflections mostly off of flat, hard surfaces? Do reflections that are scattered off a Schroeder diffusor or similar, have as big an effect?
Basically I'm asking if there is equivalent energy coming off a flat wall and a diffuse reflector, how much difference does that make in the listening tests? If you know of some info on that, I'd be happy to read it. Thanks.
(I have F. Alton Everest's master handbook, need to dig thru there again)
10ms is not 18-25ms 🙂 For a 10ms delay the speaker needs to be about 2m from the side walls. Probably achievable in a lot of rooms.
We could also add additional speakers and run them trough a delay.
There is a problem with the term "reflection" because it is actually a continuum, but we never say "100% reflection" or "10% refelection" let alone talk about if it is specular or diffuse or its frequency spectrum. These "details" are all important.
My research says that it is important that any early reflections that do occur be spectraly nuetral. That is because the audibility of a reflection at a given level will increase as the reflection differs spectrally from the direct. So this would rule out the desirability of anything that is not spectrally nuetral as being a good thing. No diffusers that I know of are spectrally nuetral.
I recently read a recent study of the literature on "The Precidence Effect" (don't recal the exact title, I will post it if I come across it). I was amazed at how misinterpreted this idea is. The author makes a strong point sbout this. He would have us stop using the term because he says that as it is used now it encompasses far too much.
Lidia and I are planning a test of exactly what we are atlking about here because after a literature search I have concluded that no one has looked at this specific issue.
To me the only effective way to surpress the Very Early Reflections is with narrow directivity aimed away from the near side wall. If the speaker is CD then the reflections that do occur off of the near wall are at least spectrally nuetral if the wall is flat and reflective (as mine are).
Some here look for a high ASW and strong spaciosness as a good thing. I see them as bad if in fact the source material is NOT very wide and is fairly dry (as much of my music is). When you have a playback system that adds a lot of ASW and Spaciousness then this effect will taint any source material that you play. Thats not "Hi-Fi".
Like Marcus says, if you want these effects then go with multichannel, but leave the mains as narrow directivity CD to maintain good imaging.
Last edited:
I don't think it would be difficult to have a large reflector on the side walls that difflected the first bounce away from the listener and over to a diffuser on the other side. That should easily get us to the 20 ms. range. Even something like foam core makes a good upper range reflector for experimentation.
David
Would reflections caused by a loudspeaker which has a directivity from 360 at low frequencies to 90 degrees > 1 kHz really be spectrally neutral? If we look at the Abbey and say that the reflected sound comes from 90 degrees off-axis from the loudspeaker, which seems likely with a toed-in speaker, the reflected sound will have a difference of about 24 dB between 100 Hz and 5 kHz. Seems like a lot.
The reflection will not be audible since it's very low in level over most of the spectrum, so it will achieve the goal of preventing image shift. But I have a hard time regarding it as spectrally neutral. Also, if the level would be about 10-15 dB down, at delays > 5 ms it will already be below the image shift threshold (in a live room). I would conclude that somewhat a somewhat wider coverage angle would also do the trick.
Thanks Earl! 
I'd be happy to know what you find. Please keep us in the loop.
Like the BBC control room that I did a simulation of a few months back. That was the idea, to send early reflections away from the mix position.
I just so happen to have 4 large panels of the stuff. I'll play around with it and post my results.
Having flat spectral content to your reflections always strikes me as one of those audiophile precepts, like flat group delay. With all the studies that confirm that flat power responae is undesirable, then at least some of the reflections will have to deviate from the direct response.
I've done some simulations and reflections sound much like their comb filtered combined response. That is, you tend to hear the widest (first) dip or the frequency just below it. The exact delay sets the comb filter frequencies and has a much greater effect than any deviation in response between direct and reflected. Certainly having a reflection with flat response doesn't mean that all is well or that the reflection has no perceivedmresponse effect.
I have seen that well above the first comb filter frequency, when the effects become spectrally dense, then you will notice the broad trends. For example, if you roll off the reflected spectra, as you would with thin damping on the walls, the perceived combined spectrum is duller than if it were flat.
I had a discussion with a designer who worked on the API Omnisats and he felt that the spectral balance with such an omni speaker needed a gentle rolloff to the direct sound as well as the power response to sound balanced. We seem to be well used to a duller and duller response for later reflections.
I would still buy what L&R concluded: flat power response is always too bright and dips in the power response are much less bothersome than peaks. A requirement for flat spectrum for reflections sounds like good marketing for those that design speakers that comply with the notion.
David S
Simply use Earl's speakers and setup recommendation (45° toe-in) then you end up with a strong contralateral reflection.
We have talked about this a lot as well. Sure it would be great to have 90° control down to 200 Hz. but that is neither practical nor, IMO, necessary. Frequencies below about 700 Hz do not have a strong influence on image perception. In most of the literature, say from Blauert and the like, they do not even look below about 500 Hz. From 700 Hz and above then becomes the real target for controlling the directivity. The Abbey has very good control above about 700 Hz. When I say "spectraly nuetral" then I mean > 700 Hz.
24 dB between 100 and 5 kHz is not the issue, because it's likely only a few dB between 700 Hz and 5 kHz. It Rises rapidly below 700 Hz of course.
It's the lateral versus contra-lateral reflections that Lidia and I will study. There is simply nothing in the literature that deals with this specific question - we looked. In fact the best paper that we found, "The Precidence Effect" by Litovsky, Colburn, Yost and Guzman, JASA Oct. 1999 (the one that I refered to earlier), does not even make a distinction between lateral and contra-lateral as if they were the same thing. The paper makes a strong case for mass confusion of the real psycho-physical effects. The three effects of "fusion", "localization dominance" and "discrimination suppression" all appear to be called "precidence" in the literature, but have very different underlying mechanisms.
I am not sure which "difference" you are refering to.
The literature makes a strong case that while "precidence" may dictate the principle source direction, the reflections stongly affect the perceived timbre or coloration and the dominance of the perceived image location. Again, I would claim that this effect is strongest above 700 Hz.
Below 700 Hz there are several more regions that must be contended with. From 200 Hz to 700 Hz there is some coloration potential from a strong reflection, but probably no image location implications. Below 200 Hz is all steady state and this is dealt with by EQ of multiple sources hence no coloration of imaging effects should be expected. The region between 200 Hz and 700 Hz deserves some study as I know that it is less important, but the degree of this is just a guess. As I said, for the most part the literature stops below about 500 Hz.
The subjective effects of a dipole from a monopole would be dominate in this spectral region and may be a reason for the strong support of dipoles. But the dipoles that I have heard have been so poor above 700 Hz that this benefit does not make me want to go down that road. A dipole and a waveguide has some appeal to me for this reason, but I can't find a reasonable solution to the problems of actually making a practical product that does not push the costs through the roof. On a weighted scale of perceptual improtance, the benefits of an improvement in 200 Hz - 700 Hz do not justify a doubling of costs.
As the gangster said to his good friend just before he executed him - "It's just business".
Don't be too quick to dismiss sidewall reflections between 200-700Hz affecting image location.
If the side wall distance of left and right speaker are significantly different (and one or both are within about a metre of the side wall) the comb filtering peaks/dips in the lower midrange occur at different frequencies, and this can "pull" the image location left and right depending on the spectral content of the music.
I find unequal sidewall spacing quite problematic for good imaging even on speakers that are relatively directional at high frequencies.. if they're less than about a metre both need to be identical distances or the imaging is never optimal due to large differences in the left-right lower midrange balance.
Simon
In no setup that I have ever done are the left and right distances unequal so subjectively I can't comment on what happens if they are. I would inherently think that this would setup would not be a "good" idea. But I would discount the potential for strong image effects if the reflections are limited to < 700 Hz. If there is significant reflection content above that frequency then I would completly agree.
In no setup that I have ever done are the left and right distances unequal so subjectively I can't comment on what happens if they are. I would inherently think that this would setup would not be a "good" idea. But I would discount the potential for strong image effects if the reflections are limited to < 700 Hz. If there is significant reflection content above that frequency then I would completly agree.
If the goal is to hear the recording and recording signal as well as possible, lateral reflections wouldn't be beneficial as long as they are arriving too early. And in almost every small room they will; no matter which wall they come from.
Whether one likes them or not is a question about preference and music material.
Omholt
Its not that black and white.
They can be vastly different in timing and level with different loudspeaker designs and room setups.
Its not that black and white.
They can be vastly different in timing and level with different loudspeaker designs and room setups.
- Status
- Not open for further replies.