Dave I also believe that this is very true as well.
But since this is not possible, I will go with what is the better compromise of the situations. Wide directivity in a reverberant space is a disaster. Narrow directivity in a dead space is pointless. Wide directivity in a dead space can be acceptable for venues where we want "I am there with a large orchestra", but will fail for the "They are here" situation. Only the narrow directivity in a reverberant space has a decent compromise on both scenarios. It is excellent for the "They are here" and acceptable for the "You are there."
"Full" figure 8 only up to 4 kHz in this case. 6 kHz is already -10 dB in the rear 🙄:
The front wall to side wall corner is drywall. No acoustical treatment at all.
Rudolf
Attachments
Hi, guys!
I agree with dr. Geddes on the matter of RT60. It is not important per se, but there has to be a sort of standard for rate-of-fall of energy in a room, so it could be whatever a group of people agree upon. So long if we do not missinterpret the data that is provided by the parameter.
If we think of a room as a system with infinetely rigid boundaries, which is not in practice but frequncy dependant, only thing for what we can use RT60 is determining the right amount of absorbing material, ie energy stabilisation. More important is the acoustic 4D geometry on which our hearing depends on, and that is the point Toole (et alii) described in his work. Stereo cannot fully describe live performances, but properly designed room can merge reproduction from loudspeakers and the room giving ambiance for certain types of music, like in the design of acoustical spaces for recording. 'They are here' isn't a very pleasant way of listening music, becase it draws too much attention to what's happening between the speakers, so too narrow dispersion should be avoided for anything else but recording studios, where you have to listen for errors. Too dry, and it doesn't come 'out of the speakers'.
In my practice, I would opt for 90 degrees from speech upwards as a practical limit, offering good driver integration in the vertical plane. Sort of like Tooles.
Directionality in lowest octaves is only an expensive and hard to achieve way of combating room anomalies. Make the room right and there is no more problem.
I agree with dr. Geddes on the matter of RT60. It is not important per se, but there has to be a sort of standard for rate-of-fall of energy in a room, so it could be whatever a group of people agree upon. So long if we do not missinterpret the data that is provided by the parameter.
If we think of a room as a system with infinetely rigid boundaries, which is not in practice but frequncy dependant, only thing for what we can use RT60 is determining the right amount of absorbing material, ie energy stabilisation. More important is the acoustic 4D geometry on which our hearing depends on, and that is the point Toole (et alii) described in his work. Stereo cannot fully describe live performances, but properly designed room can merge reproduction from loudspeakers and the room giving ambiance for certain types of music, like in the design of acoustical spaces for recording. 'They are here' isn't a very pleasant way of listening music, becase it draws too much attention to what's happening between the speakers, so too narrow dispersion should be avoided for anything else but recording studios, where you have to listen for errors. Too dry, and it doesn't come 'out of the speakers'.
In my practice, I would opt for 90 degrees from speech upwards as a practical limit, offering good driver integration in the vertical plane. Sort of like Tooles.
Directionality in lowest octaves is only an expensive and hard to achieve way of combating room anomalies. Make the room right and there is no more problem.
This is preciesely what I suggest as well, but 90 degrees is narrow directivity for most systems. Take the Moulton system described above, it is 180 dgerees. I consider 90 degree CD to be "narrow constant directivity".
Rudolf, I dunno what to think, because I remember that (with only me as testing subject) I was putting my speakers in every possible position and the result was of the same type. Same also with dome tweeters and these 4" that are dipoles. Just the test was more rich with the 4" that allows the use of lower freq.
This compromises the pure responsibility of the reflection path ??? I insist on the action of the filter in the critical zone 1200/1500. With the DCX it's evident.
The azimuth determination of a source (combined or mono) with pink noise is one of the things that our auditory system makes the best, could reflections contribute to the final judgement.
This compromises the pure responsibility of the reflection path ??? I insist on the action of the filter in the critical zone 1200/1500. With the DCX it's evident.
The azimuth determination of a source (combined or mono) with pink noise is one of the things that our auditory system makes the best, could reflections contribute to the final judgement.
This is the way it always is.
I assume this is for the high frequency noise. If you look at coherent noise between two speakers with an RTA, it is very dependent on the exact location of the mic. You will see the comb filtereing of the noise is you move even a little left or right. A dip starts at the highest frequency and then shifts downwards as you move off the center line (more dips join it). You can use the phenomenon to find the exact center line between two speakers.
Now picture the pattern of either a single HF tone or say just an octave of HF noise. From left to right across the listening distance there will be a whole sequence of level peaks and nulls. If you shift your head a little left from the center your left ear will fall into the first null. You will have no sound (relatively) in the left ear and full sound in the right. Where will your perceived image go? Hard right.
The same thing happens at mid frequencies but you must move farther in one direction to find the first null. Because of this you will have a little more seating latitude and a more solid image, but precedence will still cause the image to shift.
This effect is dependent on the angular spacing of your speakers. For example if you spread the speakers very wide (over 120 degrees apart) the comb filtering essentially goes away because your directional hearing seperates one speaker from the other (left ear hears left speaker but not right, for HF). As you shift off the center line the sound smoothly shifts from one ear to the other, rather than the "swish-swish" of comb filtering, or the rapid left-right-left HF jumping. When both speakers are to the front, not very far seperated, then the comb filtering is very strong.
Those of you having a hard time getting the high frequencies to center image should try slowly shifting left and right to see if you can find the particular spot where the two speakers are totally in phase (at least for one ear!).
David S.
Thanks for the explanation Dave that makes perfect sense 🙂 yes it was at 3Khz that I was doing this. Also my speakers are not far apart probably only about 1.8M but the distance to the seating position is probably about 3.5M.
The other thing I noticed was that when sitting centrally the image seemed to be slightly to the left. At first I was thinking it might be a balance problem, but the distance to the left wall is only about 1.8M but the distance to the right wall is about 6M (as I said not an ideal listening setup 😉 ) so it could be due to that that the image seems to be slightly shifted to the left.
Tony.
The other thing I noticed was that when sitting centrally the image seemed to be slightly to the left. At first I was thinking it might be a balance problem, but the distance to the left wall is only about 1.8M but the distance to the right wall is about 6M (as I said not an ideal listening setup 😉 ) so it could be due to that that the image seems to be slightly shifted to the left.
Tony.
This thread is moving fast! Lot's of good info this morning. I'll trhow in a few comments.
It's hard to say wrt to topic. I don't know the polar response of the system used, but I'd guess it's pretty tight. Even the Onken W with its side by side 15s has a pattern more vertical than horizontal. Most of the immense sound came from those big WE 15A horns, I believe. The very low self noise of the other sections didn't hurt.
It's true that half way up the seats the angle got pretty narrow, but the effect didn't go away. Perhaps because the speakers themselves were so large? I also think that because the spacial effect in front was so unusually strong we were simply overwhelmed and didn't think about anything else. There was so much information presented in front that the sides just didn't matter. I've had similar experiences watching some 70mm film prints. The visual information presented in front is so rich you forget that most of what's around you is just a dark cinema.
It's hard to say wrt to topic. I don't know the polar response of the system used, but I'd guess it's pretty tight. Even the Onken W with its side by side 15s has a pattern more vertical than horizontal. Most of the immense sound came from those big WE 15A horns, I believe. The very low self noise of the other sections didn't hurt.
It's true that half way up the seats the angle got pretty narrow, but the effect didn't go away. Perhaps because the speakers themselves were so large? I also think that because the spacial effect in front was so unusually strong we were simply overwhelmed and didn't think about anything else. There was so much information presented in front that the sides just didn't matter. I've had similar experiences watching some 70mm film prints. The visual information presented in front is so rich you forget that most of what's around you is just a dark cinema.
Everything you described there is accurate, but I'm not sure the "swish swish" panning effect is quite what wintermute was experiencing or describing here.
It sounded to me more like the effects of excessive toe in. The idea behind toeing in front of the listener as I'm sure you know is to offset the natural tendency for the phantom image to collapse to the side when you move left or right.
Normally if you move to the left, the right speaker becomes further away than the left, so the left speaker becomes both slightly louder, and slightly earlier in time than the right, and it doesn't take much shift for the precedence effect to collapse the phantom channel to the left.
Toeing speakers to cross in front of the listener can counteract this to a degree, because as you move left you move more off axis to the left speaker, but more on axis to the right speaker - so with fairly directional speakers this counteracts the tendency of the left speaker to become louder, and can in fact make the right speaker become slightly louder (at high frequencies anyway) than the left as you move left, (at least until you reach on-axis to the right speaker) allowing the precedence effect to be overridden to a small degree. The end result is a much wider sweet spot before the image will collapse to the side.
Toe in too much though and the increase in volume of the further away speaker and decrease in volume of the near speaker occurs too quickly and the phantom image appears to move away from you - if you move left, the phantom goes right. This is a sign that the speakers are crossed too far in front of the listener, or the directivity of the speakers doesn't suit a strongly toed in configuration. (CD is ideal here, although it seems to work reasonably well even with speakers that are non-CD but somewhat directional at high frequencies)
wintermute did the right thing by backing off the toe in angle a little bit until this effect went away. Too little toe in and the phantom image will move in the same direction as the listener, too much and it will move away from the direction of the listener.
Overlaid on top of this is the "swish swish" effect you describe where the apparent phantom location rapidly oscillates back and forth between left and right as you move a small distance sideways, and I believe is caused by comb filtering as you describe.
In general I would say if this latter effect is quite strong its a good sign that the angular separation of the speakers is insufficient, at 45 degrees on my speakers it is quite noticeable. (Can't do anything about it in this room set-up though)
At 60-70 degrees I find it quite minimal, and I think the wide separation works quite well with a crossed in front of the listener toe angle as that helps to stabilize and focus the phantom image despite the wide speaker separation. With a sufficient (but not excessive) separation angle and just the right amount of toe in I think you can get quite a stable phantom image that doesn't move much from the centre with significant lateral head movement...
Mono pink noise is very revealing of both these phenomena, and ideal for doing the initial set-up of angular separation and toe in, as well as checking the speakers are equidistant.
Last edited:
@dr.Geddes
Of course, it could be expanded to fit a particular situation for image broadening, 90 degrees is quite enough, as long as it generates diffuse reflections from lateral walls. Image depth can be added by decreasing the amount of back wall absorption. Depth is an artefact and doesn't really exist in stereo.
Moulton was in ellipsoid room, that's the key for that speaker arrangement. Kind of LEDE-RFZ. I haven't listened to it, but I think it could be viable within similar acoustic space. Not for a living room.
----
Of course there is direct to reverberation ratio.... good listening experience is not just how much a loudspeaker disperses the sound. It's all that and timing control within a room. There are no easy steps in achieving that without a dedicated, purpose built listening room. It serves it's purpose, just as a bathroom.
Gosh ! gonna see that to morrow.
Difuse lateral reflections as long as they are > 10 ms. Shorter than that will cause problems with imaging.
I have found that the depth perception increases as the back wall (behind the speakers) has more absorption. You are suggesting the opposite. I'm not sure why that would be.
I meant the back wall behind the listener. Back wall behind the speakers and side walls (up to few-10 ms, as long as it doesn't introduce linearity anomalies) should be broadband absorptive, as you suggested. Any more of the side walls and you lose the advantages of widening. It is an inherent way to limit dispersion to a maximum of 180 degrees (from the speaker plane to the listener). That, or in wall instalation, which is not good because our brain doesnt like viewing obstacles in the simulated 3D sound field. Bad Feng Shui.
PS That only apllies to monopole/cardioid loudspeakers. For dipoles there is different energy distribution.
I find the same.
I can't prove it scientifically, but I think a lot of depth perception (whether it's an auditory illusion or not) comes from low frequencies - 100-300Hz or thereabouts, and deep notches in this range can tend to "compress" the sense of depth of the sound field, regardless of whether they are caused by floor bounce, "back" (behind the speaker) wall bounce, or modal effects.
Eliminate those deep notches and I find the sense of depth improves. The amount of absorption you have behind your speakers would certainly improve the response in this region. (Have you measured the change with and without between 100-300Hz ?)
As you know, the issue is time intensity trading. Shift sideways and there is a time discrepency that shifts the image. Increase the level of the far source the right amount and the image will be pulled back to center.
How much level for how much shift?
I'm looking at a Deutsche Welle document .
http://www9.dw-world.de/rtc/infotheque/sound_perception/sound_perception.pdf
See figure 7 (Everybody read this paper because it is a good overview of what we are talking about!!). It shows a very linear dependence with 2ms of shift being offset with 16 dB of level.
I'm dubious about whether a speaker has enough directivity to create a level shift adequate to offset the lateral poistion shift, let alone overcompensate for it. 2ms of shift means you moved about a 1ft left (1ms farther from 1 speaker, 1ms closer to the other), lets call it a ft and a half since were are moving away on an angle, and in this distance the polar patern of the units has to rise/fall over 16dB to overcompensate.
Laser beam speakers.
David S.
David, as concerns the demos in Paris, I would say yes, no and I don't know.
As I was very familiar with the sound of that space - I practically lived there - the space recreated by the system was often much larger and different than any I'd heard before or since in that venue. I'd heard a lot of live and recorded music in there but it didn't sound like this. Nothing else had that immense sound of a much larger venue. There were a few amazing symphonic and choral recordings that had completely believable width, depth and height clues. It was the precise height clues that many of us found so surprising.
It was not all "you are there". Some of the jazz and pop recordings where very much "there are here, right on the stage."
Was the space anything I'd heard before in there? No
Did the size of the venue help? Yes, for sure.
Were the acoustics of the venue the main factor? I don't know, but I don't think so.
I've played some of the those amazing recordings on many other systems, some were larger systems in larger rooms, and they still never came close to the demo. Why, I don't know.
Just FYI, here is a diagram of the demo space. Much larger than a domestic listening room, but not the size of a symphony hall.
Attachments
At Snell and McIntosh we made high performance in-walls. I occaisionally ran into salesmen that said "even good inwalls never have any depth". Some people said the same about tne Mac line arrays. It never made any sense to me (lack of a back wall bounce should be a good thing) and I considered it an audiophile "bias", but it was told to me more than once. (For what its worth.)
David S.
I agree. For you and me it probably comes down to one choice, controlled directivity speakers. But there are so many folks who really enjoy omni and dipole that those choices remain valid. Even I like them, sometimes. 😉
Last year when I was a judge at the Dayton speaker contest we rolled thru a lot of speakers that were wildly different. Horns, cones and domes, line arrays, multi-way, single driver, flat panel. They all sounded different, for sure.
One that particularly struck us (Don Keele, Jerry McNutt and me) was a pair built from Walsh/Ohm parts. Insanely good and precise imaging. Don't think I've ever hear anything quite like it. We all commented on it. Don't know what the mids and tweets were doing, but I sure would like to. Bass was omni and had its flaws, but the mids and highs were pure magic. I don't know how it was done.
This is a great goal, and I agree that when it is accomplished, the results are noticeably better. The thing is, the vocal band starts around 200Hz, and male voice even goes doen to 100Hz. So there are very few systems that can actually do this.
- Status
- Not open for further replies.