I agree, but..
not "entirely", IMHO..
delay, strength and spectrum of reflections.. Is it yet fully understood?
Kill all the reflections, what do you then get? A precise but small "TV" sound. Then you need multi-channel, but maybe this is the only way to get best of both worlds.. or maybe not..
not "entirely", IMHO..
delay, strength and spectrum of reflections.. Is it yet fully understood?
Kill all the reflections, what do you then get? A precise but small "TV" sound. Then you need multi-channel, but maybe this is the only way to get best of both worlds.. or maybe not..
I use dipoles and adapt the toe in depending on the material. 45 deg for "studio" music, much less for real spaces.
I do exactly the same now!
This is actually the "genius of dipoles"
Dipoles need some air to breath around them, but you can take advantage of high directivity and strong backwave to create a "lively" acoustic image of sterophonic recordings.
One easy way to evaluate speakers' "control of directivity" in a room is to listen to mono sound reproduced with a pair. A male and a female voice from a radio station is an easy way to get that. First do a sanity check with headphones to verify mono recording and minimal reverb and echo in the source! Speech has many frequencies in the most critical range and smearing of the sound image is easy to detect. The cause of that smearing may be a more difficult problem...
hello my Finish friend! 🙂
dipoles are great, but.. to my ears they need quite a bit of distance to the front wall, >2m. treating the ceiling, floor and middle of front wall helps a lot!
I don't think a "strong" backwave really helps, it makes the recordings all sound a bit the same, although it can be a pleasant effect.
Are you going to attend the next AES meeting?
dipoles are great, but.. to my ears they need quite a bit of distance to the front wall, >2m. treating the ceiling, floor and middle of front wall helps a lot!
I don't think a "strong" backwave really helps, it makes the recordings all sound a bit the same, although it can be a pleasant effect.
Are you going to attend the next AES meeting?
Last edited:
lolo friend!
Yes we all like a little different acoustics and rooms are always different too. I am sorry to say that my living room is far from perfect, but dipoles work a lot better that conventional 2-ways with waveguide tweeters. Moving the speakers further off the front wall changes image sharper and more narrow - it suits some recording and attitudes of me.
I will participate the preliminary tutorial day of the 51st AES Conference in Helsinki. It is free for Finnish members. Sorry, I won't be listening to Genelec founder Martikainen or S Linkwitz telling about his speaker projexts the next day.
Yes we all like a little different acoustics and rooms are always different too. I am sorry to say that my living room is far from perfect, but dipoles work a lot better that conventional 2-ways with waveguide tweeters. Moving the speakers further off the front wall changes image sharper and more narrow - it suits some recording and attitudes of me.
I will participate the preliminary tutorial day of the 51st AES Conference in Helsinki. It is free for Finnish members. Sorry, I won't be listening to Genelec founder Martikainen or S Linkwitz telling about his speaker projexts the next day.
“with all respect Mr Danley, is this not the exact opposite of what the psychoacoustic research says? (project Archimedes, Moulton and Toole, IACC related concert hall literature..).”
Well no as well as possibly so.
What I am talking about is preserving a recorded stereo image, not making an enjoyable presentation in a room.
AS I have suggested a number of times, unless one listens in the very nearfleld or sets up a stereo outdoors away from reflections, one has no baseline for what is possible without these reflections.
Also it is difficult to separate the limiting factors in the loudspeakers themselves from this kind of evaluation of reflections.
For example, even outdoors or near field, most loudspeakers when fed a mono signal produce a mono phantom BUT also the R and L loudspeakers are localizable so that one has the center image AND an obvious source on the right and left sides.
If you have loudspeakers which radiate “simply”, there is less difference between what arrives at the right and left ears and so they more easily disappear in favor of the mono phantom image and it is possible to produce a mono phantom without an awareness of the R and L speakers producing it and IF one can do that, one can make an image appear anywhere between the two sources with the center position being the most difficult.
Short times delays DO mater, take the work on interaural crosstalk like these folks are doing.
3D3A Lab at Princeton University
Note they say directivity is one of the key things needed to do this successfully, the reason is the reflected sound competes with what you hear directly and such cross talk cancellation is possible by introducing a canceling signal separated by a very short time period, even shorter than room reflections.
Some friends are also working on that approach, pop on some headphones and listen to their recordings using this approach (cross talk cancellation)
ImmersAV Loudspeaker 3D Audio Format Presentation - YouTube
These are the one condition where “adding” something NOT in the recording actually helps improve the realism BUT it is a precise thing they are doing, not room dependent variable.
Another example of how very short time related things have a profound effect on what we hear. Co-worker Doug Jones made some recordings in the way old days which utilize the very short time related aspects of the outer ears to simulate position changes. These are also strongly dependent on the preservation of information reaching your ears and are harmed by reflected sound and complex loudspeaker radiation.
Online LEDR Sound Test | Listening Environment Diagnostic Recording Test
A thought experiment; Imagine recording a test signal that was a broad band impulse and you examine the “Energy vs time” view of that signal. Does it make sense that significantly adding late energy , versions or snippits of the original signal delayed in time could in any way, make what arrives at the LP more like the original signal?
I got to hear two pairs of radically different loudspeakers this spring, located in the same place, one was a “curved and amplitude shaded array” mentioned here and the other a single point CD horns.
They both had a nearly identical spectral balance and sounded VERY similar. The array my friend built made a captivating wall of sound which extended well past the outer locations of the speakers and was a beautiful effect.
The thing they didn’t do by comparison was produce a strong mono phantom image or conventional stereo image. The single point system when fed with a voice in mono, produced the impression the voice was floating dead center and produced a stereo image that extended between the two sources.
The difference in what arrives at the lp is a single arrival from the single point system and a train of arrivals from the array, each arrival delayed in time according to the distance to each source. This is evidence that how we hear also involves VERY short time delays.
These things have convinced me that following what appears to be “how we hear” is the direction I should go with my job / work.
While we don’t sell “hifi” speakers capturing and reproducing the signal IS my primary interest.
IF you had headphones on for the other videos, pop them back on and try the recordings here which is with a time preserving microphone array I have been working on . Try the parade recordings from this 4th of July.
https://www.dropbox.com/sh/hgzo98h2wqe3r9l/aqrDcSkPVl
On the reproducing side, I presume most of you have heard a large concert sound system or sound at a large sports stadium had have heard the effect of an array of a large number of sources / arrivals in time.
These videos are from a very large empty “room” (tiger stadium at LSU) and while you can hear longer time reflections, one can also hear it sounds pretty different than the concert arrays (even the largest of which cannot project this far). Also, the line arrays with all the sources radiate a complex interference pattern which is very audible when even a slight wind blows, not the one video with my daughter showing how hard the wind was blowing and how little it affected the sound.
https://www.dropbox.com/sh/nmmmdtum82lyig9/QnEaYWlnDE
Adding sounds not present in the recording may sound pleasing but it does not help preserve what is in the recording.
Best,
Tom
Well no as well as possibly so.
What I am talking about is preserving a recorded stereo image, not making an enjoyable presentation in a room.
AS I have suggested a number of times, unless one listens in the very nearfleld or sets up a stereo outdoors away from reflections, one has no baseline for what is possible without these reflections.
Also it is difficult to separate the limiting factors in the loudspeakers themselves from this kind of evaluation of reflections.
For example, even outdoors or near field, most loudspeakers when fed a mono signal produce a mono phantom BUT also the R and L loudspeakers are localizable so that one has the center image AND an obvious source on the right and left sides.
If you have loudspeakers which radiate “simply”, there is less difference between what arrives at the right and left ears and so they more easily disappear in favor of the mono phantom image and it is possible to produce a mono phantom without an awareness of the R and L speakers producing it and IF one can do that, one can make an image appear anywhere between the two sources with the center position being the most difficult.
Short times delays DO mater, take the work on interaural crosstalk like these folks are doing.
3D3A Lab at Princeton University
Note they say directivity is one of the key things needed to do this successfully, the reason is the reflected sound competes with what you hear directly and such cross talk cancellation is possible by introducing a canceling signal separated by a very short time period, even shorter than room reflections.
Some friends are also working on that approach, pop on some headphones and listen to their recordings using this approach (cross talk cancellation)
ImmersAV Loudspeaker 3D Audio Format Presentation - YouTube
These are the one condition where “adding” something NOT in the recording actually helps improve the realism BUT it is a precise thing they are doing, not room dependent variable.
Another example of how very short time related things have a profound effect on what we hear. Co-worker Doug Jones made some recordings in the way old days which utilize the very short time related aspects of the outer ears to simulate position changes. These are also strongly dependent on the preservation of information reaching your ears and are harmed by reflected sound and complex loudspeaker radiation.
Online LEDR Sound Test | Listening Environment Diagnostic Recording Test
A thought experiment; Imagine recording a test signal that was a broad band impulse and you examine the “Energy vs time” view of that signal. Does it make sense that significantly adding late energy , versions or snippits of the original signal delayed in time could in any way, make what arrives at the LP more like the original signal?
I got to hear two pairs of radically different loudspeakers this spring, located in the same place, one was a “curved and amplitude shaded array” mentioned here and the other a single point CD horns.
They both had a nearly identical spectral balance and sounded VERY similar. The array my friend built made a captivating wall of sound which extended well past the outer locations of the speakers and was a beautiful effect.
The thing they didn’t do by comparison was produce a strong mono phantom image or conventional stereo image. The single point system when fed with a voice in mono, produced the impression the voice was floating dead center and produced a stereo image that extended between the two sources.
The difference in what arrives at the lp is a single arrival from the single point system and a train of arrivals from the array, each arrival delayed in time according to the distance to each source. This is evidence that how we hear also involves VERY short time delays.
These things have convinced me that following what appears to be “how we hear” is the direction I should go with my job / work.
While we don’t sell “hifi” speakers capturing and reproducing the signal IS my primary interest.
IF you had headphones on for the other videos, pop them back on and try the recordings here which is with a time preserving microphone array I have been working on . Try the parade recordings from this 4th of July.
https://www.dropbox.com/sh/hgzo98h2wqe3r9l/aqrDcSkPVl
On the reproducing side, I presume most of you have heard a large concert sound system or sound at a large sports stadium had have heard the effect of an array of a large number of sources / arrivals in time.
These videos are from a very large empty “room” (tiger stadium at LSU) and while you can hear longer time reflections, one can also hear it sounds pretty different than the concert arrays (even the largest of which cannot project this far). Also, the line arrays with all the sources radiate a complex interference pattern which is very audible when even a slight wind blows, not the one video with my daughter showing how hard the wind was blowing and how little it affected the sound.
https://www.dropbox.com/sh/nmmmdtum82lyig9/QnEaYWlnDE
Adding sounds not present in the recording may sound pleasing but it does not help preserve what is in the recording.
Best,
Tom
Dealing with the sidewalls may be easier, but it is still the wrong thing to do.
Markus - I do deal with the floor and ceiling reflections via room treatment, but I DO NOT treat the walls. Absorbing all of the vertical reflections would not degrade the sound while doing the same to the walls would. They are simply not them same things perceptually.
Tom - I agree completely. We seem to be lock-step on these issues. Again, you should come and hear what I have done and I think that you would agree that the approach is correct.
Last edited:
Tom Danley;3586794Adding sounds not present in the recording may sound pleasing but it does not help preserve what is in the recording. Best said:
You're right - That was not the best comparison.
I was mainly focused on your suggestion that equalization didn't solve everything. I agree with that statement, and commented about it. But I think you're right that the comparison didn't work because the PT waveguide is a fairly smooth device. See the attached response file, for example. Its peaks and dips down low aren't too bad, pretty much done by 2kHz, and need no tank circuits to "tame." In truth, I like most of the PT waveguides, most behave pretty well.
But then again, the H9800 is pretty smooth too. I've attached its response curve - using the same compression driver - for comparison. I wouldn't characterize it as having jagged response. I'd say it is, in fact, slightly smoother than the PT waveguide. It probably presents a better acoustic load towards the low end.
Each has a slightly different response, so the response shaping circuits might be different, but both show a general trend of mass rolloff. The H9800 has a little more up high, showing it has begun to beam more in the top octave. But it has better acoustic loading down low, which is demonstrated by the reduced ripple below about 3kHz. I think most would be tempted to deal with its output at 2kHz with a broad filter of some sort, but it doesn't appear resonant. Looks more like increased efficiency before mass rolloff.
Anyway, that's all a digression. The point is while you can equalize out all the ripple of a peaky device, that doesn't make it sound good. Equalization doesn't solve everything. The examples you've used - a JBL PT waveguide verses the H9800 horn - is an example where both are actually pretty good.
I understand your point that you liked the sound of the H9800 better, but we're still talking about two horns with pretty smooth response. I think it would be an even starker difference if we were comparing a resonant horn with a lot of periodic ripple to a smoother non-resonant horn. In that case, I find the qualitative differences are even greater.
There are some waveguides out there with response ripple all the way through the passband, and adopters usually use notch filters well up past the crossover frequency, basically through the whole band. That kind of thing goes more to the point.
Attachments
You don't avoid sidewall reflections with a common waveguide speaker. Actually, the ones from the opposite wall may be harder to deal with because of a more accute angle if the speakers are crossed in front of you. And from the nearest wall you most likely have reflections in a certain frequency range, which leads to coloration if you don't deal with it. A cornerhorn or larger horn is better, but still needs some sidewall treatment in most cases.
Reflections from the opposite wall are detrimental to imaging too. They simply arrive later and when depends on the room. In most small rooms they will be within 7-12 ms. That something one need to deal with if accuracy is the goal. What arrives at one ear, is heard by both.
Running after speakers that need no treatment is futile. Buy headphones if one can't treat the room.
Most of the psychoacoustic research on reproduction in a small room acoustic STARTS with a near anechoic condition, and then add's simulated or real points of reflection from there. (..at least of 3rd of Toole's book is an aggregation of research in this area.)
The Princeton research *relies* on convolving the signal to binaural - which is quite different than most studio mastering process. (..and at that point you might as well be wearing headphones which are vastly superior with binaural.)
*except* for the very fact that you advocate turning your speakers inward toward the contralateral wall.
..which is of course completely contrary to what Tom was suggesting as optimal. 😉
Because of head shadowing the second sentence is incorrect, making the first sentence an exaggeration. The far wall reflect is far less of an issue than the near wall.
ScottG;3587223..which is of course completely contrary to what Tom was suggesting as optimal. ;)[/QUOTE said:
Yes. I guess that's correct. Or the head filters some of the high frequency sound.
Still, reflection from opposite wall has an importance for accuracy. I prefer redirect it and difference has always been very audible. Absorbing it is also better then no treatment for a narrow room, but does create a little bit of the dead headphone feeling.
No, but he was strongly suggesting a dispersion pattern with respect to the listener and room boundries, that advocated
-a DECREASE (both direct sound and reflections) of:
Left channel intensity into listener's right ear, and right channel intensity into listener's left ear.
"Optimal" in this case is a reduction in cross-talk, NOT a setup that is in all likely-hood is *increasing* cross-talk - which is the setup you advocate.
(..oh, and lest some think I'm being unfair hear on a product-basis - Note that Earl's design is ALSO very well "equipped" for *decreasing* cross-talk - which is why it was used in the Princeton Lab. Basically move the speakers closer together and use "toe-out" instead of moving them further apart and using "toe-in".)
Last edited:
This depends on how near the near wall is.
In fact, with respect to overall intensity difference due to reflections (..and the rate of decline in spl with distance),
-it's quite possible that a near reflection is advantageous from a lower IACC point of view.
Sort of a: "the left side stays "hotter" (or more intense) on the left, and the right side stays "hotter" on the right.
..might be an interesting area of research.
Last edited:
Nice measurements Wayne!
I though you wanted to stick with 1" drivers.
What horn would you recommand for throatless 1.5" driver like the JBL?
Regarding the PT, mine is the rectangular version (PTF95HF), with not as good a vertical directivity control, but quite a bit shallower than the square version.
My measurements do not have the same resolution as your down low (I have to use gating, so I start to loose resolution under 1.5kHz) and did not show what you are showing here.
The main grip I have with the H9800 is its strange "ultra-wide" directivity behavior in the 3kHz-8kHz range. Did you measure it?
I found that putting layers of foam on the sides of the horn (not around, but actually *inside* the horn, along the lateral walls!) did improve on that, but I did not pursue...
Those weren't my measurements, those were from Audio Heritage. Tons of charts over there for JBL gear, and plenty of 'em are of PT waveguides and H9800 horns using various drivers. Most show general agreement with other charts, with differences between compression drivers, of course. I was reasonably comfortable dredging up those particular charts because they showed general agreement with others and because they showed a PT waveguide and an H9800 horn measured using the same compression driver and same measurement system. Passed the "sniff test."
I have measured and used both of these horns extensively with various 1.5" JBL compression drivers (2435, 2452, 2450SL, 2450SL-Be), and I can tell you that the H9800 does not measure nearly as smooth as the rectangular PT waveguide, both on and off axis.
I have no anechoic measurement to show, so no "200Hz-20kHz" curves like the ones above, but I have done enough measurements at various distances, angles and gating settings to assert the behavior of these horns accurately enough on their usable range.
I have no anechoic measurement to show, so no "200Hz-20kHz" curves like the ones above, but I have done enough measurements at various distances, angles and gating settings to assert the behavior of these horns accurately enough on their usable range.
Last edited:
- Status
- Not open for further replies.