It was about time, four years since this thread and no data yet about "IMP"
I'm not too optimistic that we will get data anytime soon
.
Gary has threads on Usenet that date back to the 90s, I get the impression that he grew tired of defending his work.
The data in this thread demonstrates that Gary's designs sound good, but people tend to freak out when you tell them that Bose makes good speakers.
Basically I'd love to learn more from him, but I can understand why he keeps a low profile.
nothing fancy. class AB chipamp and simple 2 way bass reflex with 6.5 inch midbass and a 1" tweeter.
(yeppers, chipamps have transistors in them. mind me, did not point out i dislike the sound of discrete amps vs chips amps. you got the point there. it was quite late and i got tired allready.)
Doesn't anybody know how to use a search engine? 
Hopefully Gary wasn't trying to keep this off this forum or something, but here it is. Pic permissions are fixed on the fifth(?) try down. There's also his AES paper.
https://groups.google.com/forum/#!topic/rec.audio.high-end/ao0OsXRFezU
edit: P.S. the link loads unpredictably, at least on my end, but it's just a discussion format; scroll up to the top and expand the first post if it takes you to the bottom or the middle or etc.
P.P.S. my question would be this: am I right in assuming from the lack of virtual sources across the floor and ceiling indicated in the diagram/description, and from the MTM layout, that the ideal for the concept is horizontally-omni but amplitude-shaded output, with high vertical directivity? I wonder then if waveguides, or planar drivers, or arrays might be best, and if a truckload of absorptive material to toss around the current speakers would be a way to sort of find out.
Hopefully Gary wasn't trying to keep this off this forum or something, but here it is. Pic permissions are fixed on the fifth(?) try down. There's also his AES paper.
https://groups.google.com/forum/#!topic/rec.audio.high-end/ao0OsXRFezU
edit: P.S. the link loads unpredictably, at least on my end, but it's just a discussion format; scroll up to the top and expand the first post if it takes you to the bottom or the middle or etc.
P.P.S. my question would be this: am I right in assuming from the lack of virtual sources across the floor and ceiling indicated in the diagram/description, and from the MTM layout, that the ideal for the concept is horizontally-omni but amplitude-shaded output, with high vertical directivity? I wonder then if waveguides, or planar drivers, or arrays might be best, and if a truckload of absorptive material to toss around the current speakers would be a way to sort of find out.
Last edited:
I have to go to work, so i don't have time for the full review but i can tell you that the direct sound is attenuated by 6dbLooks like the new incarnation of IMP has more direct sound than the one that won the challenge which had most elements aimed towards the wall.
In the mirror can see two back elements.
How does this differ from an omni, then ?
.
The 2 rear firing panels are full power and I believe he said the wall firing panel is down 3db.
The mirrors are for stronger sharper reflections.
Doesn't anybody know how to use a search engine?
Hopefully Gary wasn't trying to keep this off this forum or something, but here it is. Pic permissions are fixed on the fifth(?) try down. There's also his AES paper.
https://groups.google.com/forum/#!topic/rec.audio.high-end/ao0OsXRFezU
edit: P.S. the link loads unpredictably, at least on my end, but it's just a discussion format; scroll up to the top and expand the first post if it takes you to the bottom or the middle or etc.
Ok I just read this link.
It seems i had the levels wrong judging from the paper.
He or I didn't want to delve into the technical side of things, It was a listening session about his theory of creating a live listening environment.
If I went on about conventional sound theory, we never would have started listening!
Lol.
K, gotta go to work.
These "IMP" speakers appear to be a variation of the DBX Soundfield 1 speakers, which I heard and liked very much. They have the same need as any open baffle speaker, which is to be away from any walls by at least 3 feet, but if you can accommodate that need, you'll have a pretty good sound.
One thing I learned from experimenting a bunch is that the tweeter really needs to be at least 36 inches off the floor, and 42 inches is better. Higher than that is good too. The addition of the second 6 inch (or whatever) driver will narrow the beam vertically and thereby reduce floor and ceiling bounce comb filter effects, plus it can handle more power while keeping the horizontal axis baffle width minimal, which gives better imaging.
One thing I learned from experimenting a bunch is that the tweeter really needs to be at least 36 inches off the floor, and 42 inches is better. Higher than that is good too. The addition of the second 6 inch (or whatever) driver will narrow the beam vertically and thereby reduce floor and ceiling bounce comb filter effects, plus it can handle more power while keeping the horizontal axis baffle width minimal, which gives better imaging.
Looks like the new incarnation of IMP has more direct sound than the one that won the challenge which had most elements aimed towards the wall.
In the mirror can see two back elements.
How does this differ from an omni, then ?
.
Most omnis don't let user adjust ratio of direct/reflected SPLs.
Also the design allows reasonably constant off axis FR profile.
I note also, the designer tends to modify his room with absorption towards the back of the listener.
https://groups.google.com/forum/#!to...nd/ao0OsXRFezU
This is necessary as reflections are too loud as compared to direct sound in small rooms.
So the design is really a speaker AND room system.
Any speaker is a "speaker and room system", unless you live in an anechoic chamber.
Since inter-aural crosstalk seriously damages our ability to perceive imaging cues in the lower midrange (below about 1kHZ), it becomes tempting to do something about that. I've built and modified the Bob Carver "Holographic generator", which is nice at the sweet spot but damages the stereo effect when you're outside of that sweet spot. I also built a Polk style "holographic" soundbar, which works very well, but again only at the sweetspot.
Bi or Di-pole, open baffle or omni speakers in a sense recreate usable timing cue information below about 1kHZ (or wideband) by working with listening room acoustics, which although artificial (imaging cues that are not in the recording) works very good if the room acoustics allow it.
I imagine that these IMP speakers are very enjoyable. I heard the DBX Sondfield 1 speakers at Mark Davis's home and thought they were as enjoyable as any speaker I've heard. The room he had them in was pretty optimal for their design. The IMPs appear to be an update on that design concept. Nice work!
My personal preference is my open baffle speaker system, because they are di-pole everywhere except the upper midrange. I wanted to retain program embedded imaging cues in the upper midrange by minimizing room reflections between 1.5kHZ and about 7kHZ, which seem likely to cause a perceived frequency response variation between the left and right acoustic energies at the listening location. It's my opinion that in the upper midrange, the stereo effect (imaging, sense of soundstage, etc.) is a function of the match of the left and right speakers as perceived at our ears. In the upper midrange, we sense image location primarily by amplitude comparisons, rather than timing comparisons, as in the lower midrange.
So below about 1kHZ, I wanted to try and bring back a sense of depth and acoustic space, which going dipole did pretty well. Above 12kHZ I have a rear firing tweeter that adds "air" to the highest frequencies.
On axis, my open-baffle di-poles have great sound. With an open baffle speaker, you get reflections more on the Z axis (off the front wall of the listening room) than the left to right or X axis, due to off axis cancellations in the lower midrange (especially below about 600HZ) due to diffraction. This gives you significant re-creation of lower midrange timing cue information, fake as it may be, but less room interaction in the lower midrange on the X axis (left to right), where the Shroeder effect may come to life and cause boomy lower midrange sound. A more restricted radiation pattern there can be beneficial for that reason.
Having said all of that, I don't know if my speakers are necessarily any more enjoyable to listen to than your IMPs. Off axis, your IMPs may well be more enjoyable. When I sit way over by the fireplace, where I often sit in the winter months, my open baffle speakers have a rolloff of lower midrange, due to diffraction, which I''m not thrilled about.
Trying to get a speaker to interact in the best way with room acoustics is what keeps this hobby of speaker building interesting. It's all about tradeoffs, and knowing that perfect fidelity is not possible for a long list of reasons. Ultimately, for me, it's about enjoyment. So I'm not afraid to use some "fake" enhancement, if I use it strategically.
If you're curious, my OB speakers can be looked at on my website:
Bob's Website
Since inter-aural crosstalk seriously damages our ability to perceive imaging cues in the lower midrange (below about 1kHZ), it becomes tempting to do something about that. I've built and modified the Bob Carver "Holographic generator", which is nice at the sweet spot but damages the stereo effect when you're outside of that sweet spot. I also built a Polk style "holographic" soundbar, which works very well, but again only at the sweetspot.
Bi or Di-pole, open baffle or omni speakers in a sense recreate usable timing cue information below about 1kHZ (or wideband) by working with listening room acoustics, which although artificial (imaging cues that are not in the recording) works very good if the room acoustics allow it.
I imagine that these IMP speakers are very enjoyable. I heard the DBX Sondfield 1 speakers at Mark Davis's home and thought they were as enjoyable as any speaker I've heard. The room he had them in was pretty optimal for their design. The IMPs appear to be an update on that design concept. Nice work!
My personal preference is my open baffle speaker system, because they are di-pole everywhere except the upper midrange. I wanted to retain program embedded imaging cues in the upper midrange by minimizing room reflections between 1.5kHZ and about 7kHZ, which seem likely to cause a perceived frequency response variation between the left and right acoustic energies at the listening location. It's my opinion that in the upper midrange, the stereo effect (imaging, sense of soundstage, etc.) is a function of the match of the left and right speakers as perceived at our ears. In the upper midrange, we sense image location primarily by amplitude comparisons, rather than timing comparisons, as in the lower midrange.
So below about 1kHZ, I wanted to try and bring back a sense of depth and acoustic space, which going dipole did pretty well. Above 12kHZ I have a rear firing tweeter that adds "air" to the highest frequencies.
On axis, my open-baffle di-poles have great sound. With an open baffle speaker, you get reflections more on the Z axis (off the front wall of the listening room) than the left to right or X axis, due to off axis cancellations in the lower midrange (especially below about 600HZ) due to diffraction. This gives you significant re-creation of lower midrange timing cue information, fake as it may be, but less room interaction in the lower midrange on the X axis (left to right), where the Shroeder effect may come to life and cause boomy lower midrange sound. A more restricted radiation pattern there can be beneficial for that reason.
Having said all of that, I don't know if my speakers are necessarily any more enjoyable to listen to than your IMPs. Off axis, your IMPs may well be more enjoyable. When I sit way over by the fireplace, where I often sit in the winter months, my open baffle speakers have a rolloff of lower midrange, due to diffraction, which I''m not thrilled about.
Trying to get a speaker to interact in the best way with room acoustics is what keeps this hobby of speaker building interesting. It's all about tradeoffs, and knowing that perfect fidelity is not possible for a long list of reasons. Ultimately, for me, it's about enjoyment. So I'm not afraid to use some "fake" enhancement, if I use it strategically.
If you're curious, my OB speakers can be looked at on my website:
Bob's Website
Which is one approach. The other "way" is to improve the quality of sound that is actually delivered into the room by the drivers; and that depends on the quality of the system overall. If this is done to a high enough standard then the acoustic qualities of the room become irrelevant, because the ear/brain has enough "clean" information to always decode what's going in, "inside" the recording - the impact of the room, subjectively, is minimal. And, the concept of "fake" enhancement seems ridiculous ...
This technique may not deliver "perfect" sound, but it will deliver convincing sound, which personally I find a lot more interesting ..
This has sort of become the IMP thread. I still don't know what's inside the IMP - is it basically a sealed box with four strips of MTM with a 2-way XO and high passed at 100Hz and the two front/side units fitted with pad dials? Is there any matrix wiring a la Nagaoka or Elias going inside? So basically any decent 5in mid woofer and dome tweeter combo can work? I am thinking maybe do this with just four good full range high sensitivity drivers like a Faital Pro 3FE25 can also work. If somebody can post a drawing of the speaker or more specifics so we can build/model that would be great as these links posted by the builder are impossible to get to either work or any useful detailed info.
The one that beat linkwitz wasn't this one.
This is a new design.
The one that beat linkwitz was based on Bose's flagship speaker.
Having been part of the study being discussed, let me point out that there really wasn't a "winner". No speaker was statistically any better than any other, they were all basically the same, which, in a very real sense was a big loss to the Orion since it should have come out on top, given its reputation, but it didn't. In terms of measurements, which I did on all three, the Behringer was clearly the best and arguably the best sounding, but the test results were not clear on that outcome.