Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
32.212.76.154
In Reply to: RE: Another First Reflection Trap (FRT) Adventure : Quasi-Ambiophonics posted by 51088 on October 05, 2013 at 10:06:54
OK, so somehow I'd missed this thread when it first appeared. As it happens I've been experimenting with an RFZ so I have some 4 x 6 pieces of styrofoam to play with, and after following the link in the 1.7i thread here, I had to try this.
The mid-tweet panels of my IVA's are already oriented to cancel the sidewall relection. Unfortunately the current setup geometry didn't permit me to try the ambio trick and RFZ barrier simultaneously so I decided to try them independently, RFZ barrier first:
Well, it hit the ball out of the park! The speakers are already 6' out from the wall and displaying good depth, but with the barriers in place, the remaining confusion was gone and the image just seemed to stretch back forever. This was despite the fact that according to the mirrors, the barrier was giving me a sidewall second reflection.
I couldn't compare the barrier with my QRD diffusors because I painted them yesterday and the paint is still trying, but I hope to do that in the next few days.
Next, I tried pulling them forward to make a quasi-ambio barrier. That wasn't as successful -- I heard a bit of an improvement in image width but not what I was expecting. I hadn't read your full post though at the time and I may have been sitting too far away -- 8' rather than in the near field as you describe. I'll have to experiment more systematically with both when I have the time.
Finally, I decided to compare that with a full ambio barrier:
I heard the expected widening of the sound stage. It wasn't actually super wide on the recording I listened to (Kurt Masur playing Tchaikovsky's Serenade for Strings on Teldec), but the sides of the stage were natural, as they never are in conventional stereo, which either truncates them at the speakers or provides an unnatural fuzzy widening from early sidewall reflections. But what really got me was what happened to the phantom image -- it gained definition and turned rock solid, like three-channel stereo but without the lacunae that occur between the center and the sides. (I'd then only skimmed your post and hadn't read the first paragraph, where you mention the same phenomenon -- I was delighted to see when I came back downstairs that your observations confirm mine.)
So very promising and I'm looking forward to more experimentation. I'm not sure how practical the ambio barrier is though I love the effect of the full barrier and I'll definitely experiment more when I have the time. But the RFZ barrier is far and away the best results I've had with these speakers in my 12 x 14 room, doesn't block the window like diffusers, and would be behind the speakers where you don't really see the barriers. So if this proves the best setup, I can see making a couple of Plexiglas gobos and making the installation permanent.
Very impressive, particularly since I haven't had much luck in my ambiophonics experiments in the past, either with a foam mattress barrier or with crosstalk cancellation, which always seems too tweaky. (Of course, this makes me think that planar technology is ideal for making a phased array -- and that with head tracking, you could aim the higher frequencies at the appropriate ears and then use crosstalk cancellation at the lower frequencies where the wavelength is long enough for it to work well . . . )
Follow Ups:
I didn't give up on the barrier just that I did not get to that stage of experimentation yet after having moved to the short wall for wall loading positioning schemes and Limage type setups along with moving to time aligned 3rd order LP crossovers in the reverse split setup. I guess I can just put up the boards and see what happens though nothing is really completely optimized .
On the long wall placements I settled down with 1st order equidistant and used a pair of barriers like yours but flush to the front of the MT panel. That is where the observation came from about the opposing sidewall.
Well I tried the RFZ barrier again last night with the boards somewhat further back and the results were at least as spectacular. I'm beginning to think that more is going on here than the M-T first reflection, I think it's blocking the higher frequencies from the woofers (lower will just go through and around) so that essentially I'm hearing only their first reflections. But whatever happens and it will be a few days before I finish rebuilding my computer and can look at the impulse response it's basically large room performance in a small one with the kind of lateral and z-axis localization I'm after.
Before that, I'd tried the ambio barrier again, this time from MGBert's recommended listening distance, but again with only partial effect. I'm not sure how much of that has to do with diffraction but some of it is just that I didn't have time to change the baseline after I moved closer so I was getting a hole in the middle at that distance.
I ended up listening to the sample files on the ambiophonics site with the barriers still in the RFZ position. As I said to stchelvam, the effect was rough because the speakers were at the wrong angle but I was nevertheless impressed -- there was an awesome immediacy to the sound that reminded me of a binaural recording, or the front part of one anyway.
This stuff -- nulling room acoustics and crosstalk cancellation -- really takes things to a different level.
From Josh358: "This stuff -- nulling room acoustics and crosstalk cancellation -- really takes things to a different level."
And that, sir, was the drum I was beating when I was posting about this a few years back. It really dwarfs the improvements you get from fancy cables or amps. In fact, it's worth just getting long lengths of plain zip cord to use as speaker cable until you get placement nailed.
And yes, the MMGs are probably better candidates since, according to Floyd Toole's book "Sound Reproduction: Loudspeakers and Rooms" which has an entire section (9.1.3) which he calls "An Important One Toothed Comb - A Fundamental Flaw in Stereo" which relates to the crosstalk phenomenon, the critical frequency is about 1,300 Hz. So the MMG tweeter is definitely in play there. The link to the Toole book I posted earlier seems to no longer include that section; pity.
Of course, it also has zero WAF. ;-)
MG-bert
Fortunately I have Toole's book so I'd already read the chapter. I'd recommend that anyone curious or even not curious get a copy, it's well worth it.
I tried more ambiophonics experiments this morning, this time with the recommended 20 degree angle, but couldn't get it to work very well -- the same experience I had with the MMG's. Perhaps the side-by-side arrangement of the midrange and tweeter interferes with the timing of the crosstalk cancellation? I could hear decent spread beyond the speakers but it was no wider and probably narrower than the usual image with the speakers at 60 degrees. It just didn't match the true ambio barrier. That was just plain right, and if I could get my speakers to do that without pink styrofoam touching my nose my life would be complete.
I take it you're still using your original setup at home?
Josh358: "I take it you're still using your original setup at home?"
Short answer is yes, unfortunately I don't get to listen to it as often as I'd like.
Longer answer: I have tried to do the setup shown by st.chelvam, and got pretty good results. I used my 2 FRT panels as the barrier, stiffeners touching so the total barrier was about 4" wide, and using those foam pipe insulators on the forward edge as a headrest. I had my Gunned MMGs touching the FRT panels, about 4 feet away from my ears. So the angle between the tweeters was between 10-20 degrees, which should be ideal. I was thinking, though, that the fact that the FRTs are wooden panels did the configuration a disservice, since the high frequencies bouncing off the hard surface might have had a smearing effect. So I have a plan to glue some insulation to that reflective surface (thinking the textured foam used as mattress cushions) to reduce those reflections. These reflections don't seem to be a problem with the separated configuration I wrote about initially, btw. Haven't had the opportunity to do that yet.
And besides, the separated configuration allows for a TV screen so I can control the music selection or watch a concert video. Can't do that with the barrier dead center. Although a VR reality set of goggles would make that possible... ;-)
MG-bert
" So the angle between the tweeters was between 10-20 degrees, which should be ideal."
The crosstalk effect only effective for frequencies below 4000Hz. You may get a better result with the midrange separated by 10 to 20 degrees.
STC: "The crosstalk effect only effective for frequencies below 4000Hz. You may get a better result with the midrange separated by 10 to 20 degrees."
The crossover of my MMG speaker is actually at 1,000 Hz, so the tweeter is handling those frequencies. Working out the geometry I used when I had the speakers as close together as the 4 inch wide barrier would allow, the included angle between tweeter quasi-ribbon centers was 16.6 degrees, and between woofer panel centers was 23.5 degrees. Actually a bit less because I had rotated the panels a bit to reduce phase differences between the woofer and tweeter at the listening seat.
I agree; if I sat further away from the speakers (I actually had my forehead up against the panels I used as the ambio barrier) the results may well be better, especially since the high frequency "smearing" should be directed more away from the ears. I will try that later; too much going on now to devote time to reconfiguring the system.
Which angle should govern: if getting the woofers within a 20 degree included angle means the tweeters are closer than 10 degrees included angle, is that worse than the tweeters being right at 10 degrees but the woofers being greater than 20 degrees? I just had a thought of a new barrier configuration to try, with my 2 FRT style panels touching each other in the plane of the MMG speakers and spreading out to 4" to 6" wide closer to the listener. In other words, a wedge-shaped barrier vice a standard rectangular barrier geometry, as seen from above. That SHOULD direct any smearing early reflections from the FRT panel surfaces further still from the listener. Also, that would allow me to bring the MMGs even closer together; hence the angle questions. Thank you for your interest and assistance.
MG-bert
MG-bert wrote "Which angle should govern: if getting the woofers within a 20 degree included angle means the tweeters are closer than 10 degrees included angle, is that worse than the tweeters being right at 10 degrees but the woofers being greater than 20 degrees?"
I see your point. I guess you can only tell which is correct by experimenting.
Thank you for sharing your setup.
Yes, the ambio barrier articles I've seen all mentioned absorption. But since it isn't practical and I was able to hear the effect with just those slabs of pink styrofoam I'm not going to try it.
Need something more practical! I'll give your barrier setup a more careful try, and also try the RACE setup again, although I'm stymied because even if I can get the Maggies to work I have the same problem you have -- I need to put a screen between the speakers. In a perfect world, I'd have a huge room and a motorized acoustically transparent screen, but realistically, the only way I can think of to do this would be to push the speakers together for music listening and use the laptop rather than the projector.
I think you are right about the driver differences in lateral placement f'ing up the crosstalk cancellation from working. When we first talked about it here a decade ago I was thinking of how polk did its SDS processing, which I heard, and figured that if you did that electronically it would not work with laterally splayed drivers covering the critical frequencies. I didn't say anything since I didn't want to work out the physics explicitly, I just didn't expect it to work unless the mids were left to cover the critical frequency range in its entirety. Something you can do with the Neo8.
You could also do it by biamping and using separate parameters for the midrange and tweeter. I believe ambiophonics software usually allows you to split frequency ranges, in fact, I recall reading that the optimal angular location for the woofer and tweeter are different.
Another issue that occurs to me is that the test recording would have been made with an inverse square point source in mind rather than a 1/D line source. Again, the software would allow adjustment.
That said I've had some very interesting results with the speakers in their normal position. I'd be playing more with it now if I hadn't just learned that Amazon has delayed the deliver of my new table . . .
I'm not sure if this pertains to your set-up but in MG-bert's post the Left Speaker - Right Ear (and v.v.) cancellation is the basis for Carver's Hologram Generator (C-9 as well as in other of his products). It can be engaged/disengaged by the press of a button, so I'm able to make instant comparisons. I liked the C-9 since first available in my pre-Maggie days (before 1975) when listening to Acoustic Research 3 box speakers and in many, but not all recordings it also adds to my listening satisfaction with Tympani IV-As.
I remember the Carver! I never had one myself, but I did get the demo disk. It is very much related to the electronic versions of ambiophonics, in fact, as far as I know, it's the basis of them -- I don't know if Carver invented crosstalk cancellation, but he was the first to sell a consumer audio product that does it electronically. There have also been some loudspeakers that do it by directing an out-of-phase signal to drivers on the opposite channel loudspeaker.
The ambiophonic barrier does the same thing mechanically but isn't really practical in its usual configuration because it's practically touching your nose.
Come to think of it, I haven't tried the ambiophonics software (you can get it free online) with the IVA's yet. I did try it with my MMG's without much success. You might want to give it a try, it's more advanced than what Carver could do with the tech of the time.
Really interesting that it works with your IVA's, you've inspired me to give it a try.
The deflectors you put up look to protrude a great deal ahead of the tweeters, I wonder if you tried this without the protrusion, with the deflector board put back to the edge of the frame where it does not interfere with the front output.
Also wondering about how far you need the back portion of the board to extend to get the full effect.
As to how far forward the board extends- do you leave it just short of intruding into the line of sight to the tweeter? When I experimented with this before, the results were very dependent on how much reflection you had from the opposing sidewall before you placed the deflector. If the opposing sidewalls were not uniform and reflective, placing the board ahead of the tweeters made less of a difference. So I proceeded largely with control of the backwave alone, without redirection of the front output away from the opposing side wall.
I didn't try them further back because the woofers were in the way, but I did try another angle in an attempt to minimize sidewall reflections. It didn't work as well for reasons I don't know. I think the next step is to experiment nore systematically, beginning with MGbert's ray tracing diagram, but of course I'll have to figure out how to accommodate the woofer panels.
Extension of the board behind the MT panel -- I'm guessing the more the merrier because at lower frequencies you're going to get diffraction so the barrier will be effective only when it's large compared to the wavelength. I just had them extended sufficiently to completely block the reflection of the panels in the mirror, so a few inches beyond the actual driver.
Extension of the board in front of the panel -- if you look at the photo from the listening seat, you can see that they were extending in front for the RFZ experiment they didn't occlude the drivers at all. For that purpose, flush with the front would presumably be better because you wouldn't get diffraction at the proximate edge.
In the parallax barrier experiment, I had the boards positioned so that I could see the tweeter with the proximate eye but not the distal one. One of the reasons it may not have worked as well as MGBert's setup is that he was using MMG's, which have a lower XO to the tweeter -- my barriers were only effective from 3 kHz up. Interesting about the effect of the sidewalls. Mine are both irregular and reflective, not sure how to characterize them.
Why did you give up on your barrier? I still have some other options -- Fresnel reflector (which however is a major production), polycylindrical diffusors (but again, that's a major production since to preserve the window, I'd need to bend Plexiglas), and of course the QRD's. Or I could just move to somewhere with a bigger listening room. :-|
Hello,I am using Ambiophonics method for about 5 to 6 years. I just got to know about this forum and thread through Google alert.
I use Sound Lab ESL speakers. I first discovered Ambiophonics early 2000s and experimented with a mattress. I am currently using AmbiophonicsDSP with JRiver.
MGBert approach is rather unique. Reading MGBert's original post, I suspect he did not get the full effect of crosstalk cancellation with the barrier due to the speakers spread or angle.
It is very important that the speakers should be around 20 degrees for the Ambiophonics to be successful. Even at 30 degrees the effect would not be good enough.
In MGBert's method, what is the recommended speakers angle?
Thank you.
Edits: 03/04/17 03/05/17
I don't think he said but according to the dimension on the diagram, the adjacent should be roughly 68" and the opposite 35", so that would be about +/- 27 degrees or basically the standard Blumlein equilateral triangle.
I assume the reason you want them closer in an ambiophonic setup is the HRTF? This is a practical issue for me since I have a projector and can't put my speakers at the recommended angle.
Your system must sound pretty spectacular, what with the Sound Labs and ther controlled directivity! I tried some of the sample recordings on the ambiophonics site last night with the backwave barriers in place and was surprised that I got a lot of the effect since I was listening at +/- 30 degrees. Not a 180 degree spread, but definitely outside the speakers and the sense of three dimensionality that I've heard from ambiophonics in the past, with sounds vastly in the distance or right up to the speakers depending. It was rough -- occasional buzzing and a hole in the middle -- but it was impressive nonetheless, particularly since my earlier ambio experiments with Maggies have failed.
Sorry for the delay. I am still unsure how to navigate here.
If you are using a barrier than the ideal separation should be around 10 degrees. Like in the picture below. It must not be more than 30 degrees. Unfortunately this is not practical to most.
The ambiophonics prepossessed files in the website can be listened without the barrier. That's how it should be listened. If you add the barrier than the cancellation signals would not reach you other ears and that explains the lack of soundstage.
Try again without the barrier. Alternatively, you can use any stereo recording and use it with a barrier. It requires no preprocessing or DSP. If you are using the ambiophonics plugins than do not use the barrier.
My main speakers are about one foot apart. There are no barriers. I use AmbiophonicsDSP with JRiver. I also use additional speakers for convolution. Basically, total speakers in my system are 26 units. Except for the main speakers, the rest are low priced small/micro home theater speakers.
IMO, even with the standard 2 speakers they sound so much better than stereo.
Hi STC,
Sorry if I wasn't clear but I listened to the preprocessed files without the barrier -- I used the barrier to listen only to unprocessed files. The barrier worked brilliantly even though the speakers were at the standard 60 degree angle, but I couldn't get the preprocessed files to work very well even though I tried the panels at the recommended angles and tried listening at various distances. I did get a spread beyond the speakers when I was in the sweet range, but it was actually less of a spread than what I get with the speakers at 60 degrees and no crosstalk cancellation.
I have Tympani IVA's in split configuration so the woofer panels are separate, and timed differently. However, the crossover from woofer to mid-tweeter panels is 300 Hz third order and as I understand it in the RACE algorithm the bass below 400 Hz bypasses the crosstalk cancellation. The separate panels also didn't have a deleterious effect when I used the barrier.
Maybe I should try the panels even closer together? I thought I had them at about 20 degrees or less. As others have mentioned, though, the Maggies have side-by-side mids and tweeters whereas you're using the crossoverless Soundlabs and I'm thinking that the lateral displacement of the drivers may be interfering with the crosstalk cancellation since the acoustic centers are laterally displaced by about 5":
Also am I correct that the greatest effect of the XTC is at about 1 kHz? In which case maybe I should optimize the distance for the midrange units -- will try it again using that as a parameter. Meanwhile, I ordered a copy of AmbiophonicDSP to play with.
I'm impressed that you're doing convolution! I'd love to do that but I didn't see a way to make it practical. Don't the inexpensive home theater speakers reduce the quality of the sound? I assume you got the impulse responses from the source mentioned on the Ambiophonics site, but I'm curious about what you used for amplification and DA conversion and what you're using for the convolution -- will JRiver handle that many channels? Have you written this up somewhere? I'm intensely curious -- it's something I'd love to do.
> > Also am I correct that the greatest effect of the XTC is at about 1 kHz? In which case maybe I should optimize the distance for the midrange units -- will try it again using that as a parameter. Meanwhile, I ordered a copy of AmbiophonicDSP to play with.
I forgot to address this point earlier. The barriers will do the isolation from above 1000Hz. However, the separation gradually declines from 1000Hz to 400Hz. it was mentioned in the many AES papers that you will find on the website.
Just a word of caution here. Ideally, the main speakers should be line array type. So it is possible that you are right.
I hope you get your AmbiophonicsDSP working well. Please note that the values would change if the sampling rate of the source changes. Initially, it would be tter to experiment with big ensemble classical music and only then move to other genres. The "zentrum" should be set around 2 to 3dB. With convolution speakers, I set the "space" to zero. However, your mileage may vary and set them according to your taste for non convolution approach.
And lastly, here is the link which was made earlier with 8 convolution speakers.
It is difficult to evaluate the sample as my stereo and Ambio with 12 speakers were recorded with binaural microphone and meant to be heard with headphones. The original file is inserted between for you to compare all three versions.I could have done better with the stereo recordings but it wasn't easy to move the speakers. :)
"Just a word of caution here. Ideally, the main speakers should be line array type. So it is possible that you are right."
I think I may have made a mistake by using the central axis as a guideline. At 3 kHz, a wavelength is only about 4", so relative phase loses significance. There would still be an intensity difference that in theory needs crosstalk cancellation, but I don't think it can really be done effectively at that frequency since even a 2" head shift will change the effect from cancellation to reinforcement. So I plan to try again with the midrange as the baseline, and also with the software although I'm currently building a new HTPC so it will be a few days before I can try that.
Wow, huge difference in the sample! I'm just listening on cheap earbuds now, I'll head upstairs at some point tomorrow and find my Etymotics. But it's obvious even with the cheap earbuds. No wonder you're enthusiastic. How have you positioned the convolution speakers?
"How have you positioned the convolution speakers?"
It is recommended by Ralph that the rear half of the side wall to be given priority and followed by the rear and front half.The recommendation was based on numerous research.
Since I am limited to only one IR, I placed mine at 90 to 130 degrees, covering the rear half. I use St.Cecilia concert hall's IR for 45 degrees. With my current speakers arrangement this yields a better integration over the 60 degrees IR. Over this weekend, I will be trying 90 degrees IR and see if it improves further.
While the proper way is to use all 25 IR of St.Cecilia or any other IR by placing the speakers at the actual measured location, I opted to recreate my own "wall" due to space and lack of technical knowledge for pro sound implementation.
It is addictive as playing around with the convolution creates the very best listening experience. The method works for stereo setup too.
Thanks, I was reading up on inexpensive speakers. It looks like there are some decent ones though hard to know what would work. But man, once you add all those DAC racks and amplification it wouldn't be cheap! Also, I'm not sure how I'd feed the DAW with my usual source, JRiver -- haven't looked into it but I wonder whether you'd need to run two machines.
OTOH, no reason I can't do what you did and try a simpler setup. I have some Monsoon computer speakers and I could use the motherboard audio to experiment, with the impulse response in JRiver . . .
Josh,You need not spend a fortune on the small speakers. Mine are a mix of Yamaha, Panasonics and Sony. Speakers the size of a CD (see picture). The role of the convolution speakers is to act as sound bouncing off the wall and therefore the quality, in a small room, doesnt matter. Moreover, the I use JRiver's PEQ to cut all frequencies below 200Hz and above 4500Hz.
Basically, you only need one amplifier. I was driving 4 speakers with one amplifier. Now, I am driving all 22 speakers with just 3 amplifiers. You can drive all speakers with one 100Watt Amp or better still get a Behringer or Crown to drive all the speakers as the output volume is very low.
I am using the http://www.gearbest.com/speakers/pp_39924.html amp to drive the other 8 speakers. You would probably need less than 10 Watt for all the speakers. You can see the diagram in my profile page. My total budget for all the speakers and amplifiers is below US$200 as the HT speakers were bought used.
ST
Edits: 03/07/17
Wow, at that price, you can't afford not too!
I'm definitely going to try this once I get my new HTPC up -- I have the box built, but the monitor I was planning to get is backordered so right now I'm scrambling for a substitute.
You raised an interesting point. Will forward your query to Ambiophonics founder, Mr. Glasgal. You can also email him at glasgal at ambiophonics.org.
Meanwhile, about the my convolution setup, it was a simpler version with one IR only. I would do another two or three IR in the future but frankly, this is already very good considering my room size. You just need to be creative. I also think the convolution would work with stereo setup as well.
The small HT speakers output are low and ithey don't affect the SQ. I have the actual recording of 2L Magnificat where you can compare the original and Ambio with 12 speakers. I used Roland CS10EM binaural microphone. Unfortunately, I couldn't insert the SoundCloud link with my phone. The sample is also available in my profile page.
I will try to put in the link tomorrow. It is way past bed time ....:)
ST
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: