Ambience tweeters using small BMR drivers

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I started writing about this over in the fullrange forum, threadjacking a thread about the Techtonic Elements TEBM35C10. So, to start this off, I'll copy out the rather windy text and photos I put over there to explain what I'm attempting to do.

Short version: First tests have been very positive to my ears.

__________________from the fullrange forum____________________
My living room (main floor of house) is quite small, but it's the most comfortable room in the house and where I mostly hang out. I have a 'listening area' in the basement, but it's a basement, it's dark, it's too close to the noise furnace/Air conditioning unit, and isolated away from my wife. So I've been concentrating on improving the listening situation in the living room.

The only practical way to arrange speakers in here (and that would also satisfy my wife's preference for room arrangement) is with speakers in the long wall. But in that arrangement, the speakers have to be right against the wall, particularly if the listening seats are going to have any space behind them to that wall back there. Dipoles or speakers that need space behind them to sound even passable aren't an option. So, I've taken that as a challenge to get something that fits in there and works well against the wall. Something small is a plus, both to not dominate the room visually and to keep the wife happy. And keeps me out of the competitions for speaker designs that might be the 'best there could be under any impractical conditions in a room no one would want to spend much time in' type of thing that seems pretty well covered already.

So, I've come up with miniaturized Synergy/Unity type waveguide speakers (the SmallSyns, documented elsewhere on this forum), which are great with handling energy off side walls and giving good overall performance. Those of you've who've built or heard well-implemented Synergy horns know how well they produce a sound field that doesn't seem to come out of the speakers but somehow just from an area -- it really takes some effort to convince a listener or even myself sometimes that the sound is actually coming from the two speakers. They even give a good sense of depth to the sound, at least compared to what I'd get with regular cone-and-dome speakers placed there.

But Synergy horns in a small room can get sort of like headphones in sound. The sound is super clear with precise location, etc., etc., but missing much of the impression that one is actually in a room where the music is being performed -- it's more like the front wall has a big opening and there are people playing in the next room there. Directivity has many benefits, but a few complications as well. Nice, but could be better. The against-the-wall positioning has definite advantages for bass response (helps minimize baffle step), but I don't get the ambience from delayed reflection off the front wall that other systems get by having speakers stuck out a meter or more from that wall. In the basement system, I have big diffusers set up behind and to the sides to gather delayed wavefronts and send them back to the seat with very good effect, but that's not an option in our living room.

So I was thinking that I could add a simulated delayed reflection, from the front (for compactness) through use of some spare channels in the MiniDSP 2x4HD I have. Put some drivers behind the speaker and against the wall, firing upward and to the side and feed them with time-diffused signal that is also delayed about 10 to 20msec. The precise sound of the Synergies shouldn't be bothered since the added signal will be out of the initial audible integration period, yet not so late as to sound like echo. The 2x4HD has FIR filters, so I can make the ambience drivers' phase response at HF quite scrambled for the desired diffuse effect without bothering imaging or tonality.

As further justification (I'm not sure about this part, but it's somewhere to start), I can use the arrangement to equalize the power response to match the on-axis response if the energy I feed to the ambience drivers is tipped toward frequencies where directivity of the speakers is higher. Even constant directivity conical horns like SmallSyns still have coverage that narrows at higher frequencies. I did some measurement and spreadsheet math and it appears that with my speakers that would happen if the ambience drivers covered mostly 2kHz and up, played about 5 or 6dB below the Synergy speaker levels.

So, that's where I got to the idea of some minature BMRs playing at upper mid to HF. An idea not too much different than Linkwitz does in his recent dipole speakers or that Audiokinesis (Duke LeJeune) does with his "LCS" floor-mount ambience speakers. Duke's run full-range (I think), though. Mine will have the flexibility to add further DSP delay and time diffusion so that any energy that does arrive at the seat directly from the ambience drivers shouldn't disturb any of the clarity of the SmallSyns. It will take a few more amplifier channels, but really, decent class D amplification is so cheap and small now that amps aren't an issue.

If I really can get away with doing this only at HF (it's largely HF shimmer that seems to be lacking) then maybe I can use use open baffle for the BMRs -- have them firing both up and down, rather than waste the energy from the back of the drivers. If not, I've 3D printed some small wedge-shaped enclouser that could get them playing down at least to the mid-hundreds of Hz.

BMR drivers should come in sometime this week and I've got two of the wedge shaped enclosures ready so some first preview tests shouldn't be too far off.

(I've had some experience with Dolby surround, and simulated acoustic from back channels. It wasn't a great effect then, but I'm hoping this will do better. The surround channels had the problem of being too locate-able, at least when one was walking through the room, which for me spoils the illusion even after I'm seated. Hoping that stuffing the sound sources at the front and all behind the SmallSyns will avoid that).
 
_______More from the fullrange forum______________

Getting there. The BMR drivers came in and I 3D printed a few little enclosures for them to try out, maybe later today. A printed box is about the only practical way to mount these little suckers if you plan to use the tiny mounting holes, the cutout is pretty complex and not much room for error. Of course, if you plan to use wood and construct by hand, you could just cut a 46.5mm dia hole and use silicone glue to mount the driver, but the mounting holes wouldn't be usable since the holes are on a diameter of only about 47.5mm! The screw sizes, btw, would be either M2 or #2 self-tapping type. About the same size as the screws that hold remote control units together. I got some on order from ebay.

These things are TINY, here is one in my not-particularly-large hand:
696328d1533841562-trials-little-tectonic-bmr-2-a-size-jpg

These are two enclosures I made to test it out. The black one was supposed to be a revised version (different side mounting slots), but I managed to choose the older file instead when I went to print last night
696329d1533841562-trials-little-tectonic-bmr-2-a-photo-1st-2-jpg

The plan (if this works out) is to have 4 of these per side mounted to the back of each speaker, firing upward and toward the sides. With one each, I don't expect to get much volume from them -- not sure I'll get enough with 4 each for that matter! Hoping that running them with a highpass will let me drive them a little harder.

This is what the 'revised' version looks like. Mostly just moved the mountings more toward the fronts and changed some angles/slopes to make things easier for the 3D printer.
696330d1533841562-trials-little-tectonic-bmr-2-a-screenshot-png
 
________More from the fullrange forum__________________
(in response to a comment from Patrick Bateman)

My guess is that two speakers cause relatively few reflections, all of which make their positions easy to locate, and make clues about how far away the speakers are, hurting the depth illusion. Controlled directivity cuts some early reflections and making more long-path delayed reflections. So more relections of various (and longer) delays helps disguise the fact that there are only two spot channels producing it all.

I'm also of the opinion that the directions of sounds matters a lot, even more than frequency response. Natural sound doesn't come from just two spots in front of you, if the instruments are spread out in a room you are also in. Even contrived sounds from other directions (but delayed enough to not mess up the precedence effect or get into the tonality integration time) are still more "accurate" than sounds that come only from two spots. Particularly if those spors are different for different frequency ranges.

Of course that's just my theory at the moment, likely to be modified or even change entirely in the future!
 
Last edited:
OK, all the stuff now copied over, here is where I am today -- after some printer problems (caused by me changing nozzles incorrectly), I've started to print the enclosures needed to install this. The plan is to have four mounted onto the back of each SmallSyns speakers. Here are the first four, which just finished printing. I have the print of the last four started now (takes about 12hrs for a set of four):

696719d1534014929-ambience-tweeters-using-bmr-drivers-flash-jpg

I've been listening to the single ones playing for a few days now. I have to say that I really like the effect. I get the precision of the Synergies, and the ambience of an omni, at least at higher frequencies. I can't say whether ambience to lower frequencies would be a big improvement, as these little critters don't play all that well into the hundreds of Hz for me to realistically try out.

So, I've only gone slightly over to the Dark Side. Not a full Bose effect (which always reminded me of the sound heard walking around in a drive-in movie theater during a show, about 40 years or more ago!). But little ambience tweeters give quite a lot more life to the presentation (without being shreaky), a way better depth illusion, and (paradoxically) better ability to differentiate instruments in recordings. And they aren't visible behind the speakers, so the room still doesn't look a like a hobby room!

Also, being behind the main speakers, they don't reveal their locations when I move within the room, which is a good thing.
 

Attachments

  • no flash.jpg
    no flash.jpg
    87.3 KB · Views: 3,608
Last edited:
This reminds of an article in Speaker Builder where the author said placing FR Phillips 5" in little wedges facing up and toward the wall was the most return for the effort and expense of anything he had tried, which was all the usual 80's multi-amp etc audiophile stuff. Maybe I can find it though I gave some of my collection to a nephew so maybe not.
Walking around at a drive-in? Really??
 
Administrator
Joined 2004
Paid Member
Nifty project Bill. I like the idea and have done a little of the same in a couple of systems. For ambience the good old Yamaha DSP-1 generates 4 channels of ambience from a stereo input. Two rear channels, and two front channels that go above or behind the mains. I have a Yamaha HT receiver that does the same thing with some of the same programs, plus some new ones. It's lot of fun and easy as the push of a button.

Interesting that you chose that little driver. I have a pair of those and they sure are tiny. I also have the bigger brother about 3.5" See photos below.
Dayton sells a similar flat woofer, also pictured below. I bought them for a shallow speaker project but ended up using other drivers. Nice idea using them for ambient channels. :up: Good use for these little guys.
 

Attachments

  • flat tweeters.jpg
    flat tweeters.jpg
    60 KB · Views: 226
  • flat drivers.jpg
    flat drivers.jpg
    90.2 KB · Views: 253
Bill have you measured them? I found they were extremely sensitive to reflections in the enclosure. Only denim insulation all the way around them made them smooth. I found out later in a discussion with Dennis Murphy he had the same experience.

No measurements yet. I did plan on stuffing the enclosures with denim fill (mostly because I have a huge bale of it sitting around already!)
 
________More from the fullrange forum__________________
(in response to a comment from Patrick Bateman)

My guess is that two speakers cause relatively few reflections, all of which make their positions easy to locate, and make clues about how far away the speakers are, hurting the depth illusion. Controlled directivity cuts some early reflections and making more long-path delayed reflections. So more relections of various (and longer) delays helps disguise the fact that there are only two spot channels producing it all.

I'm also of the opinion that the directions of sounds matters a lot, even more than frequency response. Natural sound doesn't come from just two spots in front of you, if the instruments are spread out in a room you are also in. Even contrived sounds from other directions (but delayed enough to not mess up the precedence effect or get into the tonality integration time) are still more "accurate" than sounds that come only from two spots. Particularly if those spors are different for different frequency ranges.

Of course that's just my theory at the moment, likely to be modified or even change entirely in the future!


What you are describing sounds like DML technology IMO as DML panels seem to disappear in a room with 3d like natural sound separation of instruments.
 
Bill have you measured them? I found they were extremely sensitive to reflections in the enclosure. Only denim insulation all the way around them made them smooth. I found out later in a discussion with Dennis Murphy he had the same experience.

I finally got around to measuring. I can't say I found the effect you described. Here are measurements I made today. These are from 15" away (I was working in a tight area). Windowing is at about 12msec, so lots of nearby reflections getting into it, still.

No Smoothing:
attachment.php

1/6th octave smoothed:
attachment.php

Not a heck of a lot of difference, there.

Here's an impulse response (the stuffed version):
attachment.php

Those are all on-axis, I didn't do any off-axis. Though just playing around they didn't change much when they were somewhat off.

Surprising how smooth they actually are. Also surprising is that the HF is rolled off, doesn't look too much like the plots in Tectonic's data sheet:
attachment.php
Sure didn't sound rolled off when I was listening (with them HPassed at 2k), but then my ears are 65yrs old so maybe I'm not hearing that very high stuff much anyway!

Not surprising is that their sensitivity is low, like driving bricks almost. Fortunately, I'm only using them at HF and for ambience so not all that much power is really needed.
 

Attachments

  • 5msec windowed.png
    5msec windowed.png
    13.8 KB · Views: 1,319
  • effect of stuffing full driven.png
    effect of stuffing full driven.png
    20.3 KB · Views: 1,323
  • effect of stuffing full driven no sm.png
    effect of stuffing full driven no sm.png
    18.4 KB · Views: 1,328
  • From datasheet.png
    From datasheet.png
    30.1 KB · Views: 1,342
Some system meas with the ambience

So, I have all the BMR drivers mounted behind the speakers and playing. Some DSP alignment today and measurements.
Here is an unsmoothed set of responses from a 25msec window. Black is without ambient tweeters, Red is with, Blue is the ambient tweeters on their own:
attachment.php

With 1/6th octave smoothing:
attachment.php
The BMR hardly shows in the overall response and its level looks about 10dB below the main this way.

But if I take out all windowing, then it looks like this:
attachment.php

Which shows the added ambience being about 5dB below, this sounds pretty good, too. It might be able to be pushed up a smidge, but I want to listen to more music types before I do that.

This is the Log-Impulse Response without:
attachment.php
And with:
attachment.php

Probably a better way to look at this is via Wavelet Spectrogram, so:
Without:
attachment.php
With:
attachment.php

I can't imagine anyone would prefer the sound without, at least based on what I've been playing so far (jazz, live classical, old rock). 'Chris Botti in Boston' is pretty amazing actually.

Probably some will consider this cheating or not being faithful to the recordings. But I see it as no different than placing speakers out into the room, which ehances room ambience also -- and who wants to listen in a totally dead room? Most recordings we use are only spot samples of a sound field, and were mastered to be heard in decent room, so this is just making the room more decent. But a lot more versatile than just speaker type or placement.
 

Attachments

  • wavelet without.png
    wavelet without.png
    85.3 KB · Views: 1,257
  • NoWindow 6th sm.png
    NoWindow 6th sm.png
    8.8 KB · Views: 1,247
  • NoWindow no sm.png
    NoWindow no sm.png
    14.5 KB · Views: 83
  • log IR with.png
    log IR with.png
    13.4 KB · Views: 1,240
  • log IR without.png
    log IR without.png
    12.9 KB · Views: 1,232
  • IR with ambience.png
    IR with ambience.png
    7.9 KB · Views: 82
  • IR without ambience.png
    IR without ambience.png
    7.4 KB · Views: 73
  • 25msec sm.png
    25msec sm.png
    8.9 KB · Views: 1,238
  • 25msec no sm.png
    25msec no sm.png
    13.9 KB · Views: 1,238
  • wavelet with.png
    wavelet with.png
    93.5 KB · Views: 1,242
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.