it is my favored compromise.
i actually find it easier to integrate everything with controlled directivity, narrow range, and first-order crossovers.
always experimenting, in a state of flux, and not written up anywhere.
tapped horns on bottom, mid-bass horns are ~60-200Hz, lo-mids to 1kHz, high-mids to 10k, super tweeters above that. mid horns picts see:Eleven Horns - lower midrange horn.... 140 Hertz Le Cleac'h... and MyEmia .
field coil drivers for lo-mids and mid-bass still on the drawing board using jbl 2435 and 15"s today.
trying to get back on topic--
after measuring, adjusting flat, listening, readjusting--
subjectively i tend to prefer freq resp tilted up on the bass end...
(on this or other systems i've measured).
hey sanedesign, it's not a scoop, but maybe the double bass array is a principle that could give an answer to your "crazy idea".
Absolutely. I have read that site and was thinking about the information when presenting this theory.
The idea I have presented a similar solution methodology although applied to solve a different problem. Where as the "double bass array" proposes to solve the problem of room modes at lower frequencies by "catching" the wave at the rear wall of the listening room, I have proposed to use the same concept to catch the wave off the rear of the speaker on the front wall.
Many, but not everyone of course, have agreed to the desire for at least some down tilt toward higher frequencies.
Can you share more specifics on your preferences; total lift and where, any particular frequencies where you have arrived at a preferred variation to the slope?
Perhaps you could also share which slopes you prefer applied to which type of speakers? (horn, monopole, dipole, planar etc)
Agreed, 100%. This is exactly the design philosophy of the constant directivity cornerhorn design. It removes any possibility of interference from the nearest walls, since it is packed in tightly into the corner. It also provides constant directivity, by way of matching radiation patterns in the MF and HF horns.
Where room layouts permit, when this placement option can be realized, I think it is an excellent approach. Where it can't, I use flanking subs.
Conveniently I happen to have cabinets with ports on both the front and back so I took a few fresh measurements. One is on the front just below the midbass driver, one is on the back in line with where the tweeter is near the top - both are diagonally symmetrically placed within the box and are the same length and diameter. (At the moment the box is lying on it's side so they're both actually at the same height but the front one is closer to the side wall...)
Normally they are both in use giving a box tuning of 43Hz, but if I block one or the other the box tuning shifts to around 32Hz, however the important thing is the tuning is the same whether the front or rear port is blocked, leaving the only difference the physical location. (I would add that this low tuning is a very non-optimal "droopy" response for this driver and box size, but it shouldn't matter for this test)
The speaker is in it's normal location but this time the measurement position is only about 1 metre in front of the speaker. Our old friend floor bounce is very much in evidence here at this close measuring distance at 210Hz, and the notch at about 130Hz will be either floor/ceiling 2nd order mode or boundary cancellation from the rear/side walls...
In any case, we have the following measurements. Yellow is with both ports blocked giving a baseline response as a sealed box, grey is with only the front port open, red is with only the rear port open. It's hard to see exactly what's going on due to all the boundary induced peaks and dips but it's apparent that the rear port has more output below about 45Hz, and they really start to diverge below 30Hz.
At the top of the graph I've normalized the results to the sealed box case, so purple is the relative gain from sealed box to front port only, while blue is relative gain from sealed box to rear port only.
Points of interest:
1) The purple line shows what we would expect from a bass reflex response relative to the sealed response - gain above the port tuning up to a maximum of around 4dB, starting to fall away at and below the port tuning - around 32Hz here, and falling below the sealed response at around 21Hz.
Normalized "gain" from the front port appears relatively smooth, because the front port is right next to the woofer and experiences almost the same room/boundary loading as the woofer, so any peaks and dips induced by the room tend to fall at exactly the same frequencies, and don't show up when normalized.
2) The dark blue line shows the normalized "gain" of the rear port. The first thing immediately obvious is how much increase in output there is at the low end relative to the front port. At 30Hz the increase is around 3dB, at 21Hz the difference is around 6dB. Surprisingly, there is some increase up to 45Hz - well above the port tuning, although this may be from other effects than dipole separation.
The blue line is very peaky with lots of peaks and dips - this is partly due to the fact that the different physical location of the rear port from the woofer means boundary cancellations tend to fall at different frequencies to that experienced by the woofer, thus create peaks and dips in a normalized response.
Something else apparent looking at the normalized rear port response is that all the main room modes are showing up as dips - 39Hz is the length-wise room mode, 50Hz is the width room mode, 75Hz is the floor to ceiling mode.
In the sealed response we can see that 39Hz and 50Hz are peaks at the listening position, and 75Hz is a notch (typical of those three room modes) and yet the net result of the rear port is a reduction in all three frequencies.
The front port only shows a reduction at the 39Hz length-wise room mode and no change at 50Hz or 75Hz.
I'm a little bit baffled by some of this. Some things make sense, others don't. Some unanswered questions for me:
1) Why is there so much gain at and below the box tuned frequency for the rear port - more than I expected. Does operating in dipole mode below box tuning fully explain this ? It's not a huge baffle either - 63cm x 39cm with a box depth of 30cm, and the rear port is 70cm from the nearest boundary.
2) Why do both front and rear ports show less "gain" at the room length mode of 39Hz than the frequencies on either side of this ? (Changes with impedance match to the room at the mode ?)
3) Why does the rear port show less gain at the 50Hz mode than the front port ? (Possible explanation - the front port is about 50cm closer to the side wall than the rear port due the cabinet laying on it's side, so may not be exciting the width room mode as much)
4) Why is the notch at the 75Hz floor/ceiling room mode actually worse with the rear port than without ? At this frequency the port should be in phase with the woofer, and both are at the same height. If they were out of phase dipole effect could account for a reduction in vertical mode excitation but they're not out of phase.
5) Why is the notch at 130Hz worse with the rear port than without ? Why is the port even contributing anything at this frequency when tuned to 32Hz ? If it is contributing something small, it should be in phase, and should be reducing the depth of the notch slightly, not worsening it.
Points 4 and 5 are probably moot in that the woofer is 70cm off the floor, with a low woofer both notches probably wouldn't be nearly as bad in the first place.
Other incidental observations are the yucky pipe mode resonance of the ports are clearly visible at 600-800Hz, which is something that needs taking care of...
The front facing port induces a notch as well as a peak, while the rear facing port is only inducing a peak. The notch is from out of phase radiation from the front port at ~650Hz partially cancelling the woofer on axis, while the peak in both cases is the resonance of the pipe inside the cabinet affecting the direct output of the woofer.
Anyone have any thoughts on the interactions of the rear port with room modes, or the apparent low frequency gain ?
Attachments
Last edited:
Deep null caused by reflection in the wall behind the speakers.
I have the same in my setup with my dipole speakers, a deep null at about 130Hz.
This null moves in frequency when I move the speakers from the wall behind them so it seems that it is caused by interference with the wall.
What about put some bass absorbers like tuned membrane absorbers at that wall behind the speakers, this should damp the reflecting soundwawe mostly around the tuning frequency.
Has someone tried this approach?
I have the same in my setup with my dipole speakers, a deep null at about 130Hz.
This null moves in frequency when I move the speakers from the wall behind them so it seems that it is caused by interference with the wall.
What about put some bass absorbers like tuned membrane absorbers at that wall behind the speakers, this should damp the reflecting soundwawe mostly around the tuning frequency.
Has someone tried this approach?
Seems to me that the best solution is to use a cardioid woofer and keep them reasonably far from side walls, the rear wall distance would be immaterial. This should eliminate the rear wall bounce issue. On second thought, a cardioid woofer pair, separate from the M/T, placed closer to the walls would be a better solution as long as they are low-passed such that the distance to the side walls places the 1/2 lambda path delay bounce frequency in the woofer's stop-band. Eliminates the back wall issue and minimizes the side wall issues. You'd still have the ceiling bounce issue, but that's there regardless of woofer type.
Dave
Last edited:
Double bass arrray is a great way to get uniform bass response. It just takes more drivers and space on the front and back walls.
Im very interested in reading more about flanking subs. Any links and pictures to that setup?
Wayne, after looking at your corner horns I wonder whether you've considered doing something with the backwave at floor level in a two-way design as a way of filling a floor bounce hole?
The whitepaper linked in my last post as some information on this approach in the last few pages. The link below has a description of the process, and of multiple subs in general:
You can also find a lot of information by doing a search on various discussion boards. In fact, there is discussion of it here in this thread, over the last several pages.
- Search for "Flanking Subs" on AudioRoundTable.com
- Search for "Flanking Subs" on DIYAudio.com
Here is a photo of flanking subs used with matched-directivity two-way speakers at a recent audio show. The rigid hotel walls (concrete on all six sides) are notoriously difficult, so the multisub approach is really helpful in that environment.
Last edited:
Yes, I've tried various approaches over the years. One was a two-way cornerhorn, and it sounded pretty good. But I do think the approach I have now is better, with a large format midhorn at ear level. That's always a problem, the competing priorities of having midrange at ear level while maintaining source position that is acoustically close to the corner.
Thanks Wayne!! Great links.
Here is my current setup..
As you can see Im a fan of the AEspeaker! I have many, many drivers from John. Love the TDM series woofers. The bass bins are AV15X and I can add more to flank the main speakers.
Im currently rebuilding the room so everything is going to change. Im changing my mains to IWATA-300 designs too.
Here is my current setup..
As you can see Im a fan of the AEspeaker! I have many, many drivers from John. Love the TDM series woofers. The bass bins are AV15X and I can add more to flank the main speakers.
Im currently rebuilding the room so everything is going to change. Im changing my mains to IWATA-300 designs too.
Hi DBM,
Interesting curves and I think I agree with all your assesments.
I once did a system for McIntosh with a woofer on the back side (XRT24). The Japanese distributor didn't like the concept and started postulating that the rear firing woofer couldn't work as well as a front firing one. At his office we took some curves of the system both front firing and rear firing and the differences were significant although minor, when rotating the speaker. However, when we carefully rotated the system such that the center of the front face of the woofer didn't move, the response was virtually identical between forward facing and backwards facing. Simply rotating the speaker around the cabinet center was moving the woofer by a foot and changing its interaction with the room. Rotating around the woofer face didn't change matters.
I'm guessing you are seeing the same thing, that ports at different locations make the difference (interaction with the room and its dimensions) and the actual direction of port facing doesn't matter.
Your port resonances are interesting also. They are usually there but you frequently overlook them unless you can "turn them on and off" as you have done.
David S.
Agreed, definitely. I find that in physically large boxes (like are usually required for high efficiency midwoofers), the position of the driver and the port are sometimes important because they set where internal standing waves line up inside the box. I also put a sheet of insulation on the braces, spanning the cross-section of the cabinet. It helps damp the lower midrange modes.
Of course, that doesn't matter much if you don't take care to mitigate similar problems outside the box, the standing waves that line up in the room.
Last edited:
If you look at the wavelengths involved (15 feet at 75 Hz, and 5 feet at 225Hz, for example), it gives you another way of looking at it. When the direct and reflected path length differential causes a cancellation at 225Hz, it may well be because of a half wavelength delay between the two, relative to the listening position. Or is it? Bass frequencies below 200Hz may be different. Maybe it's the distance differential relative to the nearest acoustically amplifying corner of the room, or whatever significant room boundary exists (?).
It would be interesting to see what happens when you put an adjustable digital delay on a second woofer (flank or perhaps even rear mounted if is isolated internally from the front woofer), and vary the delay time while watching for any effects on the notches at the listening position. A lot of delay might make the bass sound out of sync. If that worked well otherwise, one could consider digitally delaying the upper-band drivers to bring things perceptually back in timing sync. Just some thoughts.
In my case, I haven't actually measured it yet, but my 8 inch closed box woofer I put under my video screen (my center channel speaker) may well be filling in these notches pretty effectively, because the lower mid sounds way better to my ear (drums especially) when this 8 inch is fully enabled. Is it about vector mode like at higher frequencies or "pressure mode", and what does that mean?
It would be interesting to see what happens when you put an adjustable digital delay on a second woofer (flank or perhaps even rear mounted if is isolated internally from the front woofer), and vary the delay time while watching for any effects on the notches at the listening position. A lot of delay might make the bass sound out of sync. If that worked well otherwise, one could consider digitally delaying the upper-band drivers to bring things perceptually back in timing sync. Just some thoughts.
In my case, I haven't actually measured it yet, but my 8 inch closed box woofer I put under my video screen (my center channel speaker) may well be filling in these notches pretty effectively, because the lower mid sounds way better to my ear (drums especially) when this 8 inch is fully enabled. Is it about vector mode like at higher frequencies or "pressure mode", and what does that mean?
I quite like the idea of a woofer on both the front and back of a large 3 way system to help ease the boundary cancellation problems - essentially a built in "flanking sub". (Configured as a bipole not a dipole)
Instead of having say one 12" woofer, have 2x 10" woofers both crossed over at 250Hz, one on the front in the normal location and one at the back, and also have a baffle width that would put the baffle step frequency at the crossover frequency - around 45cm wide.
At bass frequencies both drivers work together in phase but due to their offset the front wall notch of each driver will be smoothed by the other. Baffle step is mostly taken care of since the rear woofer will help the forward output below the baffle step range but progressively less as you go above baffle step. In theory it shouldn't need any baffle step correction for the woofers.
Crossed over at 250Hz there shouldn't be too much problem with the rear facing driver muddying up the midrange, at least no worse than a flanking helper woofer crossed over at the same frequency.
The two woofers could even be braced against each other with a strut between their magnets to minimize panel vibrations, and there should be no net rocking of the cabinet with bass. I'm sure there must be a few commercial designs like that but I haven't come across them.
So in summary then,
1) Significant increase in output below the box tuning frequency of the rear port vs front port is probably due to dipole action from having the out of phase sources separated on opposite sides of the cabinet instead of next to each other ?
2) All other differences between front and rear port response at and above box tuning frequency (up to about 200Hz) can be explained by different room modal/boundary interaction due to the significantly different physical location of the port resulting in different room interaction ?
They are audible too. Not horrible or immediately obvious but I had noticed it in measurements before and done listening tests where I blocked one port, applied EQ to "correct" the bass response back to the same as when both ports are open, then listened to the midrange in particular with front or back port blocked.
With only the rear port active the midrange is quite noticeably smoother and cleaner, and it's the main reason I favour rear ports, particular on a 2 way. (Unfortunately this test cabinet was made years before this realisation, or I would have put both on the back...on the other hand it has been quite useful as a test bed for these sort of measurements) The ringing of the pipe mode resonance of the front port shows up quite clearly on CSD plots, while the rear port one doesn't (when measured anywhere in front of the cabinet) even if the speaker isn't far from the front wall.
Last edited:
Would this reduce the 80Hz to 200Hz issues similar to a "Flanking sub"?
Last edited:
- Status
- Not open for further replies.