Playing with tremolo to deal with a standing wave cancelation in a bass horn

[Tim Link]

I thought I'd share my latest experience in attempting to do something about a deep notch in the response of my bass horns at 85Hz. The problem is (I'm pretty sure) the mouth on this horn is too small. This results in a reverse phase backwave traveling back to the driver, and that works out to a very strong cancelation at about 85 Hz. Play a test tone at that frequency and the fundamental disappears. I just hear some distortion if I turn it up.
This is not something that can be fixed with DSP, or by stuffing the horn, which only moves the cancelation frequency. However, short burst come through loud and clear. So a couple years ago I made a narrow band filter and applied a tremolo effect to 85 Hz. That worked, but sounded mighty rough and growly. That was too much of an artifact for my ears to accept, so I shrugged it off as a nice try.

More recently, I've been pondering ways to make my own stereo upmixing software. While thinking about how it might work, and how different methods might fail, I considered the case where each speaker is playing a slightly different frequency. This would be hard for a frequency based decoder to pick up on, and it might decide they are both playing the same frequency and just going in and out of phase with each other. What would it sound like if the decoder moved the sound to the center when they were in phase, and put them back to the sides when they were out of phase? I tried it, and it sounded pretty much the same as when both speakers just played independently. However, if it moved them both to the center even when they were out of phase, it created a strong cancel and thus a very obvious beat.

That got me thinking again about the beat, or tremolo I was trying to use earlier to beat (he he) the null in my bass horn. What would happen if the two speakers beat out of phase? One would be loud when the other was quiet, switching back and forth. Would it be more of a stereo effect and less of an obvious beat? So I tried it. And yes, it's a far less growly beat, more of a stereo effect. So this is a pretty decent cheat! I can hear that 85 Hz filling in now and I don't really notice any obvious beating. One issue though is that it requires a very steep filter, on the order of 96 dB /Octave to keep the tremolo effect from leaking into other bands where it's not needed.

Any thougths on this? Anyone ever heard of a method like this being used to make a disappearing note reappear?

 

[Hayabusa]

You basically make a big phase 'jitter' that is most likely sounds worse...

 

[Tim Link]

You basically make a big phase 'jitter' that is most likely sounds worse...

I don't hear a phase issue. It just sounds like the tone that was missing is back. I think to really hear this stereo tremolo effect you have to either play only the tones that are doing it, or have a fairly large bank of tones doing it together. When a narrow band is doing it alone along with all the other frequency bands that are not, I'm hard pressed to notice it even when I'm trying to hear it. Occasionally I do wonder if I'm hearing real bass vibrato or if it's my stereo tremolo that's creating the effect.

 

[Tim Link]

I posted on this issue back in 2021 here: https://audiosciencereview.com/foru...the-mystery-of-the-missing-frequencies.22832/

I've tried many things since then, including using kHorn bass bins to cover from 60 Hz to 200Hz. That fixed the problem but they really didn't sound that good compared to my main bass horns other than that they didn't cancel at 85Hz. I thought I was hearing a lot of distortion from them, and measurements confirmed that to be the case. They may need their woofers replaced.

 

[DVDdoug]

You probably have regular-old "room mode" standing-wave cancellations. Were the speakers without the cancelation in the same locations?

Have you tested the speakers outdoors?

85Hz shouldn't be that hard (ignoring room modes) and it seems like you're going through difficult-convoluted work-arounds to fix a speaker problem that you shouldn't be having... If they are DYI speakers, maybe you should think about a rebuild... Or find some speakers that are OK at 85Hz (and lower).

If your room is the problem, bass traps may be the solution. And if like most people you don't have space for multiple large bass traps, there are membrane bass traps. Or, more subs in different-multiple locations if you have the budget and the space, etc.

 

[Tim Link]

You probably have regular-old "room mode" standing-wave cancellations. Were the speakers without the cancelation in the same locations?

Have you tested the speakers outdoors?

85Hz shouldn't be that hard (ignoring room modes) and it seems like you're going through difficult-convoluted work-arounds to fix a speaker problem that you shouldn't be having... If they are DYI speakers, maybe you should think about a rebuild... Or find some speakers that are OK at 85Hz (and lower).

If your room is the problem, bass traps may be the solution. And if like most people you don't have space for multiple large bass traps, there are membrane bass traps. Or, more subs in different-multiple locations if you have the budget and the space, etc.

The speakers have been in three different rooms now, always exhibiting the cancelation problem at 85Hz. I tested outside once and it was consistent. It's definitely a fault of the DIY horn design. They should be re-built, but that's not going to happen. Relegating them to subwoofer duty is a great idea, and adding a new horn to cover from about 60Hz on up to 200Hz or so is what I was doing with the KHorns, making my speakers into 4 ways instead of 3ways. I'm not sure what's wrong with those KHorns that's causing them to distort so much. When I figure it out I may put them back in service.

I sell bass traps so I'd love for that to be the solution to this problem. The room actually does need bass traps, but it won't help with this problem. I need to get rid of some things and maybe sell some stuff I'm not using so I can make room for bass traps, and get some cash to buy them. I sell them, but I don't get them for free.
This "difficult convoluted" work around I just found is actually easy to pull off with software, assuming you have the software. What's surprising is how well it seems to work subjectively. I'm very happy with the way the bass sounds on everything I've listened to so far. It also seems to balance out the entire presentation through the midrange. For now, I'm frankly amazed that this method does what it does for a seemingly intractable speaker design problem.

This has got me thinking about other ways that DSP might be used to do things that might not seem possible at first. I recently bought some 2" dome midrange drivers that have a sharp peak at 10 kHz. That can be filtered out with DSP, but it doesn't stop the driver from showing higher distortion as a result of overtones from lower tones also peaking as a result of that resonance. It can be tamed with a properly implemented passive filter that will work against the resonance. At first thought, this seems like a case where DSP can't work but a passive filter can. On deeper thought, I realize that it's conceivable to make a DSP correction that will calculate the resonance stimulation from lower frequencies that are below the crossover cutoff, and generate a countering signal to silence it. This could be done for all kinds of harmonic distortion. For this to work the DSP would have to be informed of the driver's distortion characteristics at different volume levels, and also know what volume level it is actually being driven to at any moment. So it would need to be informed of the volume knob's position and how that relates to how much the driver distorts at that position, and in relation to the signal level. This could be complicated to set up properly, but then very easy to use.

If you think about what a tremolo is, it's actually dividing a single frequency into two frequencies, one slightly lower and one slightly higher. To our ear, it sounds like the average of the two frequencies being amplitude modulated. As far as the speaker is concerned, the troubling frequency is not being played, and instead two adjacent frequencies on either side are replacing it. So the speaker does not cancel. It just tremolos. If the two speakers tremolo in sync, the tremolo effect is very obvious. If their tremolos are out of sync, it's not so easy to hear if it's only happening in a narrow frequency band. I've been playing around with this multi-tone generator page: https://onlinetonegenerator.com/multiple-tone-generator.html
Even tempered music scale increases each note by 2^(1/12). A half tone step is just enough that when playing pure sine tones it starts to sound like two notes and not just a tremoloing single tone. So the difference between the two shifted notes has to be less than that or you create an audible chord effect.