Hm.. So you want high bass and lower midletone coming from the side of the speaker. How will you Integrate that with midds and tweeter? Seems like a far from optimal design, with VERY alternative directivity and uneven frequensy response. I suggest dropping that idea and concentrate on wide and shallow design.
Cheers!
I would still like to hear what is clean? What kind of distortion at what level?
This looks like quite demanding spec. For example Gelelec 1237A and S360 maximum short term sine wave acoustic output on axis in half space, averaged from 100 Hz to 3 kHz @ 1 m is 118dB.
Hi, what kind of sim and thought process you used to determine the 500mm position mentioned? I'm very interested on this kind of stuff, was it from BEM sim, the whole room? It would be very interesting to get more insight on the positioning since I believe it is paramount info to know while designing a speaker system. You seem to know this is important.
The logic behind the simple simulation using one or few boundary reflections only is to evaluate the magnitude of the reflections in comparison to the direct sound. It is very logical and simple way to use interference pattern for this. Unattenuated single reflection gives "problems" of ~12dB (+6 peaks and -6 or more dips in frequency response, interference) and what bandwidth looks to be suffering the most. It also gives real time adjustment for toe in, listening distance as well as proximity of the reflection, the path length difference. I mean it is easy to experiment what the effect of the close front wall is and what it is in relation to a side wall reflection for example, or floor or ceiling. In general, how path length difference affects the interference. There is one aspect of the front wall reflection that is different to any other first reflection boundary is that it is behind the speaker. This gives it a lot longer path length difference in relation to the distance between transducer and boundary since it is effectively a round trip.
Most important things I've observed so far is that it is hard to come up with a construct that is thin enough to push the interference (effect of the reflection) to high frequencies, above where the ear is most sensitive. Also, there is no way to go below the sensitive range either without pulling the speaker off wall (+1m). This would hint that we need to eliminate or minimize the front wall reflection where the ear is most sensitive. Assuming the effects of reflection are most audible where the ear seems to be most sensitive. Another one is that the path length difference doesn't change much with toe in or listening distance, as long as the distance from transducer to the front wall is relatively short in comparison to listening distance. For this reason the magnitude and bandwidth of "problems" is similar for most listening situations as long as it is a "on-wall speaker". Very good base level information for design process.
What the simple sim does not tell is how audible the stuff is in real room, or what is audible and what is not. The logical thing is to try and minimize the amplitude of the reflection and this can be easily measured with the simple interference pattern. Dropping the relative SPL of the reflection to direct sound by roughly 10db roughly halves the amplitude of interference pattern, magnitude of problems. I have no idea when the problems get below audibility.
However, minimizing the reflection might minimize any benefits, is there? I think Toole talked favorably for the sidewall reflection, not so with the front wall one? I think I've read it has something to do that the image or envelopment or something could benefit for early side wall reflections from particular angle. I haven't read the whole book, and only bits and bobs here and there long time ago. I should read it.
Some connection from such thoughts to reality are: It looks like easily available waveguides are able to do roughly ~20db of attenuation or more to the incident angle. For bass, we can couple with the wall and have no problems. This leaves the mid way of 3 way speaker system the opportunity to get problems (and benefits?) under control with one way or another.
Last edited:
The speaker is practically front wall only for low frequencies, on wavelengths that are relatively long to the distance of a sound source to the wall 😉 And this seems to be quite short distance, 20cm couples roughly to 200Hz no problem but starts to have interference above. I'm not actually quite sure if there more to this than the basic rule of thumb of 1/4wl for point source and the interference pattern in simulator.
Cardioid is not needed if some other construct, damping or waveguide or array or something else was used instead, to reduce sound to the direction of the wall and/or from front wall towards the listener.
Last edited:
Eagerly waiting the moment for details and what comes up from the BEM sims! I have high hopes for nice last minute details and to see what the BEM stuff enables designers to do. But as eagerly waiting the big picture system the group comes up with!🙂
Well, instead of reading the book read some of this thread in ASR where Toole is referenced. While still not knowing what in the front wall reflection makes it bad I remembered something I didn't mention earlier. Even more important than magnitude of the reflection might be the frequency response of it. Like we try to achieve with any speakers with waveguides, smooth controlled directivity.
If the frequency response of possible reflection was very different than the direct signal I suppose it would represent itself in bad light. If it was similar enough to the direct sound instead maybe the hearing system doesn't detect it as bad thing at all. This, and notion from the link / Toole that side wall reflections are found to be a good thing at least for home listening (studio work was mentioned not to prefer early reflections) would suggest waveguides and heavy attenuation to front wall, and consequently to sidewall as well, wasn't necessarily the best option for on-wall high fidelity speaker.
Up to 180 degree radiating speaker flat on the front wall could be the thing. I think it would be possible to make such thing, that does not have much reflections from the front wall (no reflection point angle at all and eliminate diffraction) could work nice as long as the response was sufficiently similar to the first reflection points as towards the listening spot. Even without toe in 🙂 Shallow and flat instead of toed in narrow dispersion physically deep thingie, as Guerilla suggested, without explaining too much 🙂 SPL capability would be the next thing to figure out, and what kind of system setup would provide similar sound to walls and to listening spot.
Eliminating reflection point on the front wall would be nice, since the simple single reflection path length difference experiments show heavy interference already at few centimeters, practical minimum depth for cone drivers. BEM could simulate this kind of flat system nicely I think? Try and find structure whose depth and edge shape around doesn't cause reflection/diffraction to happen on the front wall, and uniform dispersion to important angles.
If the frequency response of possible reflection was very different than the direct signal I suppose it would represent itself in bad light. If it was similar enough to the direct sound instead maybe the hearing system doesn't detect it as bad thing at all. This, and notion from the link / Toole that side wall reflections are found to be a good thing at least for home listening (studio work was mentioned not to prefer early reflections) would suggest waveguides and heavy attenuation to front wall, and consequently to sidewall as well, wasn't necessarily the best option for on-wall high fidelity speaker.
Up to 180 degree radiating speaker flat on the front wall could be the thing. I think it would be possible to make such thing, that does not have much reflections from the front wall (no reflection point angle at all and eliminate diffraction) could work nice as long as the response was sufficiently similar to the first reflection points as towards the listening spot. Even without toe in 🙂 Shallow and flat instead of toed in narrow dispersion physically deep thingie, as Guerilla suggested, without explaining too much 🙂 SPL capability would be the next thing to figure out, and what kind of system setup would provide similar sound to walls and to listening spot.
Eliminating reflection point on the front wall would be nice, since the simple single reflection path length difference experiments show heavy interference already at few centimeters, practical minimum depth for cone drivers. BEM could simulate this kind of flat system nicely I think? Try and find structure whose depth and edge shape around doesn't cause reflection/diffraction to happen on the front wall, and uniform dispersion to important angles.
Last edited:
Trial and error, limited practical positioning options and REW room simulation of the low frequency modal pattern aiming to minimize nulls by placement. The high vertical directivity minimizes the influence of the floor and ceiling and the first reflection points are absorbed.
As I said before concentrate on 1K down.
Nor do the more complicated ones 🙂
https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.580.3611&rep=rep1&type=pdf
Have a read of the above and read the references
Another thread with good links to look at
https://www.audiosciencereview.com/forum/index.php?threads/research-on-reflections.27532/
Exacto, this is the area we can't escape with cone drivers. Only few centimeters of distance to wall ( longer path length difference ) on front wall reflection brings the interference down below the 1kHz and has it there all the way to not being on-wall speaker anymore 🙂 Better avoid the reflection alltogether, at least minimize it.
Thanks for the links! I'll check out REW room simulation as well.
Minimizing vertical reflections takes mucho effort since they happen in relatively small angle if listening distance is few meters or more. Do you have some multi driver setup like MTM? I found out from the simple VCAD corner model that on the low mids floor and ceiling reflections could interlace with quite normal speaker positioning so that the response evens out. And there was some really bad positions as well, where the dips were rather close to each other and would form very wide dip. About and octave below 1kHz (and above) though, there was no speaker position (in 3D) that could make the response even out, other than heavy attenuation towards the reflections like MTM can do. Waveguide helped but not mucho. Of course acoustic treatment would help as well. Maybe the REW room sim gives more insight, will check it out. It is important to remember vertical reflections affect any speaker and not just on-wall speakers. On wall speakers have advantage of height adjustment though, if it was wall mounted.
Last edited:
A sine wave is only a 3 dB crest factor signal, where music has a crest factor much higher than this. 12 to 20 dB is the typical range cited, depending on the content. It varies with frequency, too, and is rarely a broadband peak with music. Think about instrument solos, that sort of thing.
Assuming a listening position average level of say 85 dB - which is quite loud! - 115 dB peaks on musical content should be achievable using high quality components and a controlled directional response. Soft domes are often a limiting factor when trying to reach these sorts of level, so a compression driver is likely a better choice. Remember the room will contribute some uplift in the lower frequencies too.
Severe interference from the front wall reflection unless there was some horizontal directivity baked in as well. Although, I'm not certain what makes it sound bad, didn't read about the hearing system yet 🙂 Peaks and valleys in the frequency response results at least and fluid reported the overall response in room smoothed out with this particular distance from the front wall. listening distance and distance to sidewalls and other possible objects in the room would affect this as well.
A bit more progress during the few hours spent over the last couple of days. Opted for acousto over bempp more because I had got it to work before rather than a considered assessment. I still have a few days of testing and writing scripts for the input and output before running sets of cases for on wall configurations.
As part of the checking I set up an Infinity Primus 160 sized box to compare against the widely shown measurements of Toole. The frequency of the dip was around 300 Hz rather than 220 Hz suggesting the experiment did not have the speakers flat against the wall but a short distance away possibly to make room for the cables and/or mounting brackets. Will perform a more careful simulation of this case later or would it be better to leave it as a task for a group member learning to setup and run BEM simulations?
As part of the checking I set up an Infinity Primus 160 sized box to compare against the widely shown measurements of Toole. The frequency of the dip was around 300 Hz rather than 220 Hz suggesting the experiment did not have the speakers flat against the wall but a short distance away possibly to make room for the cables and/or mounting brackets. Will perform a more careful simulation of this case later or would it be better to leave it as a task for a group member learning to setup and run BEM simulations?
no, I want to use side woofers playing 100 to 1000 to create cardioid response, like Kii (https://www.kiiaudio.com/)