Hehe, I'm intrigued, I think I know, but having no actual experience of FIR or FFT filters, only reading and trying to understand, I'm not sure. If it is what I think it is, it certainly "sounds" like it should be audible....
Just for clarity
The design is an meh( multi entry horn) / synergy / art syntripp etc
My understanding rightly or wrongly as it’s acting as a point source, due to the drivers location which should be coupled into the horn within an quarter wavelength but overlapping, as the horn loads/amplifies all frequencies indiscriminately at its flare rate hence brickwalling the unwanted overlap.
Hope this makes sense
Thank you very much for the feedback so far🙂
I've never seen brick wall crossovers on these synergy setups. In fact, the midrange already gets an acoustic band pass by using that midrange port. All it needs is a little help to get into the right shape.
Many of them use passive crossovers and not any of the brick wall kind if I remember correctly. Why do you think it needs brick wall crossovers?
Many of them use passive crossovers and not any of the brick wall kind if I remember correctly. Why do you think it needs brick wall crossovers?
Yes, the closer acoustic centers are between drivers with their various band-passes, the less the need for steep xovers
Steep crossovers IMO/IME, simply help with triangulation summing off-axis, for geometrically separated drivers..
Sorry for commenting past OP goals.
I was responding more to your quote "You can then use this to hear why people don't use brick wall filters for speaker crossovers."
Because I think the method you laid out is very difficult, if even possible, to make a valid comparison, against a brick-wall crossover.
The OP's case falls under the second part of my last post i think..."OK, to solve that, say you take out the crossover. You then need, at a minimum, to align the levels, time and phase, of the non-crossed-over drivers. Who's ever done that? "
IOW, a full system alignment, but without any crossovers in place, needs to be done before there is any chance of validity using the synthetic split brick-wall crossover method you describe.
I would like to see how all that works...you have a neat idea.
Last edited:
For what it is worth, the DEQX brand of DSP Processors offers serious brick wall filtering capability for whatever you need. If you need 48 to 300 dB/octave (linear phase) or 6 to 96 dB/octave (Linkwitz-Riley and Butterworth) filtering, then you can have fun with their units. Not cheap but for those who need these performance levels, you can achieve it.
What do you mean by brick wall crossover? Do you mean a crossover with a steep slope or the crossover with a step change from zero/unit gain to unit/zero gain? The simple method I described implements the latter and since it is unique all implementations must give the same result.
Why do you think it might matter where in the chain the DSP is applied? It is only slightly more complicated to route an audio stream through a DSP filter on a PC but doing it that way won't change the result. It would be a bit more flexible but would rely on the hardware being able to keep up which is likely these days but less so in earlier years.
I agree the script should include a step for relative time alignment and relative gain. It is straightforward to implement.
Why is it synthetic? It is simply applying the DSP to the audio stream and storing the result for later playback rather than immediately squirting the result out the computer to a digital to analogue converter. If the hardware is slow you have no option but to do it this way.
Not sure about the neat idea since it is how vast amounts of audio processing is done.
I was using the term brick-wall crossover to mean a conventional active crossover complete with speaker management functions (delay, level, filters, linear phase if FIR, etc) and your method I called "synthetic".
Sorry for any lack of clarity.
Synthetic .... via real time or playback, makes no difference to me.
I can see how both can send the same signal to a driver. In fact, I'd call real time a crossover....
My entire focus has been questioning how valid is it to assess the "synthetic" method properly, without actually doing a complete system alignment that you would do with a conventional x-over, only without any x-over in place.
Not very likely i think.....especially if the 'synthetic' is being done via playback due to processing/speaker management limitations...
I use a program called equalizer APO on a windows PC to control the drivers in the speakers I listen to.
I am using a very steep variable slope probably close to 1000 db/octave. To my ears this sounds much better than a less steep slope.
Beaming, lobing and cone break up can all be avoided easier with a steep slope.
IME if you can finish crossing over by about the half wave distance between the two drivers you will basically have an acoustic point source, which minimizes a lot of the inherent problems in multi driver speakers. This is much easier with steeper slopes.
I will say though that lobing and beaming are far less offensive from a distance, and some beaming might even be desirable to limit floor and ceiling bounce.
So design for your room.
Equalizer APO, REW and Rephase WOW!
I am using a very steep variable slope probably close to 1000 db/octave. To my ears this sounds much better than a less steep slope.
Beaming, lobing and cone break up can all be avoided easier with a steep slope.
IME if you can finish crossing over by about the half wave distance between the two drivers you will basically have an acoustic point source, which minimizes a lot of the inherent problems in multi driver speakers. This is much easier with steeper slopes.
I will say though that lobing and beaming are far less offensive from a distance, and some beaming might even be desirable to limit floor and ceiling bounce.
So design for your room.
Equalizer APO, REW and Rephase WOW!
Last edited:
- Status
- Not open for further replies.