Destructively Driving Modes

[amirm]

Is that the finite element analysis software about which only Stanford trained physicist may write software apps?

It is actually computation fluid dynamics. The speakers pressurize the liquid, i.e. air in this case, and the simulation then predicts the pressure at all spots in the room. The frequencies are swept and attempt is made at the most even response. This is the modeling that was done for our theater by Keith:

A more useful one but less fancy is to examine what happens at ear level only:

Here the comparison was made for one sub versus three.

The alternative to all of this is JBL Synthesis SFM which computes all of this using measurements in the room. It automatically sets the delay, level and filter for each sub.

 

[DaveC]

^That's pretty cool... can it automatically iterate to find ideal location and phase? That would be sweet... I used an engine simulation program (Ricardo) that could automatically iterate all sorts of variables withing defined bounds to fine tune intake, exhaust, camshaft design, etc... Something like this would/could be similar in a lot of ways.

 

[andy_c]

Lol!

 

[Nyal Mellor]

^That's pretty cool... can it automatically iterate to find ideal location and phase? That would be sweet... I used an engine simulation program (Ricardo) that could automatically iterate all sorts of variables withing defined bounds to fine tune intake, exhaust, camshaft design, etc... Something like this would/could be similar in a lot of ways.

My understanding (not having spoken to them directly about it) is that they are using a general purpose CFD physics program with a specific purpose written optimization model.

We're just using Boundary Element Modeling with manual optimization. It's been working well especially for multi-row home theaters and odd shaped rooms, and better than doing nothing at all. We also use it for locating two channel speakers in dedicated rooms. We can simulate sub placement, sub delays, sub levels, boundary properties and plot variance of SPL across the room. With relatively limited options for where you can put subs I've not found it overly onerous to do the iterative simulations. Normally it's more about seeing whether you need 2, 3, 4 or more subs, and where to put them. There's only a finite number of realistic scenarios when you are not using custom made subs (we are generally sticking to brands like JL or SVS for budget reasons). It's pretty easy to get past the "well that sucks" phase (as in different FR in every seat) to something with maybe +/-5dB variability from seat-to-seat. If I can get there then I'm generally happy.

Maybe Andy can write me an algorithm to automate the modeling process :b

 

[andy_c]

Maybe Andy can write me an algorithm to automate the modeling process :b

One must never presume to know the answer to a question when one's living depends on not knowing it.

 

[DaveC]

My understanding (not having spoken to them directly about it) is that they are using a general purpose CFD physics program with a specific purpose written optimization model.

We're just using Boundary Element Modeling with manual optimization. It's been working well especially for multi-row home theaters and odd shaped rooms, and better than doing nothing at all. We also use it for locating two channel speakers in dedicated rooms. We can simulate sub placement, sub delays, sub levels, boundary properties and plot variance of SPL across the room. With relatively limited options for where you can put subs I've not found it overly onerous to do the iterative simulations. Normally it's more about seeing whether you need 2, 3, 4 or more subs, and where to put them. There's only a finite number of realistic scenarios when you are not using custom made subs (we are generally sticking to brands like JL or SVS for budget reasons). It's pretty easy to get past the "well that sucks" phase (as in different FR in every seat) to something with maybe +/-5dB variability from seat-to-seat. If I can get there then I'm generally happy.

Maybe Andy can write me an algorithm to automate the modeling process :b

Nice, a comprehensive room acoustics simulator would be pretty cool, especially if it could do some automatic optimization as that would make it useful for a wider audience. There's a lot of acoustics involved in engine intake and exhaust design, in many different ways. The Ricardo engine sim program was pretty amazing, big $$$...

 

[Nyal Mellor]

Yes, maybe, I don't know. My gut tells me that's my backup option. For me, the exercise is to minimize drastic low frequency FIR filters. #becausepreringing

Why not IIR (all pass filter)? Though maybe the phase shift has to be over too narrow a frequency band for that?

 

[Don Hills]

... There's a lot of acoustics involved in engine intake and exhaust design, in many different ways. The Ricardo engine sim program was pretty amazing, big $$$...

Hah, designers have it easy these days... I used to design 2-stroke racing engine intake and exhaust systems with pencil, paper and a slide rule.

 

[dallasjustice]

Why not IIR (all pass filter)? Though maybe the phase shift has to be over too narrow a frequency band for that?

After measuring a number of sub setups with 2,3 and 4 subs in my room, the Welti 2 sub setup offers the most elegant and smoothest measurements of any other options. However, there are two ways I can improve the Welti 2 sub at midpoint, frontwall and backwall.

1. Rear sub in reverse polarity to front sub. I measure +\- 3db from 20hz to 45hz with no DSP, 1/12 per octave smoothing. This is the "Welti optimized" setup which fazenda concluded was subjectively most preferred. The measurements in my room support that conclusion.
2. 45hz low pass rear sub and 80hz low pass front sub. A multi-crossver sub array is needed in this setup due to the phase reversal from 2nd order length axial mode and choppy response up to 80hz with 2 subs in opposite polarity. DSP isn't needed to fight the phase shifts because the solo front sub measures very well; modex plates do their job from 45hz-80hz. The subwoofer array's level can be digitally reduced below 45hz to match solo front sub above 45hz.

This setup is also great because only one sub (frontwall) needs to be time/phase aligned at 80hz crossover with R/L. It's also great, because a very low rear sub crossover means that the rear sub will be impossible to locate at MLP, no matter how extreme the bass levels.

I'm not an IIR filter fan. There are phase shifts with IIR crossover and filters. IME, if one is careful to not ask too much from an FIR filter, it can't be beat. This is the whole point of the exercise for me. I want the best sub setup BEFORE I apply FIR impulse correction.

I should have the JBL 4367s setup with the above described array using Acourate for delays, crossovers and EQ next week. I hope to post some measurements then and talk about how it all sounds.

Michael.

 

[Jazzhead]

I should have the JBL 4367s setup with the above described array using Acourate for delays, crossovers and EQ next week. I hope to post some measurements then and talk about how it all sounds.

Am all ears , exactly what I've been waiting for.... The sound !!! Great thread , thanks fior taking the time .

 

[RayDunzl]

I'm not an IIR filter fan. There are phase shifts with IIR crossover and filters. IME, if one is careful to not ask too much from an FIR filter, it can't be beat. This is the whole point of the exercise for me. I want the best sub setup BEFORE I apply FIR impulse correction.

Hmm... I don't have an opinion on that, but can add this from my humble place:

AcourateDRC has the option to use (or not) IIR in addition to FIR, and splits the usage of the IIR to lower frequencies with FIR on the higher frequencies, although the IIR does gross level changes in the HF. Upper traces = FIR, lower traces, IIR.

 

[Nyal Mellor]

After measuring a number of sub setups with 2,3 and 4 subs in my room, the Welti 2 sub setup offers the most elegant and smoothest measurements of any other options. However, there are two ways I can improve the Welti 2 sub at midpoint, frontwall and backwall.

1. Rear sub in reverse polarity to front sub. I measure +\- 3db from 20hz to 45hz with no DSP, 1/12 per octave smoothing. This is the "Welti optimized" setup which fazenda concluded was subjectively most preferred. The measurements in my room support that conclusion.
2. 45hz low pass rear sub and 80hz low pass front sub. A multi-crossver sub array is needed in this setup due to the phase reversal from 2nd order length axial mode and choppy response up to 80hz with 2 subs in opposite polarity. DSP isn't needed to fight the phase shifts because the solo front sub measures very well; modex plates do their job from 45hz-80hz. The subwoofer array's level can be digitally reduced below 45hz to match solo front sub above 45hz.

This setup is also great because only one sub (frontwall) needs to be time/phase aligned at 80hz crossover with R/L. It's also great, because a very low rear sub crossover means that the rear sub will be impossible to locate at MLP, no matter how extreme the bass levels.

I'm not an IIR filter fan. There are phase shifts with IIR crossover and filters. IME, if one is careful to not ask too much from an FIR filter, it can't be beat. This is the whole point of the exercise for me. I want the best sub setup BEFORE I apply FIR impulse correction.

I should have the JBL 4367s setup with the above described array using Acourate for delays, crossovers and EQ next week. I hope to post some measurements then and talk about how it all sounds.

Michael.

It's precisely the phase shift of IIR I was suggesting you take advantage of with the all pass filter, but nevermind.

As to the rest of your plan, seems good. And maybe now you only need 2 subs you can use the Devialet D400!

 

[zydeco]

Michael

Great thread. So, basically, you’ve high-passed the main speakers and then have a fleet (of two) subwoofers that are low-passed at 80Hz each with a mono-signal. And the approach has been to get as smooth as possible response in the bass via positioning on front / back walls before smoothing via Acourate. Is this correct? And, if so, then what was the process used re. filters for the subwoofers? Did you go through the full Acourate process for linearising the drivers or just use the overall room correction function for the combined main speakers + subwoofers?