I use a computer as well. SFM is a computer algorithm NOT the use of multiple subs. Using multiple subs IS NOT patented nor could it be. HOW you set the paramaters for the subs can be, but since there is a gazillion equivalent ways to do this, a patent on any one of them is pretty useless.
YOU are confusing the issues and that's what makes you find my comments odd. There is absolutley nothing inconsistant or incorrect about any of my statements - but yours have not been all that accurate.
Duke is correct, the spatial variance reduction is global. Just because Todd only looked at the center seats says nothing about what happens outside of them. AND as I pointed out in my letter to the AES Todd's statistics were seriously flawed.
CABS is patented by Genelec. Klein + Hummel had to stop selling their DBA solution.
All SFM examples presented by Todd use multiple subs. But of course it will work with only one sub too. That's the beauty of the approach.
You're using something similar (http://www.diyaudio.com/forums/gedlee/122318-diy-waveguide-loudspeaker-kit-163.html#post2038449). I can't comment on how close it is to SFM (which is patented or in the process) but from your comments one gets the impression that there's not much room for interpretation.
Best, Markus
I think its the exact opposite, isn't it. There is a huge range of possibilities and very wide possibilities for interpretation. That's what makes a patent pointless. I know a lot about the patent process - I have over 30 - so I have a pretty good feeling for what can and cannot be effectively patented. SFM is one of those things that gets patented just to list "Patented" in a sales brochure (this can be well worth it sometimes and I may even have done that myself before
). But because there is such a wide array of possibilities, almost all of them being very close and certainly acceptable, how do you get a patent broad enough to stop any competition? You can't. Thats why I didn't bother.So I have no doubt that SFM has been or is being patented. Does that make any difference to what I do? No nothing at all. I've read what Todd does. I wouldn't want to do that.
Is that something that was verified by measurements or is it just based on statistical data from mathematically modeled sound fields?
We also should not forget what the goal is: a linear frequency response without peaks (and dips) at all listening locations. A small seat-to-seat variation is great if the frequency response is good. If the frequency response is bad all we got is the same bad sound at all seats.
I think its the exact opposite, isn't it. There is a huge range of possibilities and very wide possibilities for interpretation. That's what makes a patent pointless. I know a lot about the patent process - I have over 30 - so I have a pretty good feeling for what can and cannot be effectively patented. SFM is one of those things that gets patented just to list "Patented" in a sales brochure (this can be well worth it sometimes and I may even have done that myself before
So I have no doubt that SFM has been or is being patented. Does that make any difference to what I do? No nothing at all. I've read what Todd does. I wouldn't want to do that.
I don't want to comment too much on the patent issues because you know exactly why patents are filed by companies like Genelec or Harman. It's just to arm their legal departments. Maybe your business isn't interesting enough to be fought but the K+H example shows exactly what a patent can be used for: hurt the business of a competitor. Another current example is Nokia suing Apple.
I'm more interested why you would NOT want to do what Todd does.
I thought of a good way to explain the difference.
Lets say that we want to know the average height of 1000 people. Now if I wanted to know this with accuracy to 100 places then I would have to actually measure each and every one of those people. But if I only needed to know this to say two significant figures, then I could get this answer with a simple sample of say 20 people. I would not have to measure each one. Todd measures each one. More accurate? Sure. Necessary? Not at all. I'm happy with a dB.
Well, I would want to have the most accurate result. Measuring 1000 people translated to SFM means no additional cost because a computer does all the work for you. SFM is a brute force approach. This is exactly what computers are good at.
Looks like "cold, warm and medium" isn't always in each and every case accurate enough.
To be accurate is a theoretical concept. I see nothing wrong with that. What i learned from Geddes aproach is that we are all humans and not measument intruments to detect kosmic radiation. When you are a spaeker designer you have no choice. You have to be an expert on noise making, so "bad" are the physical measurements. What do you expect ?
A miracle out of a papercone?
A miracle out of a papercone?
Even the invention of the wheel stopped to amaze people at one point in time. So yes, I expect more than just noise from a paper cone.
I fully understand that a manufacturer has to live with the status quo but everybody else has the luxury to think outside the box.
Yay! Soon I will be able to report my own opinions regarding multiple subs. I will be reporting and asking questions in this thread, starting from post #32.
Unfortunately one of the woofers' frame is bent to hell, so it will be a bit longer than I was hoping til I can serve with listening impressions.
I'm surprised that nobody has pointed out that you CAN somewhat control individual subs' phase and XO, no matter how powered!
For phase, simply reverse the polarity for any individual driver. Really!
Given varied placements + unpredictable room behavior, a 180 deg (electrical) phase change doesn't always result in the sum-versus-cancel result one expects. While not a "perfect" phase adjustment, one setting for each driver WILL be "more correct" for smoothing response than the other.
You can also move a given sub until you find a better measuring random location. This IS a means of acoustic phase adjustment in terms of how each sub interacts with the others, the room, and you.
You can alter individual sub crossover (and phase to some extent) by simply inserting a large lowpass inductor in series with one of the subs. By experimenting with inductor values, the effect can be as subtle or dramatic as needed.
Set your initial sub-amp crossover for overlap with mains exactly as Earl prescribes, and then for each additional sub (particularly if localized by ear), adjust tonal balance as needed by trying different inductor values. An inductor also provides a way to reduce the percieved loudness of a given sub.
Inexpensive (metal-core) inductors are adequate, given the limited power the individual subs will be receiving.
Revisit the master XO and gain settings and tweak inductor(s)as needed to achieve optimum balance. With enough patience, you can get nearly the same result as you would with multiple amps. The key is to really optimize each sub as it is added; Messing with individual settings once all subs are on will just lead to madness.
So there you have it: Individual sub phase, XO, and loudness control - All with one amp!
-- Mark
Correct, to a point. The single most important aspect of multiple subs is the gain setting and this is hard to do with a single amp. Of course you could just use an L-Pad, but this gets expensive and doesn't work very well. No matter how limited your capabilities, there is always one "best" setting of the parameters. The more parameters you have the better "best" tends to be. It all depends on how low cost one is trying to get. Better to buy multiple amps and a DCX2496 than that expensive preamp or tube amp. Save money where it doesn't matter not where it matters a lot.