Hello!
A poster sent me a PM with some of questions and I thought it would be better to do the reply in public. Not only so others can read and possibly learn, but also to open up for others to comment.
Math diffusers have a near field and a far field response. Sitting too close to them will, like multidrive speakers, not allow the various reflections(virtual sources) to properly mix before they hit the ears. A semi-circle does not have anything particular to mix as it's just one single reflection. So this caution does not apply to them. A Semicircle is only providing spatial diffusion, it have no temporal diffusion attributes.
1D and 2D math diffusers sends out a bunch of different reflections that needs to mix before they hit the ear. The 3x design frequency is a good thumb rule. Though, the important thing to realise is that it depends on complexity of the diffuser. A 2D diffuser generally have much more complex response so there'll be a huge number of reflections at any point given some distance. A 1D diffuser is generally much less complex, so they typically needs a bit more distance to the listener as the consequence of being in the near field is larger. I think the 3x distance rule may be shrunk to 2x with a complex 2D diffuser, but that's a very vague guesstimate. Somewhere between 2x and 3x should be safe.
Math diffusers (with both spatial and temporal scattering) have a certain base frequency. This is the prime number used. Neither PRD's nor QRD's actually use the prime number. The physical depth is prime number minus one at most. Ie, a prime 7 PRD counts 1 to 6. So if 6 is X deep, the diffuser base frequency is (X/6)*7. A 6 inch deep PRD 7 diffuser will have a base frequency that corresponds to 7 inch depth. A prime 7 QRD sequence only use 4 as the max depth, so the base frequency is (X/4)*7. In other words, a 4 inch deep prime 7 QRD will work at the same base frequency as a 6 inch PRD. Such large "free low end" gains are only available for low prime numbers and QRD sequences. Keep in mind that the period width (each sequence) needs to be large enough to accomodate that frequency.
The poly is a different scenario. A general guideline is to keep the panel size large compared to the wavelength. As wavelength gets larger than the array, the effect of the lump of mass gets smaller. It's typically doing some good to about half the wavelength, one octave below the physical size. There are some more advanced formulas around if you want to include more details like distances from source and receiver, angle distance from edges and so forth.
Regards,
Andreas Nordestam
A poster sent me a PM with some of questions and I thought it would be better to do the reply in public. Not only so others can read and possibly learn, but also to open up for others to comment.
Math diffusers have a near field and a far field response. Sitting too close to them will, like multidrive speakers, not allow the various reflections(virtual sources) to properly mix before they hit the ears. A semi-circle does not have anything particular to mix as it's just one single reflection. So this caution does not apply to them. A Semicircle is only providing spatial diffusion, it have no temporal diffusion attributes.
1D and 2D math diffusers sends out a bunch of different reflections that needs to mix before they hit the ear. The 3x design frequency is a good thumb rule. Though, the important thing to realise is that it depends on complexity of the diffuser. A 2D diffuser generally have much more complex response so there'll be a huge number of reflections at any point given some distance. A 1D diffuser is generally much less complex, so they typically needs a bit more distance to the listener as the consequence of being in the near field is larger. I think the 3x distance rule may be shrunk to 2x with a complex 2D diffuser, but that's a very vague guesstimate. Somewhere between 2x and 3x should be safe.
Math diffusers (with both spatial and temporal scattering) have a certain base frequency. This is the prime number used. Neither PRD's nor QRD's actually use the prime number. The physical depth is prime number minus one at most. Ie, a prime 7 PRD counts 1 to 6. So if 6 is X deep, the diffuser base frequency is (X/6)*7. A 6 inch deep PRD 7 diffuser will have a base frequency that corresponds to 7 inch depth. A prime 7 QRD sequence only use 4 as the max depth, so the base frequency is (X/4)*7. In other words, a 4 inch deep prime 7 QRD will work at the same base frequency as a 6 inch PRD. Such large "free low end" gains are only available for low prime numbers and QRD sequences. Keep in mind that the period width (each sequence) needs to be large enough to accomodate that frequency.
The poly is a different scenario. A general guideline is to keep the panel size large compared to the wavelength. As wavelength gets larger than the array, the effect of the lump of mass gets smaller. It's typically doing some good to about half the wavelength, one octave below the physical size. There are some more advanced formulas around if you want to include more details like distances from source and receiver, angle distance from edges and so forth.
Regards,
Andreas Nordestam
