We have relied on the testing done by Salford University in England, and papers published by Manfred Schroeder and Peter D’Antonio (the pioneers of modern diffusion technology). Based on this research, summarized effectively in “The Master Handbook of Acoustics” by F. Alton Everest, we came to the following conclusions:
1. The quadratic-residue diffusor yields the best diffusion coefficient over the greatest frequency range of any known design. These units are mirror image by definition.
2. The number of wells and their depth affect the performance. The deeper wells and greater number of them yields the best performance over the greatest frequency range.
3. To design modules with the highest diffusion coefficient you must use high prime numbers, which generate large sized modules.
From these conclusions, it follows that the small sized diffusors popular on the market today have the lowest diffusion coefficient. The practice of placing 4 or more of these diffusors side by side creates lobing problems at certain frequencies. We always recommend the largest size module that will fit the space. The performance level of this approach has our customers grinning from ear to ear.
1) QRD's do work well. However, there are several alternative sequences which are as good. QRD's actually are NOT mirror image - by definition....close, but not quite.
2) It has been said that "in life there are no advantages, only trade-offs". That applies here. Deeper wells diffuse deeper. Thinner wells diffuse higher. But, that is really just the start and not the whole story in diffuser design. Thin wells cause problems with absorption and with low frequency performance.
3) To say: "To design modules with the highest diffusion coefficient you must use high prime numbers" ignores a massive portion of the research on diffusion today.
To create a "small" diffusor with excellent performance does take some engineering and design. But, it has been done in more than one product on the market today. Lobing is certainly an issue. There are some products that I don't like too much for that reason. But, with a real QRD design it is easily dealt with. At 4 wide it really is a tiny issue (though still can be optimized if you are a perfectionist). With larger arrays it is no problem unless you are in a stadium or concert hall in which you probably want to switch to something like a "poly".
Regards,
Jason
