First of all, “3D” –diffusers are called “volumetric diffusers” and are not normally used in studio design:

https://www.gearslutz.com/board/studi...ml#post6844769
Secondly; the choise of diffuser type (1D vs 2D) depends on the acoustic model used. If LEDE/RFZ for example, 1D diffusers are usually a better option. 2D diffusers are not “better” just because they scatter in two planes, it really depends on the application.

https://www.gearslutz.com/board/studi...design-ok.html https://www.gearslutz.com/board/studi...-help-pls.html https://www.gearslutz.com/board/studi...ml#post8295482
Interpreting diffuser performance in not easy.

The most interesting value is the normalized diffusion coefficient and since it´s not trivial to obtain these values by measurement’s (it´s quite complicated to measure diffusers), BEM (Boundary Element Method) or other high accuracy prediction models are sometimes used to extract these figures. Also, there´s a lot of confusion as to what values are considered useful and sometimes you see graphs calming to be the diffusion coefficient but actually portrays the scattering coefficient (usually close to 1 in the effective bandwidths, unlike the normalized diffusion coefficient that theoretically cannot exceed about 0,6 and if so, only for a limited bandwidth unless nested designs). Another reason for unrealistically high diffusion coefficients is if the measured sample is small compared to depth. The sample should be at least about 3 meters wide (3,6 meters is common to use) and at least 16 times wider that the total depth of the devise. If using a smaller total sample width, the diffusion coefficient is going to get beefed but it´s because of the diffracting properties of the not only the shape of the sample, but the sample itself. Normalized vs. non-normalized values is another thing to look out for. Non-normalized values are higher at lower frequencies since these values do not take the diffracting properties of any object into consideration. For normalized values, the diffraction of the reference reflector (used as reference when measuring) is subtracted in order to give a fair comparison and only reveal the gained diffusion by the diffuser shape.

As for the difference between the diffusion coefficient and the scattering coefficient:

https://www.gearslutz.com/board/bass-...ml#post7030004
Sincerely Jens Eklund