BobH wrote:It vibrates, so is no good for recording beams (unless pulsed!).
Even pulsed, I'm not sure this is true. The vibration frequency would have to be of the same order as the inverse pulse width. In other words, if the pulse width was 20ns and the diffuser moves at a rate of a centimetre a second, it'll move 0.4 microns over the period of the exposure, while the speckle dimension is of the order of a micron or so.
Joe Farina wrote:There is no doubt whatsoever that the diffuser movement will cause no disturbance to the table or other components, with careful design. So "movement" in that sense is not the question.
Not completely true. It'll move the adjacent air. If the beam path is traversing the air surrounding the diffuser, then, as the air moves, there'll be a variation of index. This will cause a phase change and so shift the fringes. In addition there are doppler effects due to the air movement. These effects may be small enough to ignore, however, the effects may be of the same order as the speckle itself.
Joe Farina wrote:Will the path length of the object beam be increased/decreased by the small vertical displacement of the diffuser?
Joe Farina wrote:Another way of looking at it might be to forget that the glass has a frosted or etched surface, and just think of it as a plain piece of glass.
A plain piece of glass is actually quite different from a diffuser. Consider first a plain piece of glass. An object beam for display holography is a divergent field, ie the rays diverge radially outwards. . Consider a typical ray impinging on the glass at some angle, theta. Assuming the glass was absolutely of uniform thickness and both sides were parallel (surface variation of zero rms and zero degrees between the faces of the glass), then the ray will emerge displaced but parallel to the incoming ray, ie theta(in)=theta(out). If there are variations in thickness or non-parallel surfaces, then the ray will emerge at a different angle to the entrance angle. Now sin(theta(out)) = sin(theta(in))/n. Now, as the glass is raised, unless the glass were perfectly flat and perfectly uniform, variations of thickness along the face would cause the ray exiting the system to wobble at a frequency of (and an amplitude given by) the variation in thickness and the disparity from parallelism. You can get a macro model of this if you consider an extreme case of using a prism instead of a plain piece of glass, or two prisms attached by their bases, and move that vertically as you pass the beam through them. You can see that the beam will wobble as a function of the prism angle. In fact, the ray deviation is given by D = (n-1)A, where D is the angle of deviation and A the apex prism angle. Thus, if you modelled the non-uniformity of the glass thickness as a set of prisms with different apex angles, the deviation angle of the ray would vary.
Now a diffuser is a rough surface modelled by amorphous spheroids. Thus, instead of prisms, consider (at a much larger scale) that you glue several glass marbles onto a glass plate and move that vertically. You can see that the beam would wobble with a frequency dependant on the diameter of the spheroids (marbles) and their population density. This effect would be added to the nonuniformity of the glass.