I was asking this same question yesterday, because I came across some information about green houses from Arizona State University talking about diffusive glass or clear glass. The clear glass performed worse then the glass that defused the light, because if you think about the natural light, a large amount of what you are getting has been defused through various mediums, atmosphere, clouds, reflecting off other plants... Point being, a large percentage of the light is not from a single point source such as a light. But how does that directly relate to this?
As Frozen points out, you face the inverse square rule relative to energy dissipation, however, that does not actually directly apply in the case of a well designed grow area. A well designed grow are is a light trap, light may defuse based on inverse square law, however, it will still be limited in diffusion by the SIZE of the space, and the REFLECTIVE properties of the surfaces. For example, a flat white paints can reflect upwards of 97% of the light back. I would argue that if the grow space is a light trip, your light will defuse, and then hit the 97% reflective white surface and lose 3% of its energy, but the light wont actually be dissipated by passing through the space or air. So diffusion will happen, but its not a loss unless it is hitting poorly reflective surfaces, in fact, I would forward that it might even be more effective, because in the case study of the green houses, the gains in plant size were not at the top of the plant but deeper in side, the plants were bushier and fuller.
All this being said, I have not experimented with this, but I think as long as you go nuts with the none toxic flat white paints you will get better light distribution by placing the lights at a reasonable distance and letting the reflective surfaces do there thing to evenly distribute the light.
