@shadyJ those pics came out great. What a snazzy studio you have there!
I want to add to the discussion, since I'm the one who foisted that Geddes article onto James. I told him I would never speak to him again if he didn't include it.
What James says about the threshold problem is really important. We should be very careful about talking about diffraction as not being audible. It's really unclear how audible it is. Direct testing of diffraction by physically causing it while holding all else constant has proven beyond difficult, so what Geddes did was surmise the effect of diffraction like this on a speaker. Group Delay. He then inserted differing amounts of group delay into test material to figure out what its audibility was like. it was a crude but accurate test, which is to say, it wasn't intended to establish thresholds. It was designed to see how it behaves.
It's not a non-linear distortion, its a linear distortion that behaves like a non-linear distortion. That is, its no different than something like an uneven frequency response or polar response, but its negative audible effect increases with volume. This actually suggests that our hearing system is non-linear. That was one of the interesting conclusions of the Geddes study, it supports one of the more current view on the human hearing model. Don't treat what he did like you would THD, it isn't that kind of distortion. It is not possible to measure it in a room. You can't take a measurement and then say, Oh you have 3% diffraction distortion, unfortunately.
As for the takeaway from this paper, I'll just say that when Geddes investigated this, his feeling was that diffraction is quite audible and should be avoided at all costs. He designed his speakers and his rooms to be minimal in this regard. Huge roundovers, very smooth and gradual transitions in the waveguides, extensive attention to detail in the mating of the driver to the waveguide, as well as optimization of the front area. He would give me a serious hard time every time I showed him my front stage and the speakers, TV, and equipment were all in line with each other. He was the one who really pushed me to stop doing what we all do.
How audible is all this? I don't know, Geddes study was a good start, his reasoning is sound enough, but there isn't enough research to conclude thresholds. It does seem likely that placing speakers inside a bookshelf or cabinet, ontop of a cabinet, against a wall, etc. are all more common problems that cause far more severe issues.
In fact, I think placing a speaker near a wall requires a special consideration. This causes a kind of acoustic mirror where by the output of the speaker is smeared in the time domain due to the slightly delayed nature of the reflection which is too delayed to be inaudible but too short in time to be filtered out by our hearing. This mirror causes a kind of diffusing of the sound along with a lot of reflections and diffraction effects which mess up the response. Placing a speaker near a wall is especially harmful.
James great article. Next! Let's measure just how significant each of these positions are on the response and then, maybe, test the audibility. Ok maybe not the latter bit, that would be pretty hard to do.
