Anybody know any clever tests to estimate flash duration without investing in the proper kit?
Advanced Photographer use a desk fan (was that running at top speed?) but don't explain how they convert that to a guestimated duration.
I also heard of a method involving observing the histogram as you increase shutter speed on an electronic shutter.
Any other good techniques?
If you dig out your copy of the magazine, you'll be able to see and read how it was done. Basically as Garry said, photograph the fan at various shutter speeds with continuous light, then use those as a comparison for the images taken with flash.
It's actually a very accurate method, given that actual shutter speeds are extremely accurate these days and so give a very precise and relevant comparison. You have to be careful with the action of the focal plane shutter which skews the shape of the blades moving with and against the direction of travel, therefore only use one blade for comparison, exactly at the top or bottom. You'll need to shoot a dozen or more images to get that in the right position.
The dificultly comes in estimating
equivalents of flash duration, because they're not the same thing at all, hence the reason for the t.5 measurement that is supposed to translate to a conversion. The problem is, it doesn't, and t.5 flatters the performance of studio heads enormously. From all these tests I've done, I think t.5 x 2 is not a bad rule of thumb but it doesn't hold good for everything.
For example, with a hot-shoe guns at lower power settings, because of the way the IGBT circuit cuts the tail of the flash off sharply. With those, the total duration, the effective duration and the t.5 duration are almost exactly the same.
The problem with studio heads is the total duration is always very long, let's say 1/200sec for sake of argument. But within that, there is a high peak that only lasts for a much shorter time, and it's that peak which provides the visual action stopping. You could work out a formula for it, including something like the t.5 time x the height of the peak above the threshold level.
If you look carefully at the images in the magazine, and the Profoto shot in the current edition shows it quite well, check the shading of the black bar at the very top. It actually shows something similar to an oscilloscope trace. On the right, you can see the flash lighting up and building quickly to a peak, shown as the darkest area. Then it fades away more gradually down the tail.
If you enlarged that a lot, or shot that at much closer distance, you could measure at a lot things very accurately, but I'm not sure what it would tell you. T.5 is a very accurate and scientific measure, but what does it mean in terms of visual action-stopping potential, compared to shutter speeds that we can all understand and relate to? That's what the magazine tests were intended to reveal, and they revealed things that some manufacturers would prefer they didn't
