High speed sync

[Johnf]

Bob

Thanks for the article regarding high speed sync.  I never really understood it!

John

[klindup]

A good article Bob but one point, other makes of electronic flashguns can do exactly the same thing.  For example Metz flashguns certainly can.  I believe that studio flash systems cannot.

Ken

[KeithB]

Bob, what kind of sensor did you use to capture the output of the flash?  Did you do any tests to see if the sensor was overloaded?

[Bob Atkins]

It was just a small area photodiode I had lying around. I don't know the specs on it. I dropped the illumination level down to where I was well under saturation by pointing the flash away from the sensor and just having it see some of the bounced light. At full power I also put an ND filter in the path. The signal was never more than about 25% of full photodiode output (i.e. the saturation level), so I suspect it was a faithful representation of the light output. I didn't measure the photdiode response time, but it's clearly faster than 100uS since that's what I measured for the 1/128 power flash. which is probably about right.

[KeithB]

Thanks, Bob.  I was mainly wondering whether you were using the full flash output to drive the photodiode.

[Frank Kolwicz]

This thread got me thinking about another possible way to get high speed synch: slow flash and fast shutter. Studio flash at full power, some at least, can be as slow as 1/750th of a second while shutters of modern 35mm cameras often exceed 1/4000 sec. With first curtain synch, won't a shutter speed of 1/1000 or more, if it coincides with the flash peak power or any reasonable portion of it, give uniform exposure?

I have such a setup and will test when I can, but wanted to get into the theoretical aspects, too.

Frank

[Bob Atkins]

Nope, won't work, The shutter at high speeds (1/250s and up) is just a moving slit that crosses the frame in 1/250s. Any flash pulse that's faster than 1/250s just results in the part of the frame under the slit being illuminated. You need a flash that's 1/250s or slower. I don't think most studio flashes are that slow, though I guess it's possible that maybe some are? I really don't know much about studio flash.

There can in fact be problems when using really slow flashes with normal sync. If the flash pulse is, say 1/100s long, normal sync won't give you uniform illumination if the sync speed is 1/200s. In that case the flash will stilll be on as the second curtian closes, so the illumination will vary across the frame as the second curatian moves across while the flash is still on.

In the latter case you'd have to shoot at 1/100s to get uniform illumination, noy the camera's normal 1/200s (which only applies with short duration flash pulses).

[Frank Kolwicz]

I don't get this answer, it seems to refer to a different question: I specified a flash duration of 1/750 that I remember from some spec sheet and a shutter speed that could be much faster than that. You postulate a whole different scenario to refute my speculation. http://www.keradwc.com/articles/studio_flash_duration.html provides a list of flash durations for various studio lights: note that several are below 1/1000 sec, even as low as 1/300sec.

I'll restate my question: Why would there be any difference between a light that's on for a duration longer than the shutter is open, just because it's a flash? If I make an exposure by room light by flipping the switch on, then clicking the shutter at 1/1000 sec, then flipping the switch off, I'm doing the same thing that a slow flash would do: flash comes on at first curtain synch, second curtain travels to end, some time later the flash turns off. The flash was on for the whole duration of shutter operation. It may not be on full-power during that whole time, but that's a different problem, at 1/8000 sec, at least, it probably isn't changing much.

[Bob Atkins]

For a camera with a sync speed of 1/200s, the shutter is always open for 1/200s (or longer). at faster shutter speeds a moving slit is used, but that (narrow) slit is open for 1/200s while it moves across the frame.

So unless your flash duration is longer than 1/200s, you won't get uniform flash illumination. Same would apply to flicking the room lights on and off (if you could do it fast enough).

[emanresu]

Hi Bob,

I was somewhat with Frank on this one, which means I probably didn't fully understand how the open slit works for shutter speed faster than the sync speed.  You said when an open slit is at work, it always stays open for the duration of the sync speed (1/200s).  So how does this slit achieve shutter speeds such as 1/1000s 1/4000s or even 1/8000s?  

Am I correct to say that for faster-than-sync speeds, for any fixed point on (actually a fixed line across) the film/sensor, it won't be exposed for more than the shutter speed (so the wideness of the slit varies according to the shutter speed), but as far as the entire film/sensor plane is concerned, each curtain still takes the sync speed to move across?

so for 35mm film and sync speed of 1/200s (5ms), the slit needs to be 1/1000s * (36mm/5ms) = 0.72mm for 1/1000s and 0.36mm for 1/2000s and so forth?  (36mm is the width of a 35mm film frame, assuming the shutter goes from side to side, if the shutter goes from top to bottom, then we use 24mm instead because that is the height).

so if the above is true, then when I take a photo of a helicopter, even though I am using 1/4000s or even 1/8000s, the rotor blades may still be distorted if they rotate faster than my sync speed?

[found a mistake in the numbers above, originally I said 0.72mm for 1/1000s and 1.44mm for 1/2000s... bad math...  should be 0.36mm for 1/2000s because it is directly proportional to the time duration]

[Bob Atkins]

Exactly. It's called focal plane shutter distortion and is related to rolling shutter problems with video. In neither case is the whole image captured at exactly the same moment.

In fact for the helicopter example, see this image (from wikipedia)

There's more at http://en.wikipedia.org/wiki/Focal-plane_shutter

[emanresu]

yes, thank you for finding this picture Bob.  I remember I saw this picture and read something about the distortion in the rotor blades because of the shutter speed a while back when I first started photography, but it didn't make too much sense at that time... but thanks to Bob's article on how the shutter works and Frank's questions, I can now say I understand how shutter works.

[klindup]

Bob I know why we use focal plane shutters - they allow very fast shutter speeds.  I wonder why the leaf shutter has not found favour.  Ok so it has to go in the lens rather than at the focal plane but with x synchronisation up to 1/500th and I assume it would not cause a similar problem with the helicopter shot I would have thought it had a lot going for it.  I seem to recall that Hasselblad used to sell a camera with both a focal plane shutter and a leaf shutter.

[Bob Atkins]

I think you answered your own question. Would you buy a DSLR with a maximum shutter speed of 1/500s, plus pay for the shutter to be in every lens? Maybe you could push a leaf shutter to 1/1000s with modern technology, but 1/500s is more commonly the limit.

I think 99.5% of photographers would answer "No". The other 0.5% of the market isn't big enough to make the development and marketing of such a camera a money making proposition. For those who simply must have a leaf shutter, there are still a few cameras out there that they can use.

The fact that most people have never even seen the distortion that can be created by using a fast shutter speed on a focal plane shiutter when photographing a moving object is proof enough that it's not much of a problem!

[emanresu]

just have a question and an idea here.

question: how do these super high speed cameras work?  the ones that capture the moment a bullet punctures an apple, or the video cameras that can capture the individual wing flaps of a humming bird?  or do they use some lighting tricks, i.e. using a very strong and instantaneous flash to shine the moment they want to capture and then let the rest of the exposure go dark?  if so, that would answer the question for the bullet situation, but still not for the humming bird.

idea: so with the shutter distortion, we should be able to do an experiment that by shooting a(n extremely) fast-moving object using a shutter speed faster than the sync speed, and that object would be distorted so much so it shortens or disappears in the photo.