Fuji x100s Flash Sync Tests

The previous post looked at how the Fuji x100s shutter syncs with an external flash. The bottom line was that the shutter takes about 1/600 of a second to close. So any shutter speeds faster than that can’t have a fully open aperture, because the shutter is already partially closed when the exposure starts. This means some light can be lost while using a flash at high shutter speeds.

Here’s an example. Suppose you set the Fuji to a wide open aperture of f/2, and its max shutter speed of 1/4000 of a second. You have a flash connected, set to manual, at some power level. If the t.1 time of the flash is longer than 1/4000 of a second, you’re gonna lose light. A Canon 430EXII flash has a t.1 time of 1/4000 of a second when its set to 1/8 power. So as long as you’re set to 1/8 power of less, you shouldn’t lose too much light.

I decided to test this idea. At the same time, I was curious how well the Fuji would sync with the flash if I used my Pocket Wizard remotes, instead of a hard-wired connection between the camera and flash.

To test this, I put the Fuji on a tripod and pointed it at a couple of white doors. Boring, but it worked. The 430EXII flash was also on a stand, with no modifiers, other than the 1/4 cut of CTO gel I always have on this flash. The camera was set to 100 ISO. The room was dark enough that ambient light was insignificant.

The image below shows some results. The left most column of images were taken with the Pocket Wizards (Flex TT5 and Mini TT1). The flash was set to manual, 1/64 power. Going down in the column of images corresponds to increases in the shutter speed, from 1/125 up to 1/4000. By the way, the only way I could get the Pocket Wizards to work at all was to set them to basic trigger mode, which eliminates any possibility of adjusting the triggering time delays. It just a dumb, basic trigger. You can see that the camera was able to sync with the flash at shutter speeds of 1/500 and slower. Above that value, not so much.


4-case comparison small

On the other hand, when the flash was connected to the camera with a cord (next column to the right), with no Pocket Wizards, the camera and flash synced nicely all the way up to 1/4000. There is some loss of light at the highest shutter speeds, but not bad.

At a power setting of 1/8 on the flash (rightmost two columns of images), I got basically the same result. The Pocket Wizards worked fine up to about 1/500. The cord synced all the way up to 1/4000, with again some loss of light at the highest speeds. But still, this is pretty cool. I can see why people are so excited about this camera.
Comments

Fuji X100s: Synching, Shutter Speed, and the Like

After reading the great reviews of the Fuji x100s, I decided to take the leap and buy one. I’ve been getting more interested in street photography lately, and this camera seemed like a good fit. Plus, its supposed to sync at all shutter speeds, which is great for flash photography outside in bright sun. David Hobby and Zack Arias both have nice in-depth reviews.

But, things are rarely perfect. It turns out that the x100s can’t sync at f/2 unless you’re at around 1/1000 or slower on the shutter. Nice but, still, I was curious why that is. So I decided to run some tests to figure it out.

Phantom to the rescue

I used a high-speed video camera to watch the action of the shutter in relation to when the flash was fired. The setup is shown in the sketch below. The x100s sat on a table looking into the lens of a Phantom v1610 high-speed video camera, which was set to record at 40,000 frames/sec. A Cannon 430EXII flash was connected to the x100s via an OCF cord. The flash was set to manual, and aimed such that a little of its light could be seen in the video. The main purpose of the flash was just to add a timing mark to the video. Most of the light for the Phantom came from two LED flashlights (it has a really sensitive sensor). I know, a real strobist would have done a better job on the lighting, but I didn’t have a lot of time to play with it.

The video camera was set to trigger when the flash fired. So, when the shutter release was pressed on the Fuji, the flash fired, and the Phantom recorded the Fuji shutter motion. I made several recordings at various shutter speeds on the Fuji, all at f/2, no ND filter active.


Pasted Graphic 4

What happens at slow shutter speeds

I’m not all that familiar with leaf shutters. But the video showed that the shutter opens when you press the shutter release partially down. I assume that when you press the shutter release the rest of the way down to take a picture it turns on the sensor to start the exposure. The leaf shutter then closes according to the shutter speed you have dialed in. Visually it looks like a fat pair of scissors closing with the pivot point of the blades down in the lower right corner of the lens when looking into the lens.

The video below shows what happens when the Fuji is set to a shutter speed of 1/125. The two bright spots at around 2 o’clock and 7 o’clock are reflections from the two flashlights. The other bright spots are various multiple reflections. When you watch this video you’ll see a bright flash at the beginning, which is the Canon flash firing. Then quite a while later (remember, we’re framing at 40,000 frames/sec), you’ll see the shutter blades closing.

The shutter starts to close at 1/145 of a second after the flash fires and is completely closed at 1/120 of a second. So that’s consistent with the 1/125 shutter speed that the camera was set to.



If you make a longer video recording (which I didn’t post here because of file size) what you see is that the shutter then stays closed for a while. Presumably the sensor is being read during that time. Then the shutter mechanism resets and the shutter opens up again. That whole process takes about 1/6 of a second, which explains the Fuji’s max burst speed of 6 frames/sec.

How long does it take for the shutter to close?

The video above (and the other videos) showed the shutter takes 1.67ms to close, measured from the time that the shutter blades first appear to when they completely close. 1.67 ms is 1/600 of a second. If it takes 1/600 of a second to close, then you can’t have a wide-open aperture at shutter speeds faster than 1/600. So it’s a bit of a stretch when Fuji says (on page 40 of the manual) that 1/1000 sec at f/2 is available. I guess they’re rounding up.

Faster shutter speed

The video above showed that at slow shutter speeds, the flash fires, then the shutter closes at the end of the exposure. So, what happens when you ratchet the shutter speed up to somewhere around the speed that the shutter can close?

The video below shows what happens when the camera is set to a shutter speed of 1/1000. In this case, the shutter is already starting to close when the flash fires. They have to do this because the shutter takes longer than 1/1000 sec to close. I assume the sensor isn’t turned on until just before the flash fires. The time from when the flash fired to when the shutter closed was - you guessed it - 1/1000 of a second.



So in this case the shutter is closing during the entire exposure time. What that means is the effective aperture isn’t really f2. I don’t know how the exposed area changes with time as the shutter blades are closing. I suppose it could be measured from the video. But, for example, if the exposed area is cut in half halfway through the exposure, you’d be at f/2.8 at that point. That’s a subtlety I hadn’t thought about until seeing the video.

Another factor to consider is the flash duration. In these tests I set the Canon flash to 1/64 power, where its t.1 time is close to 1/10,000 of a second. In that case, the flash is done firing well before the shutter closes, so all of the flash energy would be captured in the exposure. But if it were set to full power the t.1 time increases to about 1/350th of a second. So the flash wouldn’t have finished firing before the shutter is done closing, which means some loss of light. The good news is that the t.1 time is shorter than the shutter closing time for all power levels of 1/2 or less. This is likely true for other brands of speedlites as well (but studio strobes are a different matter). So at 1/1000, and slower, shutter speeds, most strobists should be reasonably happy campers.

The fastest shutter speed

The next video shows what happens at 1/4000, the fastest shutter speed available on the Fuji. The shutter is partially closed when the flash fires, again a necessity to accommodate the 1/600 sec required to close the shutter.



Yikes, in this case the shutter is about halfway closed when the flash fires. So, even though you may have the aperture set to f/2 at 1/4000, the shutter is closed considerably before the flash fires, which means you don’t really have f/2. At that point the effective aperture was probably somewhere around f/2.8. Of course the shutter is continuing to close during the exposure, and the flash is continuing to fire. So the effective aperture will end up being higher than f/2.8, and much of the flash energy could be lost, depending on the t.1 time. For the Canon flash I used here, the t.1 time is shorter than 1/4000 for all power levels of 1/8 and lower. So if you can get by with 1/8 power on the flash, you won’t lose too much light. However, the aperture really isn’t f/2, so don’t expect the nice soft backgrounds you’d normally get from f/2.

Hope that answers some questions.
Comments

Canon 430EXII Flash Duration

Evidently Canon doesn’t publish information on the duration of their flashes, and that information seems to be hard to find on the web. So I decided to make some measurements for one of their flashes - the 430EXII.

Flash duration is usually specified by the so-called t.1 and t.5 times. t.1 is the total time that the flash intensity is above 10% of its peak value. Similarly, t.5 is the time spent above 50% of the peak intensity. These numbers are important if you’re trying to stop action in a photograph, or if you want to understand what happens when you push a camera to its sync speed or beyond.

I pointed a 430EXII at a photodiode, which was connected to an oscilloscope. Both the flash and the diode were secured so they wouldn’t move during the tests. The figure below shows the signal from the photodiode, which is proportional to the flash intensity.


Pasted Graphic 5



The 430EXII, like most small flashes, operates by charging its capacitor to a fixed voltage. When triggered, an IGBT cuts of the capacitor at some pre-determined time. If the flash is set to full (1/1) intensity, the capacitor fully discharges. That corresponds to the black line in the figure above. The intensity rises fairly fast to its peak, then slowly decays over a time of a few milliseconds (ms). When set to 1/2 power (red line), the IGBT cuts off the discharge at about 0.7 ms or so. For progressively lower power settings, the discharge is cut off earlier. The total area under any curve represents the total energy emitted by the flash. You can see that as the power level drops by factors of 2, the area under the curve is cut about in half.

The t.1 and t.5 times can be determined from these curves by measuring the time spent above 10% or 50% of the peak intensity. The measured values are shown in the table below. For most of the power settings, the t.1 value is about a factor of 3 larger than t.5. But for some power levels, like 1/2 for instance, the two values are much closer. That happens when the intensity curve is nearly a square pulse. In that limit the two values would be the same.

Power level

1/1

1/2
1/4
1/8
1/16
1/32
1/64

t.1 (ms)

2.93

0.69
0.37
0.24
0.17
0.12
0.10

t.5 (ms)

1.03

0.59
0.28
0.14
0.09
0.06
0.04

So, there are the numbers. The t.1 time varies from 0.1 ms (1/10,000 of a second) at the lowest power setting to 2.9 ms (1/340 of a second) at full power.

Comments