high-speed flash sync / auto FP .. vs .. normal flash
There have been a number of questions about high-speed flash sync (HSS), and how it affects the output from your flash. There were also some questions asked about high-speed flash sync with this recent post where we tried to reverse-engineer a photo.
I decided to do a series of comparison photos, so we can actually see what happens before, at and beyond maximum flash sync speed. And we can also see what happens with high-speed flash sync. To do this, I set up very simple portrait lighting using a single speedlight and a large umbrella. A simple white paper-roll backdrop, and our model, Rachel. Here is the setup in my dining room …
The large (60″) umbrella to my left, and a small reflector to the right to add a little bit of fill. Ideally the reflector should be larger and closer – but this isn’t so much a portrait session as it is about photographing sequences of images at various shutter speeds. I stayed close to the umbrella – keeping it as close to the lens’ axis as possible – so that the light is as even as we could manage with a single speedlight setup. Of course, the reason for it only being one speedlight is that we can now observe its behavior.
For all the images in this article, the speedlight was set to give full output in manual mode. This way we aren’t bringing in the uncertainties of TTL flash. TTL flash is an automatic metering mode, and there will be variations in output if we change our composition. Manual flash is consistent and predictable. I worked at full output so that we can more clearly see some of the effect of going to high-speed flash sync.
Initially, I triggered the flashgun / speedlight with radio transmitters. PocketWizard Plus II units, to be exact. They don’t allow high-speed flash sync. So we can now see what will happen then. For the final sequences where we go to high-speed sync (HSS), I used an on-camera speedlight as a Master, to fire the Slaved speedlight mounted with the umbrella. The Master’s output was disabled, so we’re just dealing with one speedlight’s effect.
In the spirit of keeping much of the info on this site system-agnostic, we’ll look at how the Nikon D3 and the Canon 5D works.
how does the camera’s shutter work?
Before we get deeper into this, we need to cover some basics first – how the camera’s shutter works. The focal plane shutter found in D-SLRs, consist of two curtains that open and close. They open and close with a certain timing, to open the sensor / frame to light. When the shutter is tripped, the first curtain opens, revealing the sensor (or frame) to the light, and then the second curtain closes. The time interval between the first curtain opening, and the second curtain closing, is the shutter speed. It can be 1 second, or it can be 1/60 of a second, or as short as 1/4000 of a second.
Looking at the top part of the diagram, we can see the timing of the flash pulse. With normal flash sync (shown here as Low Speed flash sync), the flash is a near-instantaneous burst of light. (Around 1/2000 of a second, although this varies on the design of the flashgun, and how the flash pulse is controlled.) For description here, let’s just regard the flash burst as instantaneous. Therefore we need the entire frame open so that the entire scene / subject is revealed by the light from the flash. Otherwise, part of the scene / subject will be obscured by one of the shutter curtains. (We’ll see this in another sequence of images here.)
There will be some shutter speed which will be the highest shutter speed at which the first curtain has just stopped moving, and the 2nd curtain has not moved yet. One notch of a shutter speed over this, and you will have part of one of the shutter curtains obscuring the sensor / frame. (More about maximum flash sync speed here. It is one of the essential things we need to know about, and understand, when we deal with flash photography.)
With high-speed flash sync – a truly amazing bit of engineering – the flash’s output is released as a rapid series of light pulses. The flash is now effectively continuous light over a very short duration. Now, when we go over maximum flash sync speed, the flash is released as that short period of continuous light, and we can take our shutter speed much higher than maximum flash sync speed.
This change from a high-energy near-instantaneous burst of light (normal flash), to the short period of continuous light (high-speed flash sync), does imply a loss of effective power. It makes sense in that a lot of the light from our flash will just hit the shutter curtains, and won’t hit the actual sensor. In other words, much of the output from the flash in high-speed flash sync mode will be lost.
And this is where the rest of this article comes in …
maximum flash sync speed
So let’s look at how the flash output appear using old school radio slaves which don’t allow us to go to high-speed flash sync. This is how studio-type shoots are usually set up .. various flashes and light modifiers set up, and balanced .. and then tripped by radio slaves. Normal flash sync. In other words, no high-speed flash sync.
Here is the first sequence, using the Nikon D3. The next sequence is with the Canon 5D. I wanted to show that the behavior of normal flash sync and maximum flash sync speed is universal for all focal plane shutters found on D-SLRs.
The Nikon D3 (like most of the bigger Nikon D-SLRs) have a maximum flash sync speed of 1/250th. Below that, and if there is not much ambient light, the choice of shutter speed has NO effect on the flash. We just need the entire sensor / frame to be open to the blitz of flash. This can be at 1/8 or 1/60 or 1/125 … as long as it is at, or slower than maximum flash sync speed.
However, as we can see there, at 1/250 we start to see the edge of the one shutter curtain. This is due to propagation delay. (More about this in a short while.)
Here is the same sequence for the Canon 5D. The 5D has a maximum flash sync speed of 1/200 and we see the same effect with the higher shutter speeds obscuring the flash exposure.
btw, if for some reason you want to see slightly larger versions of those, click on the two sets of images for the larger side-by-side version.
propagation delay of the flash trigger signal
As you could see there with the sequence of images for the Nikon D3 and the Canon 5D – even though you might be working at maximum flash sync speed – you might still get the edge of a shutter curtain. This is due to something called ‘propagation delay’.
For these images we were using radio transceivers to trip the flash. (Pocket Wizard Plus II units.) As mentioned earlier, they are simple devices that just trigger the speedlights. There is no intelligence there between the camera and speedlights.
So this is where slight synchronization errors can creep in. Where we are working right now at maximum flash sync speed, we’re on the very edge of any specific camera’s capabilities. So when we trip the shutter, the camera has to fire the transmitter mounted on it; which then trips the receiver connected to the speedlights / flashes; which in turn fires those speedlights / flashes. There is a whole chain of events that take place within finite time. And that is where any slight synchronization error will show up … just like it does here with the edge of the shutter curtain showing.
This is a common-place occurrence, and not just inherent to the two cameras I used here. Any studio shooter will have experienced this problem and will know to use a shutter speed lower than maximum flash sync speed when shooting in the studio. A shutter speed like 1/125 or 1/100 or 1/60 is fine when working in the studio where ambient light levels are low. The shutter speed has no effect then on the flash output, so any shutter speed lower than maximum flash sync speed is fine in the studio.
On-location though, I do use maximum flash sync speed, even if there is the chance of propagation delay. The reality of on-location photography is that we are much much less likely to see the effect of flash exposure at the edge of the frame. For example, I mainly do portrait photography on location. My subjects are usually more centrally placed. Hence, propagation delay doesn’t affect me. I shoot at maximum flash sync speed to get the most efficiency from my flash.
high-speed flash sync (HSS) / auto FP
Okay, finally we get here. Let’s see what happens when we go past maximum flash sync speed, with HSS enabled.
Here is what the Nikons do:
(Actually, this is what pretty much every camera does that has HSS capability.)
Here is what the Canon 5D does. Very similar to the Nikon D3, with a slight quirk around the point we get to maximum flash sync speed.
The Canon 5D, and the Canon 5D mark II, both have the same quirk at maximum flash sync speed. If you are at max flash sync (1/200), with or without HSS, the output is affected differently. This is something I mentioned in an earlier post on maximum flash sync speed. It is just something to be aware of with the 5D bodies.
So let’s look at the implication of those two sequences of images.
The moment we go over maximum flash sync speed, our output from our flash drops considerably. It makes sense:
If we are in normal flash mode, then the flash is an instantaneous burst of light. We just need our entire frame / sensor to be open, for flash exposure to be consistent from edge to edge.
The moment we go into HSS mode, then the flash output is essentially continuous light. And … continuous light is affected by shutter speed choice. Think of ambient light. If we change our shutter speed, we change our exposure. This is exactly what happens with flash in HSS mode.
linear response of High-Speed Flash Sync
As mentioned, since HSS flash acts like continuous light, it should have a similar linear response So let’s see what happens when we change the aperture in relation to the change in shutter speed.
And there is the linear response, clearly to be seen.
1/500 @ f4 … 1/1000 @ f2.8 … 1/2000 @ f2
This is also obvious then why we need to be at maximum flash sync speed when working in bright light and trying to overpower the sun with flash. (At, or just below, maximum flash sync speed.)
For these examples, we had 1/250 @ f11 … which would, for ambient light only, translate to 1/2000 @ f4 as an example. Yet, here we have an equivalent of 1/2000 @ f2 when going to HSS.
According to the examples here, we lose about 2 stops in comparison, if we were considering ambient light. eg, going from 1/250 @ f11 to 1/500 @ f4
Now, when we go even higher, the linearity starts to flatten out. Instead of 1/4000 @ f1.4 it would appear that we lose about 1 third of a stop of light, and have to open up by dropping the shutter speed slightly to 1/3200 to get the same brightness. But this could also be partly due to working at wide open aperture on this lens.
Let’s see if the same linearity exists for the Canon 5D:
Indeed, there it is:
.. with the same tapering off in the linearity when we go super-high on our shutter speed. I had to open the shutter speed by a third of a stop to get the same exposure again. Although, similarly, this might be partially due to working at full aperture on the Canon 85mm f1.8
the implications of High-Speed Flash Sync
So the output drops considerably in HSS mode. This implies several things:
1. When we work in bright ambient light, then we have a sweet spot at (or just below) maximum flash sync speed, if we want the most efficiency from our flash. This is because the higher shutter speed implies a wider aperture .. and this wider aperture allows our flash to reach further, or work less hard for the same exposure.
2. If we want correct flash exposure with HSS flash, then we need to move our flash much closer to our subject if we are using flash as the dominant source of light. Or we need to be aware that our flash will be merely fill light, since the output is affected dramatically.
3. If we want to control our depth-of-field, we are much better off using neutral density filters, than going to HSS mode.
4. You can’t “overpower the sun” by going to HSS. If anything, you should not be in HSS when you are dealing with bright light. You need to do something entirely different to overpower the sun with flash. This idea that you go to a much higher shutter speed to control the available light when you use flash, is one of the biggest fallacies I’ve come across on the various photography forums. It simply does NOT work that way. The next section will show exactly why.
5. Because of the loss of effective power with HSS, you need to double up, or quadruple up on speedlights to compensate if you are shooting in bright light. Alternately, you can move the flashgun much closer, and use direct off-camera flash, as with this photograph. This is Aleona, one of our models at the flash photography workshops, leaping into the air. I wanted to completely freeze any movement, and went to a high shutter speed.
settings: 1/2000 @ f3.2 @ 400 ISO;
Manual flash at full power from an off-camera Slaved speedlight,
controlled wirelessly with the on-camera flash set to Master.
high-Speed Flash Sync – the final word
Hopefully this article will help illustrate what happens with high-speed flash sync. What I wanted to show here is that there is something very specific happening, and that it is predictable. When we are aware of what is happening, we can work with it, or around it, with a clearer understanding. Once we understand what is happening with our equipment and the technology we are using, then it becomes much easier to do what we really want to do … create inspiring photographs!