EOS 5D + Speedlite Flash - potential bug(!)

Hi all,

<br />

While doing some 'frozen flash' images I discovered some unexpected

behavior of the EOS 5D:

<br /><br />

Some background first:

<br />

FP Flash, also known as High Speed Sync and denoted with the H(Bolt)

icon is normally used for fill-in flash in bright conditions where

your shutter speed exceeds the XSync limit (1/200 on the 5D). That way

you can keep using wire apertures to achieve a nice background blur

and still flash your subject for fill-in. It actually works by firing

a seemingly continuous beam of light for the length of time that the

shutter is open (avoiding in that way the banding that you would

observe otherwise). This feature requires a much higher energy output

than normal flash and therefore your flash range is dramatically

reduced.

<br /><br />

Well, all this bg info is to be able to explain the bug in simple terms:

<br /><br />

Even when FP Flash mode is active on the flash (H[bolt] icon active on

the flash LCD or red led lighted up on the ST-E2), the FP mode should

only activate when your shutter speed exceeds XSync speed. The

problem I found is that the 5D is activating the FP mode also when

used at XSync speed. This behaviour is unexpected and certainly

undesired, given that it will dramatically affect flash exposures. It

reduces the flash range, increases recycle times and consumes

batteries faster.

<br /><br />

To make it very concrete: If you have FP flash active on the flash and

your shutter speed is 1/200, the flash will fire in FP flash mode

(also called 'long burning'), so instead of having a typical flash

duration of 1/10000, the flash will fire for 1/200, an exposure 5x

longer!

<br /><br />

The tricky side of this bug is that there's no apparent visible

difference between normal flash and FP flash. You can only observe

this bug by photographing a rapidly moving subject, like running

water. In my case I was photographing an hourglass running and noticed

the bug when I was unable to freeze the sand grains falling, even with

the lowest flash power setting (that is supposed to have a duration of

about 1/35000 of a second)

<br /><br />

I've submitted this bug to Canon and I'm still waiting for a response.

<br /><br />

<b>It would be nice if some of you, using an EOS5D and a Speedlite could

try reproduce the problem. That way I will know if it's only an issue

in my camera. I'm using Firmware version 1.1.0</b><br /><br />

<I>The best test is to photograph a running shower with your camera in (M)anual

mode 1/200 f/whateveryoulike ISO100. Take one with FP flash activated and one

with FP flash off. They should look fairly the same (no bug) or one with frozen

droplets and one with a blured flow (bugged!)<br />

On the 580EX the FP flash button is indicated with a (bolt)H icon, on the 550EX

you activate it pushing on the (+) and (-) sign at the same time. This process

cycles between OFF-FP Flash-Second Curtain Sync. Make sure that the (bolt)H Icon

appears on the LCD</I>

 

<br />

<br />

You can see some images illustrating the problem on my website:

<br />

<A HREF="http://www.gerardmaas.net">www.gerardmaas.net</A>

<br /><br />

BTW, there's no reason for panic as the workaround is fairly simple: just turn

FP off when you don't need it. (Unknowingly it played havoc with my exposure in

recent photo shoots when I was using my Speedlites with softboxes and in that

case you really need all power you've got!)

<br /><br />

 

greetz, Gerard.

Hi Gerald, I may have missed something, I would have thought the 1/200th maximun sync speed as an approximation and that some bodies may be slightly faster or slower, what happened in your test when shooting below 1/200 with high speed sync activated?.

<p>The cited maximum X-sync speed should not be an approximation; in fact, the shutter would be capable of syncing at least marginally faster than that in most cases, since at 1/200 the shutter needs to be fully open long enough for the longest, highest-powered flash burst possible, which is certainly more than 1 ms and probably more like 2 ms with at least some flashes. So if you're firing at less than full power, the shutter is physically capable of syncing at somewhat higher speeds (but the firmware won't allow it, to be safe).</p>

 

<p>The correct behaviour is that if you're shooting at or below the cited maximum X-sync speed, the flash should fire in normal mode, regardless of whether you've enabled FP mode. If the flash is in FP mode, you should be able to set (or the camera should be able to select, depending on what mode it's in) a higher speed, in which case the flash should fire in FP mode. If the flash is not set to FP mode, the camera should not set/allow you to set a speed above maximum X-sync*. This is as the original poster expected but not as the original poster found.</p>

 

<p>I'm glad you specifically mentioned using M mode. In a mode in which you tell the camera what shutter speed to use, the shutter speed it uses is exactly what you set (to within the best of the camera's ability). In a mode in which the camera decides the appropriate shutter speed, the speed displayed is <em>not</em> necessarily the speed used; shutter speeds (and apertures) are set in finer increments than the camera can display, and it rounds off to the nearest value (full stop, half stop, or third stop, depending on the body and how you've set it up) for display. So if you were to shoot in Av and notice that the displayed shutter speed was 1/200, that doesn't mean that the <em>actual</em> shutter speed is 1/200; it simply means that the actual shutter speed is closer to 1/200 than to the next higher or lower shutter speed. That means that the camera could correctly set the flash into FP mode even though the displayed shutter speed is not higher than maximum X-sync. But your use of M rules this out as a possible explanation in your case.</p>

 

<p>I'd be happy to test this out if someone would kindly provide me with a 5D and a 580EX :-) I haven't tried this with my 20D and 420EX (or with previous bodies or EX-series flash) and since I don't typically shoot anything that would show a difference, I haven't noticed whether my equipment has the same bug or not.</p>

 

<p>When you're doing your testing, does the high-speed sync indicator in the viewfinder come on at 1/200?</p>

 

<p>*: assuming that it's a fully compatible intelligent flash such as any Canon EX-series flash unit. If you're using a dumb old flash that only uses the PC connector or the one big pin in the middle of the hot shoe, the camera doesn't even know it's there and cannot adjust its behaviour accordingly - and in this case, you <em>can</em> sync at above maximum X-sync if the flash burst is short enough and your shutter speed is only marginally above maximum X-sync.</p>

I just tested FP on my 20D (camera in manual mode, 550EX with FP turned on). It definitely does not shift in to FP mode until the shutter says 320. At 250 (max sync speed) it is still using one fast pop. At 320 you can see the blur, at 250 you cannot. Just what would be expected.

Jim,<Br>

Thanks for the test. I also tested my old trusty EOS 10D and it doesn't show any problem as well. Still looking for people with a 5D as I think it's a camera issue.

<br><Br>

-greetz, Gerard.

Gerard, I got similar as Jim earlier 20d+550, I know its not what your lookin for, sorry spelt name wrong before

 

good luck

I have a 5D and 20D with a 580EX. Tested both. 5D was same as yours, Gerard. 20D was not. It must be a body issue.

I think this is the kind of issue that will cause Canon to provide a firmware update. That could be fairly rapid if Canon has no other pending issues that they are working on for 5D firmware, but it will probably take someone notifying Canon's Chuck Westfall to get some action.

Nadine,

thanks for taking the time to do the test. So your 5D is buggy as well :-)

-gerard.

Mark,

 

I think too that it can be solved with a firmware update.

 

Any idea about how to get in touch with Chuck?

 

-kr, Gerard.

Yup--apparently. It was hard to tell since I shot a stream of water from a faucet and it didn't have separate water drops, and the difference was slight, but there WAS a difference--the spill around the drain was blurrier with HSS on while using 1/200, f5.6, ISO 100 as you specified, on the 5D. The 20D shots were both exactly the same--very sharp and "frozen". To be sure, I repeated the 5D shots with the same results.

Thanks Gerard:

 

I will give this a try with my 5D and an older 550EX.

Chuck can be contacted at cwestfall(at)cusa(dot)canon(dot)com or via the link at the bottom of this article:

 

http://digitaljournalist.org/issue0705/tech-tips.html

 

Another method of diagnosis is to set f/22, ISO 100 and use a 100mm+ focal length to photograph a wall at 10ft at night. Where X sync is being used, that should result in a correct exposure and a full pop of the flash, but HSS will give significant underexposure because of the lower effective guide number. A series of shots at increasing shutter speeds should prove conclusive.

Mark,

<br><br>

Thanks for the reference. I'll try to contact him, given that Canon support is really taking too long in getting back to me.

<br><br>

Regarding the power issue, I also did that test a few days ago and casually with conditions almost identical to the method you propose. I used a 580EX in an UmbrellaBox (a cool hybrid umbrella-softbox from Lastolite) EF 100 Macro, 1/200 f/16 ISO100. You can see the results on this link: <A HREF="http://www.gerardmaas.net/?p=36">http://www.gerardmaas.net/?p=36</A>.

<br><Br>

greetz, Gerard.

FYI

 

As expected, same results using a 550EX, at 1/200th. With FP Mode set, and using a shutter of 1/160th, the camera does not shift into FP mode.

 

Any luck contacting Chuck Westfall?

Hi all,

 

does this bug also happen on the 400D?

 

Thanks,

Michael

Hi everyone,

yes, the same bug 5D & Speedlight 550EX. We just got our camera 3 weeks ago and noticed blurry photos using our strobes in studio at 1/200 when the model moved slightly. So we did some tests using running water and here are the results:

1/30 - 1/200 without high speed sync off (EX 550) - pictures show stopped motion.

1/30 - 1/160 with high speed sync on pictures show stopped motion, however 1/200 is lost, becomes blurry all the way up to 1/8000.

There is NO high speed sync for this camera! And if Canon has no solution we will most likely return it.

We are planing on doing test with our White Lighting Strobes and will post the results. We had blurry photographs at 1/200 as we mentioned above. If 1/160 and below does not fix it, the camera is no worth buying. Too bad, Canon...<div></div>

the flash high speed off - now you can see the bubbles<div></div>

1/2000<div></div>

"There is NO high speed sync for this camera!"

 

Denisa, I think you fail to understand the nature of high speed sync (also called FP flash for "Focal Plane" flash). It operates by firing rapidly to imitate a continuous stream of light for the whole duration of the shutter opening. It is a way to sync the flash with the higher shutter speeds, as both shutter curtains are not otherwise fully open at the same time during higher speeds. This is NOT a way to stop subject movement. All cameras that have a focal plain shutter, and offer high speed sync, do this, not just Canon.

 

The "bug" here is that, normally with a Canon camera, even with HSS/FP turned on, it doesn't start engaging until the shutter speed raises above it's normal top sync speed. In the 5D we have a case where HSS/FP engages at the top sync speed of 1/200, and we don't think it should. The solution is simple, don't turn on HSS/FP until the shutter is above 1/200. It's only a nuisance, not a deal killer.

 

If you need a regular short duration flash to truly sync at high shutter speeds, you must use a camera with an in lens shutter. Other than point & shoot cameras, that limits you to some medium format systems. If you want digital with that, all it takes is mega amounts of money. Even then, you may not be able to get much above 1/750.

I believe an EOS 1D or Nikon D70 among others will X sync at very high shutter speeds with the right flash (i.e. non dedicated). With these cameras, the effective shutter at higher speeds is electronic - possible because the sensor is a CCD not a CMOS device.