mcline

In response to: "I'm wondering what would be the best way to lock the exposure over multiple shots to a fixed value with Canon speedlites...", on a 20D body and possibly others(with a 580EX), just hold down and continue holding the FEL button as you take the shot. After the shot (if you held it) it will remain FEL'ed for the next shot. If you want to keep it locked over any number of shots you have to keep holding it. (Please don't sue me if you develop joint problems in your right thumb.)

Ben and Nadine, and all the other who have contributed answers,

 

Very good and helpful feedback, I'll apply this to my next shoot and continue experimenting.

 

Thanks!

To clarify, ETTL-II does not take into account subject distance if the flash is not locked straight ahead as the camera would not be able to predict the path the light would take to the subject.

 

The problems I am experiencing are not related to the flash being under powered, but rather are related to the camera using an algorithm that in easy lighting conditions would produce good results, or with distance info and ETTL-II in difficult lighting conditions produces acceptable results, but without the distance info as is the case when the flash is not pointed straight ahead, it uses an "intelligent" matrix mode without distance info or any apparent factoring based on the active focus point that blows out the white faces of distant models wearing dark dresses. <--This is not unexpected as I understand the general algorithm it uses, but there should be a way to _force_ it to partial meter the FE off the active focus point. Agreed, if accidentally the focus point misses the 18% grey surface of the face the exposure will be completely wrong, but that is at least in my control, and the focus will be completely wrong too. I believe C08 used to accomplish this on my old EOS Elan II.

 

I'm uploading a sample partial metered off the face using AV and FEC of minus 1.333. (The astute observer will note this is using a side-mounted diffuse master and top-mounted direct slave flash in A:B mode set to a ratio of about 1:1, another factor that disables distance information being used by the camera, but not the "problem" in this case. You will note it produces pleasant shadows and lighting). I unfortunately deleted all of my extreme examples, but this one was close enough to be usable but still illustrates the point. If she had been further away (or had a significantly smaller face), her face would have been even more overexposed. While in this specific example she is stationary, many of the models were walking along the runway towards me slightly beyond the zoom range of the lens, meaning as they get closer their faces get bigger, and with dark dresses the FEC needs to be dialed up slightly since the camera's flash evaluates the entire scene rather than partial metering off just the face. Also, in several cases there were two or even occasionally 3 models walking at the same time, not always wearing consistently light or dark dresses, meaning the FEC needed some adjustment for each shot, which is near impossible to keep up with.

 

By the way, the comment about the spot lights is dead on as well, these caused a few of my pictures to be overexposed while I was shooting M exposure, but that is understandable and avoidable with slightly less ambient exposure and an accurate fill flash for the darker areas or clicking up the shutter speed when necessary. Catching others flashes killed a few as well, but that is also understandable.

 

I've experimented with CF14=1 but found while it makes it more predictable this just contributes to the problem.

 

Does anyone know why Canon does not allow FE to be tied to the FP or even support partial FE metering without using FEL on the 20D?<div></div>

In difficult lighting situations such as a walking white model

wearing a black dress in a large darkly painted room, I have been

frustrated to no end with trying to get even remotely correct flash

exposure for skin tones using a 20D and 580EX using non-direct

flash. (The example does not need to be as extreme as above to be

problematic, and the above is not uncommon)

 

Surprisingly, with slight FEC the 20D handles the above situation

reasonably well with ETTL-II and direct flash, but when bouncing or

using any sort of attachment which forces standard ETTL, it misses

(in this example on the high side) by several full stops. Using

only direct flash simply is not an acceptable option unfortunately.

 

I would like to track the model's face using one of the focusing

points other than the center and have the flash partial meter off of

this point. I don't mind some FEC corrections, but want the camera

to at least get it close when I'm tracking a perfect metering

surface. I would prefer to set the camera to manual exposure and

rely on the flash to bring up the exposure to just the right amount

based on the skin under my active focus point. It doesn't sound

like too much to ask for.

 

Two problems I have run into is that first I can't find any way to

force the 20D to partial meter the flash based on the selected

focusing point or even key the FEL off of a non-center point, and

second, because the models are in constant motion, using FEL even if

it could be keyed to an off-center focusing point would be very

difficult.

 

Questions: Is there no way to partial meter the flash on the 20D

based on the active focus point? Other than FEL is there any way to

partial meter the flash using the center area? Does the 1DmkII or

any of the newer canon digitals such as the 5D or 30D cameras

support this? If anyone else here regularly shoots runway models,

what method have you found to get consistent flash exposure if not

firing the flash directly?

I have extensivly used the Canon 10-22 with a canon 20D and am extremely satisfied with it except for one thing: One significant negative trait mine has is that in test pictures shot at f3.5 versus 1/3 stop down, the center of the image is f3.5, but the edge is about 1/3 stop down. This is apparent over about 50% of the total image area, so it is very significant in my test pictures. However, even 1/3 stop down and this lens performs great. That was a complicated way of saying it is very critical vignetting at f3.5, but even 1/3 stop down and it is great. Very little distortion, and great depth of field, but please stop it down at least once click.

 

For comparison, I also own the Canon 16-35mm f2.8L which performs consistenlty over its entire range, but blasphemously prefer the performance of the 10-22mm ef-s.

I have not used the Tokina.