EOS 30D/550ex and ETTL2

[jasona1]

I am under the impression that exposure should be the same with ETTL2 regardless

of whether a flash is bounced or not. Obviously shadows will look different but

the exposure should be the same should it not?

[jim_larson1]

Well. . .yes. . . .

 

. . .but the picture is going to look WAY different. And probably better.

[picturesque]

Theoretically the two exposures should be the same value. In the real world, it probably doesn't happen that often.

[jimstrutz]

There are a lot of reason why they would not be exposed the same. In many situations, bounced flash tends to light up the background more, and with evaluative metering this is going to change the way the camera calls for lighting the subject -- perhaps underexposing it. In other situations, the fact that the light would be coming from the ceiling, and possibly leaving shadows on the face, might cause overexposure of the hot spots on the well lit areas. And then there is the deep secrets in the minds of Canon flash engineers that have programed the camera to behave differently with different patterns of light and dark. Who knows how that will effect things.

 

So while the simple principle of TTL metering (E-TTL II or otherwise) should meter them for the same exposure, it probably won't happen. The same ideas would be true of auto flash metering, but it would be different again.

[jasona1]

Here is a quick example. White/Black backdrop. First shot is direct and second is bounced. White walls and white ceilings.

[jasona1]

Bounce

[steve_dunn2]

<p>If your lens does not return distance information, and you frame the shot exactly the same way, and the flash is powerful enough to provide full illumination even when bounced, the exposure should be pretty much the same.</p>

 

<p>If your lens <em>does</em> return distance information, then all bets are off. The distance information cannot be used when you use bounce flash, because while the lens can tell the body how far the light travels on its way from the subject back to the camera, there's no way of knowing how far the light travelled on its way from the flash, off some mysterious object outside the field of view, to the subject, or how much light is lost by bouncing off that mysterious object. When you use direct flash, on the other hand, the distance information <em>is</em> used, because the distance from flash to subject is known to be the same (to within a reasonable approximation) as the distance from subject to body.</p>

 

<p>As I discovered <a href="http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00GPLS">when shooting a subject with the same body, same flash, and two different lenses, only one of which returned distance information</a>, the use of distance information can make well over a stop's worth of difference to flash exposure with E-TTL II.</p>

[jasona1]

But why should a shot with a direct flash be underexposed like that. It makes no sense.

[jimstrutz]

<I>"It makes no sense."</I>

<P>

It's an imperfect system. They all are.

[steve_dunn2]

<cite>But why should a shot with a direct flash be underexposed like that.</cite>

 

<p>Because Canon seems to have put a bias towards underexposure into E-TTL II? As I mentioned in the thread to which I linked, I have three lenses which return distance information, and E-TTL II tends to underexpose with all three. The fourth one doesn't return distance, and I rarely use it with flash, but on the one occasion I can think of when I did use flash as the main light source for this lens, the pictures were overexposed. (This is all with direct flash; I haven't put enough thought into the results I've had with bounced flash to have figured out if there's a pattern.)</p>

 

<p>Presumably the reason for the underexposure bias is that with digital, you cannot recover blown highlights at all, whereas you can recover from underexposure (albeit with more noise, and in the case of severe underexposure, more serious appearance issues), so Canon decided to err on the side from which you can recover.</p>

[picturesque]

Steve is right--ETTL II seems to have a strong tendency to underexpose greatly whenever there is a white object in the scene, even bounced. To confuse yourself even more, put the camera in averaging flash metering (custom function 14) and try the test again. Unfortunately manual flash is the only way to have total control and consistency over your flash exposures. Otherwise, you learn how the flash metering works in different situations and compensate accordingly in the field.

[DickArnold]

Wow. I wish I understood my 550ex on my XTi better. I actually bought it four years ago to go with my D60. Inexplicably, I have gotten some less than satisfactory results which I cannot explain. I found I do a lot better by using XTi on manual and ETTL. I do a lot of bounce flash which seems to expose better than direct flash. I have three L lenses. At least two of them measure distance. I used to shoot weddings with Vivitar 283s on medium format. I felt I had much more precision and control than I have with the 550EX. At least I could predict the outcome and I had a lot of confidence in the product. I previously had a 420EZ, 430 EZ and 540EZ flashes which I used for press photos and weddings. I have not wrapped my mind around this because with digital you can look at the screen and adjust what you do on the spot. You all have shed some light but there is no accounting for algorthyms. I sometimes miss my Bronicas because almost everything was manual and I had only myself to blame for bad pictures.

[jimstrutz]

Of course with print film there is a lot more latitude, and the corrections were done in the lab. So if it wasn't right, they usually made it look right. Most of the time we never even noticed moderate exposure errors. With digital, the photographer is stuck doing the processing adjustments, and doing it with less latitude. Almost like slide film, but they weren't adjustable at all.