Bounce Flash Question...

Ive been shooting for about 6 months now with 5Dmk2 and the last couple of indoor gigs i found myself having to put my flash on manual 1/16 or 1/8 power to get a nice exposure when using bounce flash (turned upward to low ceiling). ISO @2000, Aperture 4.5 and shutter 80 or 100sec. matrix metering
Is this normal?

When i bounce it in TTL its always underexposed...

How about a few examples with exif info?

Are you in HSS when you're having exposure issues? Are you unable to correct with FEC?

Keep in mind the angle of incident equals the angle of reflectance. It could be your bouncing the light behind or in front of the subject and not onto it.

ETTL seems inconsistent when bouncing for me. I sometimes add some flash comp, sometimes I go manual.

Mark1 wrote in post #14245165
Keep in mind the angle of incident equals the angle of reflectance. It could be your bouncing the light behind or in front of the subject and not onto it.

First, the whole "angle of incidence" rule only applies to direct reflections (like bouncing off a mirror), not diffuse reflections (like your typical ceiling covered with flat white paint).

The key to effective bounced flash is to created a bright spot on the ceiling (or wall or whatever), and that bright spot becomes the apparent light source for your subject. You aim the light where it will create the effect you want, from your subject's perspective. I aim my flash head straight up, more often than not, but it's driven by the situation.

As for the OP's question, here's my theory: Bounced flash tends to light up the background more than direct flash. This affects the E-TTL metering system as it reads the entire scene and tends to result in underexposed subjects, and the remedy is to dial in some Flash Exposure Compensation.

"Matrix metering" is not an option on Canon DSLRs. Regardless, the ambient metering mode you choose has no bearing on flash exposure. Try switching the E-TTL metering mode to Average (the default is Evaluative). You'll have to dig through the camera's custom functions to find this option.

When you are bouncing off the ceiling back down to the subject the effective distance from flash to the subject will tend to relatively constant. Thus a manual flash power setting will be close for all the shots.
With TTL metering every shot is metered based on the chosen metering modes. This means that a darker subject in the metering zone will suggest more light and a lighter subject in the metering zone will suggest less light.

Curtis N wrote in post #14249581
First, the whole "angle of incidence" rule only applies to direct reflections (like bouncing off a mirror), not diffuse reflections (like your typical ceiling covered with flat white paint).

Yes and no. While it is true some light will go in all directions. The bulk of the light will follow along the path. It does not become omni directional just because it reflects off a cieling.

Curtis N wrote in post #14249581
First, the whole "angle of incidence" rule only applies to direct reflections (like bouncing off a mirror), not diffuse reflections (like your typical ceiling covered with flat white paint).

I gotta side with Mark1 on this one. I set my meter on the floor 10' from my position...

• shot with flash at full power straight up to 8' home textured white ceiling ...f/2.8 +0.4 (ISO 100)
  
• shot with flash head at 45 degrees...f/4 +0.2 (ISO 100)

Mark1 wrote in post #14250070
Yes and no. While it is true some light will go in all directions. The bulk of the light will follow along the path. It does not become omni directional just because it reflects off a cieling.

It absolutely becomes omni-directional. Take a beamed continuous light like a flashlight (torch), and set it on a table so it points straight up at the ceiling. See that bright spot up there? Now go and walk all over the room. You can see the bright spot from everywhere, because the light is being reflected in every direction.

Even non-glossy surfaces have a certain percentage of direct reflection (which PacAce demonstrated to me years ago). But with flat white ceiling paint, or textured ceiling tiles, that percentage is small.

Moreover, direct reflections create a small apparent light source. If you look at a light source reflected from a mirror, it still looks small. The beauty of bounced flash comes from the large apparent size of indirectly diffused light.

Curtis N wrote in post #14251253
It absolutely becomes omni-directional. Take a beamed continuous light like a flashlight (torch), and set it on a table so it points straight up at the ceiling. See that bright spot up there? Now go and walk all over the room. You can see the bright spot from everywhere, because the light is being reflected in every direction.

Even non-glossy surfaces have a certain percentage of direct reflection (which PacAce demonstrated to me years ago). But with flat white ceiling paint, or textured ceiling tiles, that percentage is small.

Moreover, direct reflections create a small apparent light source. If you look at a light source reflected from a mirror, it still looks small. The beauty of bounced flash comes from the large apparent size of indirectly diffused light.

But Curtis, the test results which I published do show that angle has a contribution in intensity of bounce flash...in the situation I cited, the angle of flash head made 3/4 EV difference in intensity at the subject. That supports his earlier statement, "While it is true some light will go in all directions. The bulk of the light will follow along the path", doesn't it?! It is somewhat directional (as illustrated by the 45 degree angle intensity being greater), but also less directional than a mirror (as confirmed by the fact that the subject got some illumination while the flashhead was pointed straight up.

Wilt,

Your description of the experiment setup doesn't give me enough information to identify possible flaws in the methodology. There is more involved than angles. But I can take your results at face value, and anyone else reading this thread can do their own testing. The extent of direct reflection is irrelevant to the OP's problem.
- We both know that the 580EX II has plenty of power for bounced flash at f/4.5 ISO 2000 in a typical residential situation.
- The OP proved it's not a power or efficiency issue. He switched to manual flash (at low power) to get proper exposure.

Hopefully we can agree that it's an E-TTL metering problem, changing the angle of the bounce won't solve the problem, and Mark's advice regarding the flash angle was without merit in this situation.

Curtis N wrote in post #14251617
Wilt,

Your description of the experiment setup doesn't give me enough information to identify possible flaws in the methodology. There is more involved than angles. But I can take your results at face value, and anyone else reading this thread can do their own testing. The extent of direct reflection is irrelevant to the OP's problem.
- We both know that the 580EX II has plenty of power for bounced flash at f/4.5 ISO 2000 in a typical residential situation.
- The OP proved it's not a power or efficiency issue. He switched to manual flash (at low power) to get proper exposure.

Hopefully we can agree that it's an E-TTL metering problem, changing the angle of the bounce won't solve the problem, and Mark's advice regarding the flash angle was without merit in this situation.

Curtis, I agree that flash head angle was likely not the direct remedy to the insufficient exposure issue. ETTL metering preflash could have been fooled by a brighter background or subject, so that insufficient light was commanded by the camera. And, I my own testing confirmed, there was sufficient light at 1/16 power and ISO 2000, for the subject to have been sufficiently illuminated for f/4.5 aperture.

On the tangent issue, though, 'total non-directionality' of bounce I believe to have been disproven by the fact that bounce flash illumination of the subject was increased when the flashhead was aimed forward at 45 degree angle.

I am wondering why you are shooting at ISO 2000 when using a flash? Seems to me that with a flash (especially a 580 exII) that you should have plenty of power to shoot at ISO 200 or even lower, even when bouncing. I would expect you to be overexposing, not underexposing with an ISO of 2000. Have you tried shooting at ISO 200 or even 100 and checked your results? Sounds like a metering problem with the camera, or the flash. Check to see how brightly the flash is firing when you are getting the underexposures. I had a problem with my 580 exII firing at full power (all the time) and severely overexposing my images even when shooting ETTL. You could see that it was firing way too strong...you should be able to see if it isn't firing with enough power to get a proper exposure. At ISO 2000 it shouldn't take much flash to get a good shot! Make sure your hot shoe connections are clean, as a bad connection between the camera and flash will result in a bad exposure as well. Hope you can figure this out!

Point-n-shoot-n wrote in post #14256482
I am wondering why you are shooting at ISO 2000 when using a flash?

Nothing wrong with cranking up the ISO. It brings up the ambient exposure, rather than rendering the background dark. It also requires less flash power so you get faster recycle and longer battery life.

E-TTL flash will take the ISO setting into account. One setting won't give different results than another, as long as the flash power required is within the range of the flash unit.

Huh? An underexposed image is pretty strong evidence that your flash unit isn't firing with enough power. The key is to figure out why.