Urgent bounce flash question 760D owners

walkaround wrote:

WilbaW wrote:

walkaround wrote:

See my response in the 70D forum. You have done zero research to understand what you are doing wrong (yes, it is user error). The 70D works exactly the same for bounce flash metering as all the other advanced Canon cameras, which 2 minutes on Google would have shown you.

How do you explain this? - "If I flash exposure lock on the subject before snapping the bounce flash picture with the 70D, all the photos are exposed perfectly. If I don't FEL before snapping the photo, they are all very underexposed."

Because FEL uses an entirely different method of metering the subject.

70D manual

It's metering the subject off the center zone, I hope everyone realizes that.

Which words say that, and what else do I need to know to interpret them that way?

And when not using FEL, Evaluative E-TTL is metering the whole scene:

When you press the shutter button, the ambient light is metered by each sensor. Then a pre-flash fires and this is also metered by all the sensors. The ambient and flash readings for each sensor are then compared. The area of the scene that shows a significant difference between the readings is the area where the subject is most likely to be. This is because the main subject is likely to be closer to the camera than the rest of the scene and will reflect more of the flash. However, if the difference between the two readings is significantly high, the camera will ignore this area on the basis that the flash is being reflected back to the camera by a very shiny surface, such as a mirror. The pre-flash meter readings from accepted areas is weighted and averaged. It is then compared with the ambient light reading before the main flash output is calculated and stored in memory for the exposure. If the lens is able to provide distance information, this is used to determine the closeness of the subject and any highly reflective areas relative to the background. This information is used to refine the flash exposure. The result is better flash exposure for difficult subjects, such as white wedding dresses.

I'm not seeing what you're claiming is said there. How do you know that FEL doesn't use evaluative metering?

And if we look at the Troubleshooting section of the 430EX II manual, we see the very first point about underexposure mentions reflective objects. (also warns about HSS and range decrease)

Err... you think that the problems reported with the 70D are because they have a mirror in the shot?

I can't explain (yet) why it seems to work differently in Rebels. They use E-TTL II also, and should behave the same, although they could be using a different algorithm and we would never know. It's entirely possible that, as an entry-level camera, it is making assumptions about flash output that are not deemed appropriate for the enthusiast and pro models.

Sure, but what evidence do we have that they do behave differently? What differences are observed?

There are a lot of variables, which of course can be further disrupted by modifiers, room color, room size... it's a complex dance. It's kind of absurd to expect to never use FEL or FEC, and receive perfect exposures in all use cases.

Sure, but I don't hear anyone saying they are expecting that in reports about the behaviour of the 70D. I haven't specifically tested this lately, but in my experience you get similar brightness with or without FEL, in evaluative or CWA, with the flash is direct or bounced. Not "very underexposed" without FEL as reported for some 70Ds. That's not normal.