Evaluative vs Average Flash Metering

I found out that for indoor shoot the Evaluative then to underexpose the subject while for Average Flash metering it'll overexpose the subject if the subject has fair skin tone.

Even ISO crank up to 800, the evaluative still underexpose the subject.

May I know what's the logic behind this behavior?

If your using the camera to AE (Auto Exposure) your image then cranking the ISO up wont change the end exposure.

Is your exposure compensation set to -0- ?

As TomTom noted, changing the ISO won't affect flash exposure because the meter will account for the higher ISO and reduce the flash power.

It's just like changing the aperture in Av mode, the camera automatically shifts the shutter to compensate.

Evaluative and Average metering change the way the camera looks at a scene. Typically evaluative will see white or shiny objects in a scene and it will try to not blow them out. This leads to dark pictures. Use of evaluative at a wedding for instance requires positive FEC because of that white dress.

Average will generally deliver a more consistent lighting of a scene, but if the background is beyond flash range the camera may blow out near objects futilely trying to light it.

Once you know how the camera behaves, you can dial in some FEC headed into a shot and get consistent results. I personally find average flash metering to be more predictable.

Those are interesting points about Average, I might give that a shot because evaluative ETTL is becoming too unpredictable for my liking (though it's my user error that's at fault)

Here is an interesting set of tests of indoor flash with different ambient and flash metering modes (page down to find the tests)

http://super.nova.org/​DPR/Canon/TTL/

Basically, no combo is "the" solution. Knowing hoe your camera reacts to different situations and what to do is what is imposrtant.

danielyamseng wrote in post #5946815
I found out that for indoor shoot the Evaluative then to underexpose the subject while for Average Flash metering it'll overexpose the subject if the subject has fair skin tone.

Even ISO crank up to 800, the evaluative still underexpose the subject.

May I know what's the logic behind this behavior?

In Evaluative mode, the camera will use the ETTL-II algorithm to determine flash exposure. ETTL-II will try to determine what the subject in the scene is and bias the exposure for it. So, if the subject is fair-skinned, then the exposure will tend to come out a little underexposed unless FEC is used to increase the exposure. (The opposite would be true if the subject were dark-skinned).

Avarage metering means exactly that. The light reading from all parts of the scene are averaged without giving the subject any special consideration. So, if the subject is a small part of the whole scene, the overal tone of the scene will determine the exposure. If the scene is generally dark, the picture will tend to overexpose. If the scene is generally white, as when in a small room with white walls, the picture will tend to be underexposed. If the subject makes up a significant part of the picture, then it's overall tone will determine the exposure.

If I use FEC, I won't know how much of stop I need to add in without doing several attempt.

Would light meter in this case would safe my time?