Originally posted by skyoftexas I guess the only way I could be sure of what you are saying would be to shoot with you and compare notes. I don't discount that the histogram has value, but I think its value really is to make sure you aren't clipping the high lights. Otherwise, I don't think you can look only at a histogram, and predict the quality of picture you will get. There are too many situations when the histogram of a very successful picture will be crowded on one end of the spectrum or the other. Shooting in a dark room is one of those situations. Shooting into the sun is another. If one only accepted photos with a nice, spike in the middle of the histogram, a lot of interesting photographs would be overlooked. To support my uncredited opinion, I will refer to page 190 in Scott Kelby's book, The Digital Photography Book, Vol. 2, in which points out the limitations of histograms and concludes: "–so don't get hung up on histograms---"
Getting correct exposure is not that difficult. But blending flash into your photo is. If you slap a flash gun on your camera, set it to TLL or even Automatic, and point it at the subject, you will probably get a correctly exposed photo. Washed out, and not very interesting, but correctly exposed. But if you want to see the background behind your subject, and try aiming the flash behind you or bouncing off of something, you need a visual confirmation to see if your flash setting is correct. A histogram might be of some use, but I still believe that visual confirmation is necessary. But maybe I'm wrong. I'm willing to be corrected.
Kelby is right that we get too hung up on histograms. I give the histogram a quick look, and as long as it looks right compared to the scene, I don't get too bent out of shape.
Most of the time, as long as you aren't too stacked up at either one side or the other, then you will have a good exposure.
The obvious exceptions are either high hey or low key where the histogram will be very much on one side or the other.
So, the histogram is important, but it has to be viewed in relation to the scene in question (but not necessarily the review screen.)
Flash exposure is the one place where the histogram doesn't tell you enough since, as you have surmised, it is possible to have a glorious histogram and a totally wrong fill exposure.
Again, experience pays dividends, and in my case, a total lack of ability to run anything more complicated than a simple auto flash has saved me more than once.
I don't use a TTL (or pTTL) flash.
I just use a simple but very powerful auto flash, and I set everything manually.
I don't like to leave that sort of thing to chance, or technology.
I set my own aperture and shutter speed, and then I set the fill ratio by biasing the flash to either a higher or lower aperture number to give more or less flash output.
I suppose I should try to figure out how to run my 540 flash sometime, Lord knows I paid enough for the thing and it has only been used once to make sure it worked at all, but even then, I set it on auto, chose an aperture and used it that way rather than on TTL.