Upload & Sell: Off
cgardner; thanks for the great reply. Very appreciated.
I've got a section on Canon flash on my tutorial site: http://photo.nova.org
I bought into the Canon system in 2004 with a 20D camera and a pair of 580ex. E-TTL vs E-TTL II metering is determined by the model body and the 20D was one of the first to use the E-TTL II zone based evaluative metering. There was very little information about how it actually worked, so I experimented, gleened infomation from Canon white papers, and e-mailed Chuck Westfall of Canon with questions.
I did tests to visualize what the A:B ratios produce and others to compare all the possible ambient / flash metering modes. I concluded that the quickest route to correct exposure with detail everywhere in ETTL mode is:
1) Start by setting shutter / aperture / ISO to keep the brightest ambient highlights under clipping. Why? You'll blow them even more when the flash overlaps.
2) Set FEC = 0. That's the default "I think this is right" guess. Will it be right? Only some of the time, but always startting from the 0 baseline makes it easier to grasp when and why it doesn't sometimes.
3) Set ratio to A:B= 1:2 Master A / Slave B. That should match scene range to sensor indoors on most camera sensor ranges, or be very close.
4) Take a shot and chimp. Adjust FEC until solid whites are just at or below triggering the clipping warning. It's not possible to objectively judged midtones and shadows in the playback if the highlights are not exposed the same way consistently with detail, ideally with an eyedropper reading of 250.250.250 on white paper.
The only thing that should be clipping are the specluar reflections. Why? When you look at a white object what provides the clue to 3D shape besides the shadows in the low parts? The specular highlights on the high points. If the solid highlights clip those 3D shape clues are lost, which is why blown highlights look flat in photos.
5) Once the highlights are dialed in, examine how well the A:B=1:2 starting baseline did in the shadows by looking at the detail in the playback. Adjust A:B ratio to 1:1 for lighter shadows, to 1:3 and greater for darker. Once FEC is dialed in for the highlights correctly under clipping the metering wiill keep them the same when you change ratio. But if you start with clipping highlights in the ambient or the flash before shifting ratios the metering will get confused.
I shoot the majority of shots with 1:2 to capture the full range of detail, tweeking as needed in Photoshop when I see the results in the RAW file.
What is seen in the playback is a JPG. RAW results will have more "headroom" but if you plan to create JPGs at the end of a RAW workflow for the Internet the 16 bit > 8bit JPG process will clip highlights and shadows. So the camera playback on the camera is a fairly accurate rendering of what you can expect at the end of a RAW>JPG workflow. By way of analogy if your end product is an 8' long board you need to cut the tree into 9' logs to allow spoilage during finishing.
If the RAW highlights are exposed on the "bleeding edge" of the solid white clipping at capture they will clip the the JPG in ways you might not notice, such as the red channel in skin highlights clipping which gives them an odd, waxy, yellow look.