Date: Thu, 07 Aug 2003 09:54:13 +0900 From: Jim Davis <[EMAIL PROTECTED]> Subject: EOS 10d metering and shutter rants
Here's some thought to get things moving here perhaps...
I'm convinced the metering in the 10d could be better, or there is a bug.
Nope/nope. The error is behind the camera. The secret lies in understanding what the camera does.
Several times, while shooting a scene, shooting a series of 3 or 4 shots, maybe a couple of seconds apart, I've had one of the images be way underexposed. I check the EXIF info and find it's 2 or 3 stops
underexposed from the others with basically the same scene.
The miracles of matrix lottery.
Now with partial metering I understand how this can happen, but I'm now using Evaluative mostly for the reason that I want more consistancy when I'm blazing away at a really important time.
I use full manual all the time. No consistency problems.
I thought evaluative broke up the scene into many segments of a grid.
Approximately. And then they try to guess what is in front of the lens.
http://www.photoquack.de/tutorials/ematrix.htm
If indeed they do, the centre weighting of it would seem a bit extreme. Lately I took 3 photos each about 2 seconds apart. There was a slighty shift in some
background elements,
But certainly no change in lighting. Manual would have given you perfect consistency.
How can 2 of the shots be so nicely
exposed, while one way under?
On the first ones the camera guessed right, on the others the squelch was crossed into another field in the lookup table. Which resulted in a misguessed exposure.
I like evaluative most of the time, but there seems to be some kind of bug in the camera exposure, or else evaluative is more like simple centre-weighted than we've been lead to believe.
Isn't there some kind of database of scenes in camera memory that are being compared?
There is. But tell me, how do you think the camera can decide wether there is a black door in bright sun or a white door in deep shade in front of it? Try it, both will have the same partial meter reading. One needs correction plus, the other minus. But how to decide? And what if the door is indeed gray? The camera can't recognize the nature of a subject, you have to do it.
Does anyone else wish the shutter button needed to be pushed down a little more to shoot continuous shots?
No.
I find myself taking at least 2 most of the time in continuous drive mode, even though I only want
one, and I'm feathering the shutter very softly.
Obviously not easy enough. I shoot single frames even on EOS-1 bodies set to continuous.
I seems unless you push and release very quickly you'll take more than one.
More training will do the job.
I'm going to try manually setting my exposures and see how it goes.
If you just follow the arrows and don't understand what the camera is doing and why, manual won't be any help. Get Ansel Adams books on the zone system and learn the concept.
It might be interesting to see if the camera meter reading in Manual mode using eval metering varies much in this type of scene,
It will just like before. Evaluative is the most silly thing to use (besides AWB), because you can't predict what the camera is going to fuck up next.
I do need consistant metering while using compensation based on experience.
(white bird go down one stop, light grey, half stop down)
In AE mode. Meter off a reference object and you can use the same settings for white/gray/black birds.
If I use Manual mode, I can only tell how much proper exposure has changed by the metering scale, I must change exposure now manually, plus add my compensation.
You need to learn more about the concept of metering and why the camera does what it does.
Or I could use partial metering and try and find a medium grey value in the scene and set Manual with comp to that.
If you have a medium gray (or similarly bright) reference subject, you need no compensation. Unless you want to manipulate the tonal values in your subject.
What do you guys do?
Set the camera to (M)aster mode and break off the dial. Just to make sure.
-- Michael Quack <[EMAIL PROTECTED]>
* **** ******* *********************************************************** * For list instructions, including unsubscribe, see: * http://www.a1.nl/phomepag/markerink/eos_list.htm ***********************************************************
