dmward wrote in post #15691319
This may just be semantics, but a reflective meter is reading light reflected from the subject. The reflectivity of the subject surface directly influences the meter reading. That's why a light in a mirror has a disproportionate impact on the meter reading. The meter's light reading algorithm is based on the assumption that the total reflected light will be middle gray (18%) Thus a reflective reading from a white wall results in a middle gray wall image and a reflective reading from a black wall results in a middle gray wall image.
I would say that the assumption is done by the photographer who meters off an 18% gray card. The assumption is that the subject (e.g. person in neutral clothing) shares the same reflective value as the card. But when metering based off a card and basing your camera settings on the reading, the scene will be accurately captured insofar as getting the luminosity of the recorded image to match what was 'experienced'. But like many, I choose to implement ETTR when recording the image, so this is not an option for me.
Furthermore, how do you qualify "proper exposure" after metering the gray card? Is it when the needle hits the center gradation? Well, what I'm saying is that the center gradation is that 3 stop under sensor max that I mentioned. Just shoot a neutral surface like a piece of paper in Av or Tv, evenly lit, then go into manual mode and take additional shots increasing EV by 1 with each successive shot. At +3, you should be clipping the image.
Canon describes the various ambient meter setting options available to bias the meter reading from the field of view the meter sees via the lens. The ETTL metering biasing options are limited to two, average and evaluative. How they read the pre-flash as reflected by the subject is described as well. i.e. evaluative center weights and does some other things to bias the reading toward the center of the field of view.
Thanks David. I'll look for this, but if you have a link handy I'd really appreciate it.
Your description of the meter "backing" off 100% to get to mid-tone is a bit misguided. Its more accurate to say that the meter is calibrated based on the premise that the sum of the light striking the meter sensor is the amount of light required to properly expose a subject that is middle gray. The way photographers correct for that presumption is to use Exposure Compensation (EC) correct for subject reflectivity that is greater or less than middle gray.
The "middle gray" concept is based on the card being in the same scene as the subject. The camera meter knows nothing about this concept. How could it? It can only take a reading of brightness and spit out a value. And I think this computation is based on the sensor. If you take two sensors, each with different dynamic ranges maximums, I would guess that the settings of an autoexposure of the same scene would differ between the two.
I don't know if there are any still around, but when I was a young photographer shooting film Sekonic made a reflective meter that was very popular and one option was a Zone System dial. The way to use it was to read the light reflected from a specific part of the scene and place that reading in the proper zone on the dial. The result was the exposure required for the film selected (ISO rating) and development planned, (Pull, normal or push) to get a negative that had the metered area of the scene at the proper density to print in the zone the photographer wanted.
With current digital cameras and raw file processing software proper exposure for an image is both more complex and more forgiving. One technique is Expose To The Right (ETTR). Using this technique, taking care not to clip highlights, provides much better shadow detail and minimizes noise. Metering for the technique requires evaluating the scene and using EC based on total reflective value relative to middle gray which is where the meter is placing the exposure.
So in both cases, we're mapping the scene's dynamic range to that of the medium, be it film, digital sensor or print. I understand the technique that you describe and I think the digital equivalent might be to spot meter the brightest part of the scene and use EC - I say + 2.67 for the perfect exposure.
Wilt wrote in post #15692181On this point...
Some folks think that Av mode 'uses flash as fill' (as opposed to other modes, where flash is NOT used as fill, I suppose, or that 'flash as fill' is different than 'flash as main source'). In fact, tests confirm that
- the amount of light output by the flash onto the subject is identical for all automation modes, whether Av vs. Tv vs. M...Av mode behaves no differently than other modes in the amount of flash intensity!
- the amount of light emitted by the flash onto the subject is not different for 'fill flash mode' vs. for 'main light mode'...there is only a single metering algorithm that reads the intensity of light returning from the preflash
There is professional who has made a statement essentially representing Canon USA (but who is not a Canon technical employee), whose flash statement accuracy CANNOT be proven true, but which is actually DISPROVEN via tests.
Can I trouble you for links? This sounds really interesting.!
Curtis N wrote in post #15692367Keep in mind there are three different ways you can use the data from the various metering zones.
Evaluative E-TTL is the default. That's where it attempts to identify the subject by comparing the flash and ambient readings from each zone.
By custom function in your camera you can switch to Average E-TTL mode (This is totally separate from the ambient metering modes on the camera).
Finally, the Flash Exposure Lock uses only the center of the frame, and is akin to spot metering the flash exposure.
Thanks for the info Curtis. I can't imagine any situation where average would work better than evaluative. FEL, perhaps.
dmward wrote in post #15692963As I predicted...
Here is the link to the Canon page with the description of using flash for fill;
http://www.learn.usa.canon.com …ash_use_EOS_article.shtmlWhether the individual writing the article is a Canon employee or not is irrelevant. This is Canon educating us on the capabilities of its technology.
Enough said on this extraneous topic that has no relevance to this conversation.
I think I read this a while ago, but I'll have a look at it. I remember Canon expounding the virtues of using flash to brighten your subject a little. But in the Photonotes article, the author describes how flash power is automatically reduced at certain EVs. I always found this suspect.
digital paradise wrote in post #15693104That is interesting and I have said that myself. I know the flash meters separately than the camera's light meter so I wonder if it is just a way to describe something. When shooting indoors in dark conditions and you put the cam on M the ambient will probably show underexposed thus using more flash - dominant light source. If you switch to Av then the meter will centre which gives you almost all the light you need so you require less flash or fill. So it is not Av that is telling the flash to work less hard, the light that reflects back after the pre flash tells the flash to work less hard thus interpreted as a fill. Outdoors it is basically a fill in M or Av if metering to balance with ambient. Only if you kill the sun then the flash becomes dominant and you probably won't be doing that in Av. It is possible if you are only going a few stops but my guess is most people will choose M.
Are these tests equal? By that I mean Av centres the cameras meter. When testing in M is the camera's meter centred as wel,, same subject and both in the same lighting conditions? If so then yes if would only make sense the flash output would be equal.
Are you questioning if the meter is the same in both M and Av? I always assumed it was.
