D4 vs 1Dx (Imaging Resource), RAW, ACR, 6400-12800

Started Sep 4, 2012 | Discussions thread
bobn2 Forum Pro • Posts: 59,036
Re: D4 vs 1Dx (Imaging Resource), RAW, ACR, 6400-12800

Lance B wrote:

bobn2 wrote:

Lance B wrote:

You didn't bother to read it did you? If the EV (f-number and shutter speed) and the scene luminance is the same then the exposure must be the same, whatever the ISO setting or 'accuracy'. ISO is not a part of exposure.

I understand that shutter speed and aperture are the exposure, but it really is rather a silly terminology,

ah, silly, is it? Well, that's the way ISO uses it when the set up the Exposure Index standards, so I suppose they are silly too. What we really want is a recursive standard that defines itself to be part of itself.

but you still need to get the correct luminence for a scene

the scene has the luminance that it has, unless you are using flash or other controlled lighting - I don't suppose that is what you are talking about when you say 'get the correct luminance', though. The silly terminology that I'm using is that 'scene luminance' is the amount of light given off by the scene that you photograph.

I well understand that. I never implied that the scene luminence was controlled by the camera!

How then are we to interpret 'you need to get the correct luminence for a scene'. How do you get it if you don't control it?

and this is what I am saying. Yes, exposure is the same, but one shot at ISO6300 and one at ISO100 would result in very different outcomes, ie one either way too dark or the other way too bright. The fact is, you still need to get it to look correct.

While that is true, it is irrelevant to this discussion, which was about the need to ensure equal exposures when comparing cameras if you want a good indication of their performance at some given light level. For the constraints are on the actual exposure parameters, the scene luminance (which you generally have no control over at all), the shutter speed (generally constrained by motion blur concerns) and f-number (either the largest aperture you have or minimum DOF you can stand). Those are the things you can't change. It's easy to change the ISO to get the tonal range you want in the output image (if you are using the camera processing), or even just to change the 'brightness' in the JPEG defaults if you don't like what the camera is giving you. Strictly speaking, by doing that you have changed the 'ISO' that the camera yields at that setting.

Yes, I also well understand that.

For the sake of the discussion, my response to this whole thing was to Tedorian2's suggestion that the cameras were not using the same "exposure". The point is, the cameras were being tested for ISO noise at certain camera indicated ISO's and therefore they were set to the same ISO, ie ISO6400 or ISO25600 or whatever, which is what many people would do "out in the field" and, for the sake of this particular example, then set their aperture and get a resultant shutter speed by the camera's exposure meter to obtain "correct exposure". This appears to be how this test was conducted, and most other tests that test ISO noise.

Yes, but it yields a silly result - one where a camera can appear better in low light simply by gaming the meter to give more exposure at some set ISO. that really doesn't help, because if you need 1/30 sec, say, you'll raise the ISO till you get it, even if that results in a different setting on different cameras, so this really isn't a good tets of low light capability.

As you say, a truer indication of ISO noise

there is no such thing as 'ISO noise'. Get into terminology of that sort, you'll get very confused.

would be to set the aperture/shutter combo and then set an apropriate ISO for each camera in order to get a correct result depending on the luminence of the scene.

That was what tedorian was suggesting, or else set the same ISO and the nominal exposure for that ISO, then you'll see how 'slow' or 'fast' that ISO setting is.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow