Okay. I picked up a Zeiss Distagon 21mm f/2.8 and Cameraquest adapter. I have no problem with the concept of manual focus and limitations of metering. I just am not quite sure how to execute stop down metering. I *think* I need to stop down to the aperture I want to shoot at and meter manually. Is that the way? Can I use any of the auto metering modes (e.g. matrix, center weight, spot, etc.)? Can I use any of the auto metering modes (Av, Tv)? Any pearls of wisdom?

Thanks in advance.

Set camera to Av mode. Focus wide open for greatest accuracy; then stop the lens down. Recompose and shoot; the camera will set correct shutter speed. (Check histogram; you might need to set exposure compensation).

tom in mpls wrote: ... the camera will set correct shutter speed. (Check histogram; you might need to set exposure compensation).

I've found this NOT to be the case on some cameras. I've tried a 5d right next to my 1Ds2, and the more you stop down on the 5d, the worse (over exposing) the exposures get, even as far as 2 or 3 stops to bright. While the 1Ds and 1Ds2 remain perfect and repeatable exposures at any aperture. We tried different metering methods, to the same result. I've come across some people who saw this problem in their 5d, and some who didn't see any problems like this at all. I'm honestly not sure what causes it.

So, if infact your camera does this as well, the only way we found to correct this was to focus and meter both wide open, then stop down (counting how many stops down you go) and calculate your shutter speed manually. (if you stop down 3 stops, move your shutter speed slower by 3 stops, and so on). For some reason some cameras seem to meter differently when the aperture is closed down. I'm not sure if this has to do with the mirror, or what the heck is going on, but it was not as quick and easy as my cameras. Even my 10d meters just fine with manual lenses stopped down. But the 5D didn't.

If using Av or Tv modes, to reduce the effects of vignetting affecting the camera exposure calculations as you stop the lens down, you also need to use centre weighted or spot metering rather than full evaluative metering , otherwise you will get overexposure problems as you stop down caused by the camera calculating the exposure including the now much darker image edges.

The preferred method is to mimic what the camera exposure sytem actually does - ie meter with the lens wide open. Use Manual mode, take exposure reading, count the aperture clicks as you stop down and adjust the exposure by the similar amount of f stop increase.

On modern cameras all exposure calculations are carried out with the lens at full open aperture, there is an overide calclation as you reset the aperture with a standard automatic lens, but the metering is still calculated from the full bright image.

On older cameras using stop down iris lenses the exposure was calculated from the reduced brightness image - it was rare to have full evaluative metering, most cameras therefore used centre weighted average or spot metering precisely to overcome the edge darkening issue confusing the exposure calculation - it was only when the full auto iris lenses appeared that full evaluative metering started to appear in SLRs

Daniel

What you are seeing in the differnemce between a crop sensor camera and a full frame dslr is a reduction in the edge vignetting caused by the aperture iris closing - the edge darkness is being cropped out, so on a 1.6 crop DSLR the metering is thrown less than with a FF dslr

Mike

Mike, both the 1Ds2 and the 5d have the same (physical) size chip, and I correct? We went back a second time and tried all the metering modes on the 5d, including spot metering. The 1Ds2 meters just fine on any of the modes, and the 5d did badly on all of the modes. I don't recal if I can change metering modes on my 10d (it's an infrared camera only, and I don't use if often) but whatever mode it was on, it metered just fine.

Edited by Daniel Buck on Mar 27, 2008 at 11:53 PM GMT

Ditto Daniel. 1Ds2 and 1Ds3 spot on; 5D and 40D a train wreck. The 5D was typically 2, sometimes 3 stops over exposed (around F8 or F11). I didn't shoot enough with the 40D to determine if it was consistently wrong, or just a fluke. The 5D was with AF confirmation + non-confirmation adapters.

Sorry misread your comparison, I thought you were comparing crop with ff.

I agree odd behaviour and foibles - re-enforces the need to experiment with all your lenses when using exotic glass to gauge what metering effects you get and the compensation required to overcome overxposure problems.

Mike

John, sounds familier with the 5d. I forget how many stops it took to get 2 and 3 stops over exposure, but it did seem to be consistant, at least in the lighting we were in. It wasn't erratic, as soon as we clicked down one more stop, it would over expose by exactly one stop. Stop down one or two stops more, and it would over expose by yet another stop (now two stops) and so on.

Kinda makes me wonder if it's a computer error, like it's trying to calculate for a dim/slow lens that still has yet to be stopped down to it's shooting aperture, and not expecting a lens that is already stopped down to it's shooting aperture? Not quite sure if it's to smart, or not smart enough. :-( The 1Ds2, my old 300d, and the 10d seem to meter no matter what you throw on it, including pinhole setups. I even had it shooting on a 3 foot long box with a small hole at the end of it (a telephoto pinhole camera!) and at 1600 ISO I got fairly accurate metering, within a stop or so.

So, when in doubt, just meter wide open, and calculate. Once you've calculated for a given lighting area, then all you need to do is do an educated guess for your next exposure. chances are, it's going to require the same exposure, or if you see an area that's visibly darker or light, bump it up or down a bit. Don't always have to meter from scratch each time you take a new photo, if you are in similar lighting.

Different bodies behave differently... and at different apertures. From personal experience, the 5D and 20D are off by up to 3 stops at different apertures. My EOS 3 is fine for all apertures of f2 and slower, but overexposes below f2. A little Sekonic Twinmate-II meter is a nice inexpensive accessory... and Sunny-16 is usually much better than any TTL meter outdoors anyway (unless you are shooting gray cards)... so if you bought this wide lens for landscape, don't even worry about it, just go back to old school guess exposure... works great.

Daniel - Did you use the tele with an adapter? As an aside, I've notice the medium format adapters tend to under-expose about 1/2 to 1 stop across the board. They just need added EC most of the time.

Canon increased the 1Ds3 sensor zones from 21 to 63, and I'm not sure it really helped since it feels like one bright point can skew the exposure easier before. This technical difference makes it hard to compare the 1Ds3's behavior to the 1Ds2's.

And here's something to wrap your head around - how will highlight tone priority behave with a manual lens? I haven't tried it yet. If the system is communicating with the lens in some manner to calculate the exposure, then HTP might really bugger-up things.

You ever feel like we spend 1/2 our time trying out-think Canon's thinking?

I always counts the number of stops I'm stopping down and adjust exposure accordingly. Haven't noticed any problems with my 5D with stopped down metering.

the old way to do this is to meter the light when you enter a scenery or room and set the camera accordingly. today you can make a test shot and look at the camera histogramm. once you know what lens at what f stop you want to use you can set the correct exp compensition.

After a while everything gets easier. My 2.8/28 would require a - 2/3 setting between f5.6 and f11.

Am I glad I read this. I've got the 5D and 20D and whenever I use my Oly lenses I over expose. Generally I used compensation based on how far I stop it down and then look at the histogram. Unfortunately, sometimes I can't dial in enough compensation and end up going to manual mode.

Thanks for the information; I'll follow others advice and life should be easier!

The only CZ lens that ever gave me wrong exposures at Av metering was the 35PC, but that is not a surprise on a shift lens. All my other lenses metered correctly. Perhaps there is some "gremlin" in the Canon hardware/software that causes inconsistency between bodies?

With the CZ 35-70 f/3.4 Vario-Sonnar, my 20D underexposes at f/3.4 & f/4, exposes perfectly at f/5.6 (well, as perfectly as the 20D's metering system ever will ...), and overexposes at smaller apertures (I haven't done a systematic test to see how it changes from f/8-16). I need to try the work-around suggested above of metering wide open (or, in my case at f/5.6) and then calculating the manual exposure settings. For now, I have been guesstimating the EC needed, and then adjusting it based on the histogram.

Interestingly, my OM Zuiko 50mm f/1.2 on the same 20D body results in no change in metering performance when stopping down. The OM Zuiko 24/2.8 has only very small variability -- nearly within the range of what I expect from the 20D anyway, so I just check the histogram and make small adjustments as necessary or figure I can fix it in raw processing with that lens. I haven't been using the OM Zuiko 85/2 much lately, so I don't recall what effect that one has.

So you need to experiment with each lens (and with each body, it sounds like) to get an idea what to expect.

The metering issue seems overblown to me. Never have I had a manual lens that caused more than about 1 & 2/3rds of a stop of error. It is easily correctable and becomes predictable very quickly, IMO. One chimp is all I ever need, if at all.

I have the CZ21 and the 5D and I have no problems with over-exposure with stop down metering with std metering. Spot and center weighted metering is sometimes a little off but that is due I believe to my brightscreen split screen.

I put the 5D in matrix metering model and aperture priority, compose and focus wide open, then at the last second, spin the aperture ring to the desired setting before hitting the shutter release. I frequently do have to use the exposure compensation but most of the time no more than 1/3 to 1 stop but this is just to move the histogram left or right depending on what I am trying to do.

The metering system is located in the prism after the focusing screen. The choice of the screen affects metering. The stock 5D screen is optimized for the visual brightness which means that most of the light is directed toward your eye. With higher aperture values the metering system gets disproportionally less light. If the camera "knows" the lens it can compensate and calculate the right exposure. Quite understandably, the camera doesn't know alternative lenses and has no clue about the aperture, and this usually results in overexposure. The Canon 5D EES screen distributes the light much more evenly. After I switched to this screen the exposure inconsistency was reduced to a very manageable level.

Alex

I appreciate all the replies. I'm doing something wrong. It's not working at all. My shots are horribly over exposed and my 1Ds doesn't seem to operate as others have said it should. Here's what's happening with my CZ21 on my 1Ds III.

I put it in Av mode and focus wide open. The 1Ds LED says f/2.8. I just ignore that reading since I know the camera can't recognize the lens.

I stop down to f/8. I've tried to meter with both evaluative and center-weighted.

When in Av I just take the picture. The picture is horribly overexposed. It's even moderately overexposed with a -3 EV comp.

When in M mode I can't adjust the shutterspeed on the camera at all. WTF? The big wheel and top dial just moves the aperture.

I know I'm doing something wrong. Just don't know what. Help!

Update: Not sure what happened but after making my previous post I picked up the camera and flipped the top dial around and voila! The shutterspeed starting moving in M mode. So, I went out side and just based the exposure on the sunny 16 rule and I got perfect exposures. But, the cameras meter is worthless on my 1Ds III. I'm going to need to figure this all out manually or get a light meter.

You are using an AF-confirm chip adapter (which is why the camera thinks there's an f2.8 lens on it).

I'd try a non-AF-confirm adapter and see what happens.

You don't need to "focus" a CZ21. Just set the aperture to f11, then turn the focus ring so the infinity mark lines up with the right-side "f8" distance marker. You'll find that there is a combination that allows everything from 6ft to infinity to be in focus, which may vary slightly depending on adapter thickness, etc. Do some systematic tripod work for an afternoon and get to know her. For landscape work, I would just set it to infinity at f5.6 and push the bloody shutter release for a guaranteed perfectly (frighteningly) sharp photo.

P&S with stunning results!

Sorry for the size, but I think it is necessary in order for you to really understand what it is capable of. This is from a 350D, f5.6.

Holy cow the contrast and detail in that image is marvelous. Of course that's EXACTLY why I bought this lens!

Thanks for the info. I didn't think the Cameraquest adapter was an AF confirm chip though. ??

Isn't that easier than all the variable EV correction?



>>So, I went out side and just based the exposure on the sunny 16 rule and I got perfect exposures.

cogitech wrote:
You are using an AF-confirm chip adapter (which is why the camera thinks there's an f2.8 lens on it). I'd try a non-AF-confirm adapter and see what happens.
I don't have any AF-confirm adapters, all I have are dumb metal ones