Exposure - Setting your exposure to -2/3 for most shots

Always shoot -1/3 stop on my D200 because it tends to just give me better exposures. Think the sensor is a bit keen at times. Mind you, that's only for automatic modes - shoot in manual and you don't have any of those problems.

Histogram is also important for me but I generally use it in conjunction with gut feeling as to what i think is right. I usually shoot dark anyway - expose for the highlights, process for the shadow detail and all that - as I like to do a lot of post-production and make sure I always have a lot of highlight detail to work with.

Well after a bit more research I found this: Pretty much blows all I've been taught out of the water but if you look at the bit on the RHS the outcome is pretty much the same.

Never used a grey card in vengeance - never seen the use of them in the decade or so I've been doing it professionally. You use your eye to judge the right shot so why not use it to judge that the exposure is right?..
 
Never used a grey card in vengeance - never seen the use of them in the decade or so I've been doing it professionally. You use your eye to judge the right shot so why not use it to judge that the exposure is right?..

So how exactly would you have done that pre-digital?
 
Well that’s cleared some fog for me, nice find :thumbs:

So in my understanding, digi meters do read at 12% ish reflected grey card, (at a 30 degree angle to ensure reflected light only) but the manufactures have inserted a correction (K) into the output for everyday use .... also reading from an 18% grey card will most often show an underexposure of half a stop (not 2/3 as I said earlier) ... but will give the best overall exposure because of the need to avoid blown highlights and usefully because a digital sensor can collect more data at the bright side of the histogram than the dark side, so little to be lost by boosting it afterwards in PS, but loads to be gained since the highlight information is as high as it can be without being lost....the same highlights wouldn't be so easily lost on film, so no correction is necessary. Hence the reason your Lunasix reads correctly.

Maybe. :thinking: :
 
EdBray, in the studio I used to use a lightmeter or Polaroid back, and outdoors just either taking a general meter reading using the camera, polaroid back on a 5x4" or guessing it. Mind you, I've been digital since 2001 so it has meant I no longer need shares in Polaroid... :)
 
Interesting. And I can only add further confusion, I think ;)

My first reaction was like Forbiddenbiker's reference to exposing to the right of the histogram for maximum dynamic range and minimum noise, and then pulling the brightness down in post. But if I've got my brain around this correctly, that would suggest the opposite to Ed's experience? (See here for reference - http://www.luminous-landscape.com/tutorials/expose-right.shtml )

I have tried some simple tests based on that link and did find that adding a bit of exposure helped very slightly in the final output, but along the way it means that all the JPEGs displayed on the LCD are too bright and therefore not nearly so helpful for generally checking the image immediately after shooting. Not checking for optimum exposure of course (use the histogram for that) but a bleached image is hard to assess pictorially. I digress.

I don't think 18% grey has any automatic right to the tablet-of-stone status it has. Not sure it was strictly true for all films anyway, and surely digital sensors have different characteristics both from film, eg no reciprocity failure, and between themselves.

Before drawing conclusions based on Ed's good work (and sending the camera back with a rude letter :eek: ) it would be useful to eliminate potential errors. What happens in very bright light, in more regular light and in low light? What happens with other lenses? I'm thinking maybe the iris doesn't stop down as it should. I can't think that the shutter would be running anything like that far out, but different speeds would eliminate a fault there. Unless it's just a faulty meter calibration issue, in which case another camera body would sort that. What exposure does a separate handheld meter indicate? Is Canon building in a margin for error? Possibly. But I agree that is an extraordinary thing to do with a 5DMk2 and bad news for those using studio flash and a handheld flash meter.

Just thinking about it now, none of this really stacks up to me? Is it possible that not all sensors/components are the same, and that there's sample variation here? I could believe that might be true, but for Canon to let a camera through QC being 0.7 stops out is unforgivable, unless it's a new camera thing that hasn't been caught yet?

None of which is very helpful. Sorry :( :D

Richard.
 
How does that work then?

You are saying that if I spot meter from a Grey Card, then set the exposure IN MANUAL and take a picture of a black door it will not come out black?

Please enlighten me why?

I apologise, I misunderstood: I thought you were claiming that your camera, in evaluative metering, underexposed a white subject---something that is obviously to be expected with reflective metering. I hadn't noticed you were shooting the white door in manual based on your grey card metering.
 
I have found on my 400D (as mentioned above) that shooting with no exposure correction can lead to burnt out highlights on Av (clouds, faces, in bright sunlight etc).

I always shoot on -2/3 and this tends to correct it, although PP can be used to fine tune individual images.
 
When I have some time I will do some further tests in bright sunlight at infinity focus with different lenses and see what happens.

I do think that Thom's suggestion that the cameras meters are calibrated to approximately 12% grey (ansi standard) as opposed to the 18% grey we thought they were is probably near the mark.

The main thing I have found is that it's all about fully knowing your camera and getting to know how you want to use it and how it will perform when you need it to, but if you are happy with the results that the camera turns out as standard and you don't feel that you need to know all it's ins and outs to enjoy your photography then 'good enough'.

For the majority of people photography is supposed to be a pleasurable pursuit and do they really need to know exactly what makes their cameras tick, for those professionals, semi-professionals and serious amatuers then knowing all your cameras foibles is really a must to assist you to get the best out of it.
 
.....
My first reaction was like Forbiddenbiker's reference to exposing to the right of the histogram for maximum dynamic range and minimum noise, and then pulling the brightness down in post. But if I've got my brain around this correctly, that would suggest the opposite to Ed's experience? (See here for reference - http://www.luminous-landscape.com/tutorials/expose-right.shtml



Actually Richard, I didn't reference exposing to the right. ;) ... As you say that would be the opposite.

I am guestimating, about the reason the digi meters provide an underexposure which will move the highlights to the left, out of blown and into the right side of the histogram where theirs a benefit of the sensor being able to collect a large dynamic range (different than film which is the same all the way across)... so the raw images can be lifted afterwards in post prod without loss of the bright bits info. ...for the same shot there'd be no bright/highlight info if the meter had metered at 18%, they'd be blown out.... but if it had been film (which is where the 18% grey card kinda comes from) then the highlights wouldn’t have been blow due to the slow 'down curve' of the dynamic range at the films range limit….as opposed to a sensor which falls off instantly.
Hope that makes sense.

I'm not really thinking about blow out skies or reflections either, cos that’s just poor exposure ...more those little items, a watch face, shiny buttons, jewellery, white collars etc...those items we don't always spot or think about exposing for, without the lean to underexpose via the meter we'd be blowing bits everywhere.

edit, Shouldn't say we, I mean I would. ;)
 
I do think that Thom's suggestion that the cameras meters are calibrated to approximately 12% grey (ansi standard) as opposed to the 18% grey we thought they were is probably near the mark.

Then I guess that the acid test would be to shoot a 12% grey card, run the image through a calibrated workflow, and produce an accurate 12% grey print without needing any adjustment?

In practise, I find it's very hard to get any closer than within about 0.3 of a stop of optimum - there are so many variables in the system, many of them unkown. During my time testing cameras, a film's ISO frequently varied from that marked, f/stops and T/stops were sometimes not the same, and lens apertures were often innaccurate, especially f/11-f/22 range. But 0.7 of a stop is uncomfortably outside what I would call acceptable tolerance, and I can't explain it away.

On the other hand, what is optimum? Using the expose to the right philosophy and pushing dynamic range right to the limit of any headroom, generally gets me an optimum histogram and a better final print after a bit of post. There's about a stop extra to be had (0.7 maybe?) which is useful, but not a huge amount. However, along the way, it means the camera's LCD image is too bright so in practise this is not the best way of working for me.

I will also say that generally speaking I have found evaluative metering of 'ordinary' subjects gives a remarkably consistent 'good' result, which produces a nice print and a good looking LCD image for working reference, straight out of the box. The histogram usually shows there's a bit of room to the right - around a stop, as I've said, but I'll happily accept that as handy insurance. Blown highlights are ruinous, and what amounts to a slightly higher ISO rating is a welcome bonus.

All of which also suggests that maybe the 12% thing is the answer?

All the above relates to my experience with a Canon 20D. I've not tested my current 40D in this way, as it's never been an issue.

Sorry to ramble ;)

Richard.
 
Crossed post with yours above FB, but what you say does indeed make perfect sense. I think we're agreeing.

I also agree about your reference to accidental highlights - collars and cuffs, even the shiny reflection from a person's cheek bone, that sort of thing is very common and if we actually used spot metering 'for the (real) highlights' then the whole image would be hopelessly dark, underexposed.

Richard.
 
Back
Top