Under Exposure problem

[mayhem]

Hello,

I am hoping someone can help me here, I am experiencing a problem with my D70, in that whenever I take a shot of a car from head on with it's headlights on the result is a massively under exposed shot with just the headlights visible. I should mention that these are shots taken in perfect light, and with Shutter priority mode on (at around 1/320ths or so). The lens is a crappy Tamron 300mm lens (Nikon 200mm f2.8 on it's way thankfully), but I feel the problem is in the camera as my mate with same body but 300mm Nikon lens experiences the same problem.

Am I doing something wrong, is there a setting I am overlooking, or am I a victim of a lens that should be no more than an ugly paperweight?

[Tim Falce]

You probably have the meter set for centre weighted or spot reading, try using matrix (I think thats what it's called) or the one that takes a reading from an average of the whole area.
There was another thread on this a few months ago but I can't find it.

[924nut]

I think that it was myself that asked the same question
The problem is the sensor will always react to the bright light source and the only way round it is to switch to a manual mode. take your readings without a car, and then use them in manual mosw- then the camera wont react and change the settings as the bright lights come into view.
Hope that helps
NB I use a Tamron lens - and it got me as one of the runners up in the 2005 Motorsport News photo competition so it can do the business - yes its not as fast as an original pro lens but it isnt a bad lens.

[forzaf1]

Your always going to get problems with the exposure if the car has the lights on coming towards you because the headlights confuse the sensor.


This was taken at the LMES at Silverstone, granted it was a wet day but this was taken fairly early on in the afternoon, before it started to get dark.

Alex

[aerog]

Originally Posted by mayhem

and with Shutter priority mode on (at around 1/320ths or so).

And that's the problem. As has been said the bright lights are influencing the meter to an extent that everything else is being underexposed to compensate.

In very simple terms (and this isn't completely true in the case of the D70 and other new cameras anymore) the light meters in cameras try to make everything gray. With a center-weighted meter try taking a picture of a completely object and you'll notice the image will tend to be overexposed - because the meter is trying to make the black "lighter". Conversely try taking a picture of a bright-white object and you'll notice the image will tend to be underexposed - because the meter is trying to make the white "darker".

The modern camera's (starting with the Nikon FA I think) various matrix-metering modes try to avoid that problem by sampling different areas of the image rather than just concentrating on the center, but the "problem" still exists when faced with extremes (sometimes). The headlights probably are enough of an extreme in brightness that the camera just decides to underexpose everything else so you don't lose detail in the headlights. You've given control of the final exposure to the camera and you see the results.

Ok, so there's a complicated explanation of the problem. The solution? Simple: take back control of the camera and shoot in MANUAL. This is a generalized answer by the way, with lots of exceptions: take a picture of the track with no cars on it and look at it. Everything looks good, right? Well shoot at that exposure setting (in manual) with cars in it and you're likely to have a decent final exposure without the adverse influence of the cars. This technique also removes the influence of the car's brightness (color) from the final exposure setting. That's not always desireable, but usually it works out. It also helps to shoot RAW rather than jpg so you can make a broader change in the final image's density and contrast.

Here's a sample I shot in November. I used matrix-metering but that really doesn't matter. My baseline exposure was set by metering off part of the track area and underexposing by 1/3 stop. The image was taken in the late afternoon in deep cloud shadows with very little direct sunlight:

(click for larger-resolution image)



This was a similar situation in reverse. I had plenty of bright highlights in the car, but it was generally in shadows (as was the track). The camera was tending to add more exposure than I liked, resulting in blown highlights...so I did my own thing in "manual" with it:

(click for larger-resolution image)


Here's a white-car/black-car situation. Same general area on the track, two different lighting conditions, two different days. Matrix metering should examine the whole scene and try to set things up for something decent, but as these were originally almost full-frame images I think it would concentrate on the cars and probably under or over-expose the images, but the overall image exposure ended up the same:

(click for larger-resolution image)


(click for larger-resolution image)


Good luck - Scott

[Hubble]

Or you could see if your camera has an exposure lock function. Press the exp lock button just before the car you want to shoot comes into view, hold it down, compose, shoot.....it's as easy as that. Ideal for those of us who prefer to use TV more often than Manual. Good luck

[Tim Falce]

Originally Posted by aerog

Lots of stuff

I was just wondering if we could go back to basics and use a good old light meter with an incident reading. I have a Weston Meter that you put a cone on and take a reading of the light falling on the subject not the reflected light.
Would this work with digi cameras and also cars that could be up to a couple of hundred feet away?

[vs346]

You can revisit some of the viewpoints discussed on this forum before here:

http://tentenths.com/forum/showthread.php?t=76150

[vwpilot]

Originally Posted by falcemob

I was just wondering if we could go back to basics and use a good old light meter with an incident reading. I have a Weston Meter that you put a cone on and take a reading of the light falling on the subject not the reflected light.
Would this work with digi cameras and also cars that could be up to a couple of hundred feet away?

Using an incident meter should work as well as long as the light falling on it is the same as that falling on your subject.

You might want to do a test or two with it first though and see how it affects your specific camera, digital slrs have tended to underexpose a bit in my experience compared to film, likely to help keep the highlight detail that with digital, once its gone its gone. So you may want to test so you know how the reading on your meter comares with the output you get from you camera.

As to the original question. Here is an experiment to help you understand how your cameras light meter works.

Look through it at a normal scene and note the exposure. Now shine a light into the lens and see how its changing, that is what is happening when you shoot headlights head on. Now put a white sheet of paper in front of the camera and notice the change in exposure, now a black sheet, notice the change.

The meter is reading light reflected off subject and going into the lens. When you shine lights at it, it thinks there is a lot more light than there actually is and underexposes. Same with something white, it reflects more light, the camera thinks there is more light than there is and underexposes (very common issue when shooting in snow) and likewise black reflects less, camera thinks there is less light than normal and overexposes.

Best control is manual or try to not shoot when the lights are directly at the lens. If they are just a bit off, usually there is no issue, its when they are shining right at you that you will have problems.

[MikeHoyer]

It's pretty much already been covered, but manual exposure is the best way to solve this problem. You can sometimes correct it in photoshop, but can often create very artificial and grainy looking images.

I tend to think about what is more important from the shutter speed or aperture, then set the other one up for that... for example, if I'm doing a panning shot I would set the shutter speed to 1/250 or something, and then set the aperture to whatever gets me the right exposure. If it was head-on, particularly against a background of trees or crowd I might consider the aperture and DOF more important, and have a fast shutter (depending on the light conditions).

But generally to get my exposure I meter off the tarmac, and set up my aperture and shutter speed accordingly.

[mayhem]

Thankyou one and all, your help is very much appreciated.

[thebear]

Originally Posted by falcemob

I was just wondering if we could go back to basics and use a good old light meter with an incident reading. I have a Weston Meter that you put a cone on and take a reading of the light falling on the subject not the reflected light.
Would this work with digi cameras and also cars that could be up to a couple of hundred feet away?

You would have to set the desired ISO (film speed) on the camera and then add ~20% when you set the speed on the meter as it is probably calibrated in the old "ASA" numbers.

Testing, testing, testing. At least with a digital camera you can see your results immediately and the "film" is free.

[aerog]

Originally Posted by thebear

You would have to set the desired ISO (film speed) on the camera and then add ~20% when you set the speed on the meter as it is probably calibrated in the old "ASA" numbers.

Huh? They're the same, aren't they? The tech sheets I have on films all say "ISO/ASA" and refer to them as being equal.

Anyway, I've used my Spotmeter-V with several digital cameras and gray cards, and the "film speeds" all work out to be the same (ISO on the camera and ASA on the meter).

[vs346]

ISO 100 = ASA 100

Both refer to sensitivity of the image capturing film and/or digital CCD, just different (updated) terminology.

[thebear]

Originally Posted by aerog

Huh? They're the same, aren't they? The tech sheets I have on films all say "ISO/ASA" and refer to them as being equal.

Anyway, I've used my Spotmeter-V with several digital cameras and gray cards, and the "film speeds" all work out to be the same (ISO on the camera and ASA on the meter).

Originally Posted by falcemob

I was just wondering if we could go back to basics and use a good old light meter with an incident reading.

The old was an indication to me that it was from the days prior to harmonization between ASA & ISO standards. I didn't consider modern meters.

[bella]

my solution is as per a couple of others - learn to take pictures manually. you'll find that you'll learn a lot quicker and that just by using your brain a little instead of the cameras you get better results than it can come up with

the in-built light meter is usually fine, personally with my d500 i prefer to underexpose compared to what it tells me, but that's to do with the camera/lens combination.

also you might find that your lens is better than you think it is if you can master the art of manual exposure and focus and working around its issues

[mayhem]

The lens is probably better than I give it (public) credit, but there have been a few times when I have missed an incident on track as the lens 'searches' for a focus point. Manual Focus is great, I know because if nothing else this lens has forced me to learn how to use it, but sometimes I would just like a lens that worked a bit faster.

Thanks all for your advice on this one, your replies have made for great reading.

[Tim Falce]

The focus speed is more down to your camera than the lens. Don't forget that at 300mm there is wide focus range and I would expect the motor in the D70 isn't as quick as say the D1 or D2. Also the lens is probably something like an f4.5 to 5.6 maximum aperture which doesn't aid quick focusing. As for image quality there probably isn't that much difference between the Tamron and a Nikon unless you are going to make poster size prints. The only way round this is to buy one of the new Nikon motorised lenses and get one with a wider aperture although they are very expensive.
You could stick with manual focus and save some money, I have a Canon A1 and can't remember ever having a motor racing shot out of focus but have missed plenty of shots with various auto focus camera while they sat thinking about life the universe and whether they were in focus.

[aerog]

Originally Posted by falcemob

The focus speed is more down to your camera than the lens. Don't forget that at 300mm there is wide focus range and I would expect the motor in the D70 isn't as quick as say the D1 or D2.

You're right, on all counts... but just for information's sake on the D70/300mm combo: I shoot with a D70 (latest firmware) and Nikkor 300mm f/4D ED-IF AF-S (one of the faster, motorized lenses you spoke of), and use a Nikkor TC14E-II teleconverter to bring it up to a 420mm/f5.6 lens. The D70 with that lens combo, even with the TC14, will do a nice continuous focus of every car I've shot head-on at Daytona on the infield-track...including the GTP cars. The TC14 does slow it down a tad but it is still very useable and keeps up except in really dismal conditions. The D2-series cameras are supposed to be far superior so I've been pleasantly surprised.

Even if you're stuck with a slow-focus lens the D70 does have options though, like what people have termed "trap focus". In "trap focus" you pre-focus on an area then track the cars normally. The shutter doesn't release until the car is in focus. This technique removes any of the mechanical delays in the focusing, especially with the older non-AFS lenses.