Possibly a dumb question about EV & AWB

[groovesection]

Please be gentle with me:
I have only recently delved in to photography, I have a Canon EOS 1100D i purchased 4/5 weeks ago. I purchased a Sigma 18-125 (purely because of the 1.6x crop factor)
I feel i understand the exposure triangle and how these 3 variables apply and affect each other.
But for some reason even when the viewfinder is saying balanced exposure i seem to be over exposing (by at least 2 stops, maybe more)

I am UK based and since i have had my camera there has not really been an blue sky (mostly milky cloud or overcast) and as i mentioned it appears to be over exposing.
So to my question, Is it a metering problem or is it possible the CMOS sensor (being an 1100D) is just poor quality and it will naturally over expose? I can easily set a 2 stop -EV but the problems seems to appear in AV, TV & M modes which makes me suspect it is the sensor/metering.
Another thing i noticed is i shoot raw and AWB, most of the time the images are too read/warm.

By mistake i went out Today and i had set the WB to tungsten on my previous shooting, I did not even notice until i had taken about 30 shots, But the irony is the sky appears to be much crisper (more blue) although shadows are under exposed in a few of the images but most look better than usual using AWB.
The only thing i changed was setting ISO to auto to compliment the AV/TV modes

I am confused, Could anyone impart and share some knowledge, Am i kidding myself i understand the theory of EV or is the 1100D basically a poor camera?

Any advice, links or help would be much appreciated.
(i can add the RAW images if that would help?)

Anton

[xpatUSA]

Hello Anton,

By coincidence, I read yesterday that 3rd-party lenses can become incompatible if the camera manufacturer changes his firmware (which includes metering). It's quite an unlikely cause of your problem but I mention it anyway. The solution was said to be re-chipping the lens! Do you have another lens to compare with?

Are the camera and the lens new? +2 EV seems a lot to my simple mind!

English skies . . . I remember them well

[Manfred M]

Without seeing the images that you are shooting, we are going to have to guess what is going on there. That being said, I think that the missing part of the equation may be your camera's light meter, and understanding how it works. I assume that you are shooting in one of the averaging modes. I don’t know the camera model, but I assume that you are not spot metering (I’m not even sure if your model allows this).

Your camera's lightmeter is something called a reflective light meter, i.e. it measures the light reflecting off an object and it will determine the exposure based on that reading. All reflective light meters work under the assumption that you are shooting an "average" scene, and most of the time this works out quite nicely. If the scene you are shooting as scene that has a lot of sky in it, then the light meter would try to darken it up to make it into an "average" scene. If it is dark, like a night time existing light shot, it will try to brighten up the scene, again to make it "average".

Your tungsten WB will assume a colour temperature around 3000K, while your overcast daylight scene is going to be in the 6500K range, so your shots will have an overall blue cast, including the sky. This has nothing to do with metering.

I strongly doubt that you have a faulty sensor; and even if this were the case, I would be surprised that what you are seeing is a sensor issue. The biggest favour you can do yourself is learning how to use and read your camera’s histogram of a shot and learning to use the exposure compensation controls. My guess is that your camera’s light meter is being fooled by what you are shooting. You have to learn to recognize what is causing this and then correct by using exposure compensation or shooting on manual.

[davidedric]

Hi Anton,

No, the 1100D is not a poor camera. It's a good entry level dslr, and should be more than capable of taking decent landscapes. I don't know the lens, but, again, Sigma make decent quality glass. I suggest that you do post one or two images. It may be helpful to post both the JPEG and RAW (if you have them both). I'm sure that some of the experienced photographers on here can offer a view, with a look at the EXIF's and the images.

[Mike Buckley]

Anton,

It's not possible to overly stress the increased help you'll get if you post a couple of the photos that are typical of the problem you are experiencing, as others have mentioned.

[rpcrowe]

Many, certainly not all, but many, color balance issues should become moot points if you shoot in RAW. The editing software Canon provides allows you to do this as does just about any other editing program. I always shoot my images in RAW using AWB and then correct the images in editing. Don't worry about what the images look like until you correct them.

I have begun using a WhiBal card to get my color balance right However, shooting a white or grey reference card of virtually any type will allow your editing program to bring your color balance within acceptable limits. Along with this, you might need to calibrate your monitor. Speaking of monitor calibration, your monitor might possibly be off!

As far as the exposure, it would certainly help if you posted a couple so we could make some intelligent recommendations.

In what mode are you shooting? How is your exposure set up? What ISO are you using?

I have read more postings regarding exposure problems from new photographers attempting to shoot in the manual mode (which many expert photographers use) than from the new photographers who have elected to use Av, Tv, P or even Auto exposure modes!

[William W]

I have a Canon EOS 1100D [and] a Sigma 18-125 . . . for some reason even when the viewfinder is saying balanced exposure i seem to be over exposing (by at least 2 stops, maybe more) . . . [there has]not really been an blue sky (mostly milky cloud or overcast) and as i mentioned it appears to be over exposing.

So to my question, Is it a metering problem or is it possible the CMOS sensor (being an 1100D) is just poor quality and it will naturally over expose?

Another thing i noticed is i shoot raw and AWB, most of the time the images are too read/warm. . .
I am confused, Could anyone impart and share some knowledge . . .

Run this simple test and post the results and answer question #12.

1. Get all the materials listed below and set the camera ready, before the test.

2. Set AWB - Auto White Balance

3. Set ‘EVALUATIVE’ METERING MODE

4. Set ‘M’ - MANUAL CAMERA MODE

5. Set ISO to ISO 400

6. Set Capture to ‘raw + JPEG (L)’

Find an area of GREEN GRASS – doesn’t matter that the skies are overcast, just ensure the grass is evenly lit with no shadows across it.

Focus on the grass (directly downwards) and set the exposure such that the TTL Meter’s LED in the viewfinder lines up in the centre position – i.e. the camera indicates that it is a ‘correct exposure’.

7. Make a Photograph – it would be good if your aperture is around F/8 and your shutter speed faster than 1/125s – so fiddle to get the shooting details to be around those parameters –your photo should look like this:


Get TWO Dinner Plates or any similar platform – one is to be WHITE and the other one BLACK.
Get some different coloured fruit or other small objects – red and green and yellow is good.

8. Place the two sets of materials side by side in the exact spot you made a Photograph of the Grass

9. Ensure that you MOVE QUICKLY from Step 7 to Step 10 and that the lighting conditions DO NOT CHANGE between the two Photographs you make

10. WITH THE SAME EXPOSURE SETTINGS you used to photograph the grass – make another photograph of the plates of fruit – DO NOT CHANGE THE EXPOSURE SETTINGS, EVEN IF THE CAMERA’s TTL METER INDICATES THAT YOU SHOULD: it is critical for the test that you use the SAME exposure settings for the two images

11. Post the JPEGS of the two images on this thread WITHOUT ANY POST PRODUCTION, except for resizing for in-line web display – the second image should look like this:




12. Answer these TWO question, using the Monitor you view YOUR photos on, to look at my two pictures in this thread -
a) DO the TWO PICTURES ABOVE appear to be too red, too warm?
b) Looking only at the WHITE PLATE - what is the COLOUR of the TINT at the edges, especially the TOP of the WHITE PLATE?

WW

[groovesection]

Hello Anton,

By coincidence, I read yesterday that 3rd-party lenses can become incompatible if the camera manufacturer changes his firmware (which includes metering). It's quite an unlikely cause of your problem but I mention it anyway. The solution was said to be re-chipping the lens! Do you have another lens to compare with?

Are the camera and the lens new? +2 EV seems a lot to my simple mind!

English skies . . . I remember them well

Hi and thanks for the reply Ted,
I have an EOS 1100D (purchased new) and the Sigma 18-125 (purchased new)
The Camera came with the kit 18-55 non IS lens, So i can use that for comparison, But i do remember having the same over exposure problems (i thought it was because i was still learning)

[groovesection]

Without seeing the images that you are shooting, we are going to have to guess what is going on there. That being said, I think that the missing part of the equation may be your camera's light meter, and understanding how it works. I assume that you are shooting in one of the averaging modes. I don’t know the camera model, but I assume that you are not spot metering (I’m not even sure if your model allows this).

Your camera's lightmeter is something called a reflective light meter, i.e. it measures the light reflecting off an object and it will determine the exposure based on that reading. All reflective light meters work under the assumption that you are shooting an "average" scene, and most of the time this works out quite nicely. If the scene you are shooting as scene that has a lot of sky in it, then the light meter would try to darken it up to make it into an "average" scene. If it is dark, like a night time existing light shot, it will try to brighten up the scene, again to make it "average".

Your tungsten WB will assume a colour temperature around 3000K, while your overcast daylight scene is going to be in the 6500K range, so your shots will have an overall blue cast, including the sky. This has nothing to do with metering.

I strongly doubt that you have a faulty sensor; and even if this were the case, I would be surprised that what you are seeing is a sensor issue. The biggest favour you can do yourself is learning how to use and read your camera’s histogram of a shot and learning to use the exposure compensation controls. My guess is that your camera’s light meter is being fooled by what you are shooting. You have to learn to recognize what is causing this and then correct by using exposure compensation or shooting on manual.

Thanks Manfred,
The EOS 1100D does have 3 types of metering.
Evaluative, centre weighted average, and partial.
For some reason my camera was set to partial (instead of the default evaluative)
So this may have been causing the problem (camera by default should have been on evaluative).
I have read up on metering but i guess i need to understand what metering to use in any given situation.

[groovesection]

Hi Anton,

No, the 1100D is not a poor camera. It's a good entry level dslr, and should be more than capable of taking decent landscapes. I don't know the lens, but, again, Sigma make decent quality glass. I suggest that you do post one or two images. It may be helpful to post both the JPEG and RAW (if you have them both). I'm sure that some of the experienced photographers on here can offer a view, with a look at the EXIF's and the images.

Cheers Dave,
I do shoot raw & L-jpg
Here are 2 examples (L-jpg SOOC) ignore the blue colour cast (due to having WB set to tungsten by mistake oops)
I do however think GrumpDiver has nailed it (the metering)
I had it set to partial (from when i first got the camera and was playing with all the options etc) when it is recommended you should use evaluative by default






cr2 raw files





Edit, for some reason tinypic has converted the raw to jpgs.after upload, so i have uploaded them on RapidShare:

https://www.rapidshare.com/files/392...7/IMG_2317.CR2

https://www.rapidshare.com/files/324...4/IMG_2318.CR2

[groovesection]

Many, certainly not all, but many, color balance issues should become moot points if you shoot in RAW. The editing software Canon provides allows you to do this as does just about any other editing program. I always shoot my images in RAW using AWB and then correct the images in editing. Don't worry about what the images look like until you correct them.

I have begun using a WhiBal card to get my color balance right However, shooting a white or grey reference card of virtually any type will allow your editing program to bring your color balance within acceptable limits. Along with this, you might need to calibrate your monitor. Speaking of monitor calibration, your monitor might possibly be off!

As far as the exposure, it would certainly help if you posted a couple so we could make some intelligent recommendations.

In what mode are you shooting? How is your exposure set up? What ISO are you using?

I have read more postings regarding exposure problems from new photographers attempting to shoot in the manual mode (which many expert photographers use) than from the new photographers who have elected to use Av, Tv, P or even Auto exposure modes!

Cheers Richard,
I do shoot raw (and L-Jpg) and i do as you say, usually ignore WB and then correct in post in CameraRaw.
Interesting you mention monitor calibration, That is something i have not yet even thought of or looked in to.
Thanks mate, More reading to do

The images were shot in TV mode using auto ISO.
Most of the time i shoot AV mode or TV mode depending on the subject

[groovesection]

Run this simple test and post the results and answer question #12.

1. Get all the materials listed below and set the camera ready, before the test.

2. Set AWB - Auto White Balance

3. Set ‘EVALUATIVE’ METERING MODE

4. Set ‘M’ - MANUAL CAMERA MODE

5. Set ISO to ISO 400

6. Set Capture to ‘raw + JPEG (L)’

Find an area of GREEN GRASS – doesn’t matter that the skies are overcast, just ensure the grass is evenly lit with no shadows across it.

Focus on the grass (directly downwards) and set the exposure such that the TTL Meter’s LED in the viewfinder lines up in the centre position – i.e. the camera indicates that it is a ‘correct exposure’.

7. Make a Photograph – it would be good if your aperture is around F/8 and your shutter speed faster than 1/125s – so fiddle to get the shooting details to be around those parameters –your photo should look like this:


Get TWO Dinner Plates or any similar platform – one is to be WHITE and the other one BLACK.
Get some different coloured fruit or other small objects – red and green and yellow is good.

8. Place the two sets of materials side by side in the exact spot you made a Photograph of the Grass

9. Ensure that you MOVE QUICKLY from Step 7 to Step 10 and that the lighting conditions DO NOT CHANGE between the two Photographs you make

10. WITH THE SAME EXPOSURE SETTINGS you used to photograph the grass – make another photograph of the plates of fruit – DO NOT CHANGE THE EXPOSURE SETTINGS, EVEN IF THE CAMERA’s TTL METER INDICATES THAT YOU SHOULD: it is critical for the test that you use the SAME exposure settings for the two images

11. Post the JPEGS of the two images on this thread WITHOUT ANY POST PRODUCTION, except for resizing for in-line web display – the second image should look like this:




12. Answer these TWO question, using the Monitor you view YOUR photos on, to look at my two pictures in this thread -
a) DO the TWO PICTURES ABOVE appear to be too red, too warm?
b) Looking only at the WHITE PLATE - what is the COLOUR of the TINT at the edges, especially the TOP of the WHITE PLATE?

WW

Thanks Bill,
I will certainly do that when i get a chance.Probably will not be until the weekend, Thank you though my friend

[Mike Buckley]

What a helpful post, Bill! Why is it that you were never around when I was first learning this stuff?

Anton: Now that I see that you were using spot metering, I'm 99.9% sure that Manfred is right that that is the cause of your problem. When you meter a particular spot in the image, it's very difficult for most people to know in the beginning stage which spot to meter. I've been doing this stuff a long time and, though I may be unusually inadequate when it comes to spot metering, I don't have the skill to use it effectively.

[groovesection]

Thanks Mike,
When you say spot metering, I am assuming you mean using a certain part of the image to be photographed to use for EV?
For example if i was taking a picture of a building (in shadow) contrasted against a bright sky i should be metering the building and not the whole image? Then set my focus/recompose and take the photo?
Or am i really misunderstanding metering?

[William W]

Now that I see that you were using spot metering . . .

Hi Mike - thanks for that kind comment.

Pertaining to the thread - where do you see Spot Metering?
My EXIF reader indicates Centre Weighted Average was used for both images.

I’ve just downloaded the images and checked that with two different readers – definitely EXIF states that CWA was used for both.


WW

[xpatUSA]

. . . if i was taking a picture of a building (in shadow) contrasted against a bright sky i should be metering the building and not the whole image? Then set my focus/recompose and take the photo?

Or am i really misunderstanding metering?

No, Anton, your understanding is very good. For such shots, though, we soon find out the limited range of a camera EV versus the human eyeball. In principle you are correct but, if you expose "correctly" (using either spot, or center-weighted metering) for the shadowed part of a building against a bright sky, it's likely the sky would be blown to hell. There several ways around that problem and a search of these and other fora will reveal all . .

"Correctly" above is in quotes because there are differing opinions as to what "correct" would be in this type of high-contrast scene.

Forgive me for not suggesting a particular solution but forum life is such that, if I posted a solution, there would be some counter-posts and your thread would be effectively high-jacked and become a "the best way to do real-world HDR" discussion . . .

Just as I've high-jacked Mike's probable response to your post

[edit] slightly OT, but I've just read that the Pentax K-30 can do it all for you . . .

http://www.dpreview.com/reviews/pentax-k-30/10

[/edit]

[William W]

Metering Mode funtionality - simarlarities and differences, in Canon EOS Cameras:

‘Centre Weighted Average Metering’ (CWA) is VASTLY DIFFERENT to Spot Metering.
The closest Metering Mode to ‘Spot Metering’ is ‘Partial Metering’.

‘CWA Metering’ is more similar to ‘Evaluative Metering’ is so far as BOTH these Metering Modes take into account the WHOLE SCENE: whereas ‘Spot Metering’ and ‘Partial Metering’, do not.

As already mentioned, the EOS 1100D has only THREE Metering Modes and they are: Evaluative; Partial and Centre Weighted Average.



WW

[groovesection]

Hi Mike - thanks for that kind comment.

Pertaining to the thread - where do you see Spot Metering?
My EXIF reader indicates Centre Weighted Average was used for both images.

I’ve just downloaded the images and checked that with two different readers – definitely EXIF states that CWA was used for both.


WW

Indeed the EXIFF states CWA, But i can categorically confirm both of the images (in this thread) were shot using partial metering.
I have since set the Canon back to evaluative and i really need to read up on the 3 modes offered on my model.

[groovesection]

Metering Mode funtionality - simarlarities and differences, in Canon EOS Cameras:

‘Centre Weighted Average Metering’ (CWA) is VASTLY DIFFERENT to Spot Metering.
The closest Metering Mode to ‘Spot Metering’ is ‘Partial Metering’.

‘CWA Metering’ is more similar to ‘Evaluative Metering’ is so far as BOTH these Metering Modes take into account the WHOLE SCENE: whereas ‘Spot Metering’ and ‘Partial Metering’, do not.

As already mentioned, the EOS 1100D has only THREE Metering Modes and they are: Evaluative; Partial and Centre Weighted Average.



WW

Thanks again Bill for your valuable knowledge It is much appreciated

[Mike Buckley]

I don't know the terminology used by Canon, but Anton mentioned three types of metering: evaluative, centre weighted average, and partial. When he mentioned the latter term, I assumed incorrectly that that was his way of referring to spot metering, as I had never heard of partial metering. Now that I have done a little (actually very little) reading about it, I gather that partial metering is a standard that meters about 9% of the frame and that spot metering meters about 1% to 5% of the frame, depending on the meter.

If all of the above is Canon's view, it's understandable why Anton's camera mentions partial metering and doesn't mention spot metering. It seems as if perhaps his camera does not do spot metering.

The two cameras that I own and the one that is on order use evaluative metering (called matrix metering by Nikon), center-weighted metering and spot metering. None of them mention partial metering. The spot metering in all of them uses a 3.5mm circle, which is about 2.5% of the frame. That distinction is consistent with the apparent difference between partial metering and spot metering.

You're right, Anton, that you should review the specification in your camera's manual to understand what is and is not available to you.