400D underexposure. Please instruct idiot in how to check

Mascot

Forum Enthusiast
Messages
383
Reaction score
4
Location
NO
From what I've been able to gather from the umpteen threads on the topic (sorry for starting yet another one!), serious underexposure has in at least one instance been shown to be a faulty camera.

It's my first dslr so chances are I wouldn't have noticed this for ages myself, so far I've done little but drool over the IQ and increased low light performance compared to my P&S. But, since the issue is out there, can someone give some instructions on how I check for it from a complete dslr idiot's perspective?

Consumer protection laws around here gives me 14 days to return the bugger with zero questions asked, which I'd rather be able to do if it turns out to be defective (saves having to wait around for repairs, or argue for a replacement). But I'd be much more happy if I have nothing to worry about, and the few outdoor shots I've been able to take seem ok to my untrained eye. It's been raining continuosly since I got the camera though, so I've had little opportunity to really use it.

I don't have an 18% greycard. I do have walls in various colors and lighting conditions. From what I figure a well lit 18% should give a histogram smack in the center. Which doesn't help me much since I don't have one. I do have walls in various colors and lighting conditions, that all end up left shifted on the histogram unless I add some EC or flash.

I'm rambling. The question I'm trying to pose is this. Can I make use of what I have to do this test with a reasonable degree of accuracy, or do I have to hunt around for an 18% grey (the local store didn't have such a thing)?

Please do not make this thread into yet another arguement over whether the issue is real or not, whether people should just compensate or not, etc. There's plenty of threads for that. Just tell us newcomers how to test for a faulty light sensor.
 
If you haven't noticed that it is underexposing why would you bother?
--

 
Why bother? Because if the camera has a fault I'd rather have it corrected now than having to jump through hoops to get it corrected down the line.

Just because I'm too much of an amateur at this point to notice doesn't mean I should ignore it only to have it come back to haunt me when it's too late to get it fixed at no cost.
 
I saw that thread earlier, but didn't think it had all the info I needed.

Does this mean I can pretty much find any solid colored wall, regardless of lightsource, and the histogram should be centered?
 
for the XT/350D, the firmware fix come one after another.
and canon is good with that, maybe tomorrow it is in the jap website.
who knows ...

alien :)
Why bother? Because if the camera has a fault I'd rather have it
corrected now than having to jump through hoops to get it corrected
down the line.

Just because I'm too much of an amateur at this point to notice
doesn't mean I should ignore it only to have it come back to haunt
me when it's too late to get it fixed at no cost.
 
for the XT/350D, the firmware fix come one after another.
and canon is good with that, maybe tomorrow it is in the jap website.
who knows ...
From what I've read, someone returned it and got a better unit. That seems to indicate that some units have a hardware fault, not a software one.

If if is indeed a software fault, or just indoor shots, I don't see any reason to fuss about it myself. It's not knowing whether it's hardware or software that bothers me.
 
Okay, here are three screenshots from the zoombrowser showing the histograms. Going from 0 to +2/3 EC. They're of a white closet door in a room lit by a bulb.





 
your new girlfriend is better than the old one, or
else why you dump the old one, or keep the new one.

I trust canon's QC ....

I only believe sigma has QC problem, not canon.

sorry to offend sigma fans.

alien :)
for the XT/350D, the firmware fix come one after another.
and canon is good with that, maybe tomorrow it is in the jap website.
who knows ...
From what I've read, someone returned it and got a better unit.
That seems to indicate that some units have a hardware fault, not a
software one.

If if is indeed a software fault, or just indoor shots, I don't see
any reason to fuss about it myself. It's not knowing whether it's
hardware or software that bothers me.
 
Okay, here are three screenshots from the zoombrowser showing the
histograms. Going from 0 to +2/3 EC. They're of a white closet door
in a room lit by a bulb.
What I make of this test, please correct me someone if I am wrong, but this looks like underexposure. You are shooting a blank wall, with uniform texture, therefore the camera should get it right in the first place without any exposure compensation. In the first shot, the histogram is shifted on the left, in case of correct metering, it should have been in the centre. Once you compensated then it is properly exposed, but this shouldn't have happened in the first place had the camera metered the scene correctly with respect to its sensor.
 
Instead of looking for a problem, go out and take lots of photos. If after you print them and they show consistent underexposure, then you may have a problem.
Why bother? Because if the camera has a fault I'd rather have it
corrected now than having to jump through hoops to get it corrected
down the line.

Just because I'm too much of an amateur at this point to notice
doesn't mean I should ignore it only to have it come back to haunt
me when it's too late to get it fixed at no cost.
--

 
I'm not looking for a problem for the sake of it, I'm trying to verify or exclude the presence of a known issue that affects some units.

Waiting until it's too late to get it sorted at no cost doesn't seem like a very sound recommendation to me.
 
My results are almost identical. Plain white wall, at 0EC the histogram is shifted left quite a bit, it is only centered with +2/3EC ...

This means that I should leave my camera set to +2/3EC at all times? That's fine with me, but doesn't that mean that my camera will use slower shutter speeds for the same scene than the 350D did, all other things being equal? The 350D didn't seem to have a problem with underexposure, and if I'm interpreting this test correctly, my new 400D does ... To get a properly exposed picture I will have to leave the +2/3EC enabled, right?
Okay, here are three screenshots from the zoombrowser showing the
histograms. Going from 0 to +2/3 EC. They're of a white closet door
in a room lit by a bulb.





 
My results are almost identical. Plain white wall, at 0EC the
histogram is shifted left quite a bit, it is only centered with
+2/3EC ...

This means that I should leave my camera set to +2/3EC at all
times?
It seems to me that the amount of underexposure is inconsistent with the 400D so you probably won't need to leave it at +2/3 EC all the time. You'll want to take a test photo under using the lighting you'll be shooting in, look at the histogram and image to see if it's exposed properly and then set you camera accordingly. You might find that 0 EC is needed in brighter lighting and +1/3 to + 2/3 EC is needed in lower lighting.
That's fine with me, but doesn't that mean that my camera
will use slower shutter speeds for the same scene than the 350D
did, all other things being equal?
Yes, if the 350D didn't need any + EC adjustment, this is correct.
The 350D didn't seem to have a
problem with underexposure, and if I'm interpreting this test
correctly, my new 400D does ... To get a properly exposed picture I
will have to leave the +2/3EC enabled, right?
Again, see above.

Michael
 
From what I've been able to gather from the umpteen threads on the
topic (sorry for starting yet another one!), serious underexposure
has in at least one instance been shown to be a faulty camera.
I wouldn't worry about it. The bottom line is the 400D is less sensitive
than the 350D, which is usually about 1/3 stop "hot" out of the box.
Just use the histogram and your exposures will be fine(expose to the
right).

Regardless of the metering scheme, sensitivity can't be created out of
thin air. I use EC all the time with my XT and it doesn't bother me a bit.
If I owned the 400D, I would do the same.

Rob
--
'Don't sneak up on it - surround it'
 
I just tried both partial and center weighted avg ... At all EC levels, partial metering showed a very slight right-shift, and center-weighted metering showed a slightly higher right-shift. In other words, CWA metering was closer to accurate at 0EC although it was still left of center ...
I'm almost convinced that it's the "Evaluative" metering. I would
appreciate if someone can repeat this test in centre weighted and
partial.

Ilias

--
http://s20.photobucket.com/albums/b207/iliask
 

Keyboard shortcuts

Back
Top