Whats the big deal w/ possible underexposure ?

[sportster64]

so what if the camera underexposes ?

you set the Exposure Compensation to +1/3 or whatever it needs - it then uses a slower shutter speed or wider aperature.

If it was exposing properly in the first place - it would have STILL USED a slower shutter/APerature anyhow Right ?
what am I missing here

 

[onlyonejc]

Compensating for exposure isn't a true capture of what the lens is seeing. It is the camera's guess at what it thinks will be captured. So, in turn, it brings out additional noise in the final image.

Metering to true (or flat) exposure levels, means that the lens is allowing enough light in to do so. This means that the camera doesn't have to do any guess work, and the image should[ i] be perfectly exposed; thus, minimising noise any excess noise.

onlyone-jc.

 

[magicmikey]

Compensating for exposure isn't a true capture of what the lens is
seeing. It is the camera's guess at what it thinks will be
captured. So, in turn, it brings out additional noise in the final
image.

Metering to true (or flat) exposure levels, means that the lens is
allowing enough light in to do so. This means that the camera
doesn't have to do any guess work, and the image should[ i] be
perfectly exposed; thus, minimising noise any excess noise.

onlyone-jc.

What? I'm not sure what you are trying to say. If you compensate to get the "correct" exposure (which is subjective), you won't see an increase in noise. You'll see an increase in noise if you underexpose the shot or you use a higher ISO to capture the shot.

If the sensor is less sensitive in the XTi/400D, as many suspect, that means you'll have to let in more light than indicated to get a correct exposure. The biggest loss is that of low light sensitivity, where an XT/350D might be able to use a higher shutter speed or smaller f/stop. (Smaller f/stop meaning a larger number.)

Michael

 

[JohnAugust]

There's a couple of concerns here:

1. Since the 400d is an "entry-level" camera generally targed to those making the p&s > DSLR transition, there's a higher expectation for the camera to perform well in full auto modes. To those making the transition, a slightly overexposed photo with blown highlights might look more attractive than a conservatively exposed photo shifted far left. Forcing users to use P mode so they can make EC adjustments just to get a proper exposure in moderate light conditions may be a big deal to some people.

2. Compared to the 20d and 350d, the 400d supposedly underexposes by 1/3 to a full stop. In otherwords, the 20d and 350d expose correctly (or sometimes overexpose) at the same aperture and shutterspeed as the 400d without the EC adjustment which may imply that the 400d is slightly less sensitive. For those who are upgrading from the 350d to 400d, the loss in sensitivity (or however else you'd prefer to explain it) may be a big deal.

3. The underexposure issue seems to be inconsistent. Some people are reporting perfect exposure in their 400ds while others are reporting frequent underexposure. Since "perfect exposure" can be somewhat subjective, it's difficult to distinguish differing notions of perfect exposure from poor quailty control. If these inconsistencies in exposure are in fact quality control problems, some people might consider that a big deal.

4. The general attitude on these forums exacerbates the issue to the level it is now. Those who recently dropped a lot of money on a camera that should perform better than their previous p&s or DSLR are instantly chastised when they turn to these forums to voice their concerns. Instead of educating those people about proper exposure, we sometimes respond with offensive answers like "go back to your point and shoot!" or "my 400d exposes perfectly so you must not know how to take photos".

5. It's a new camera so it will be thoroughly scrutinized until the next new camera comes along.

 

[onlyonejc]

I'm pretty sure the compensation is applied to the 'underexposed' image initially after its capture before it is written to the card, and so any existent noise in the image will therefore become boosted.

onlyone-jc.

 

[PeaceFrog]

While I agree that it isn’t a show-stopper, it is annoying to always have the camera set to +1/3.

Greg

--



http://www.pbase.com/dadas115/

 

[Sabz]

The issue is not simple underexposure by a fixed amount. It is exposure CONSISTENCY (sorry for shouting it out).

If you are taking photographs at an event, the last thing that you want to do is to spend your time editing every single one of them for a different exposure amount (can't batch process). In fact I shoot JPEG+RAW so that I can get away by using quite a few of my JPEGs without edit (get it right the first time, what a concept).

Now, having to correct the exposure significantly will increase the noise especially if you had to shoot at higher ISO's. I for one choose Canon because of its lower noise (trust me, it wasn't the fancy body or pentaprism).

At my last event I set the EV to + 1/3 and still had to correct the exposure by 1/3 to 1-1/3.

And finally what is the point of having a metering system if you are better off guessing the exposure.

 

[Ricouk]

The problem IMHO is not the metering system, as it would seem
the 350D and 400D give the same shutter speed and aparture value
for those that have directly compared the two, so the only explanation
would seem to be inconsistency in the sesnsitivty of the sensor.

Some 400D owners sensors are ok... others are not, there IS a
problem with these cameras.

Cheers, Rico.

 

[Tom Long]

Exposure compensation will change the aperture and/or shutter speed. You can test that by taking pictures of the same scene with and without compensation. So if you use positive exposure compensation for a dark photo you will reduce noise.

I'm pretty sure the compensation is applied to the 'underexposed'
image initially after its capture before it is written to the card,
and so any existent noise in the image will therefore become
boosted.

onlyone-jc.

 

[lukewarm]

by 1/3 to a full stop. In otherwords, the 20d and 350d expose
correctly (or sometimes overexpose) at the same aperture and
shutterspeed as the 400d without the EC adjustment which may imply
that the 400d is slightly less sensitive. For those who are

Do you have any proof that the 400D underexposes at the same aperture, shutter, and ISO manual settings as the 350D?

From the (admittedly) few samples I've seen comparing the two, manual exposures on the 400D and 350D are almost identical under the same settings. There are some minor differences, mostly due to the different tone curves and the 400D's slightly higher dynamic range.

But so far I haven't seen any solid proof that clearly shows the 400D sensor being less sensitive than the 350D's. Image samples of 400D underexposure I've been seeing are nothing more than demonstrations of the 400D's very conservative metering.

 

[Smitty72]

i agree, i'm a newbie and i don't post much, but i can see why some get frustrated by all those "underexposure" posts and such. if anyone can tell me what in life if perfect i would love to hear about it. just my .02, but i don love this forum...a wealth of info
--
http://smitty72.smugmug.com

 

[lukewarm]

I'm pretty sure the compensation is applied to the 'underexposed'
image initially after its capture before it is written to the card,
and so any existent noise in the image will therefore become
boosted.

There's a big, big difference between adjusting exposure settings before taking a photo, and adjusting exposure after the photo is taken. Basic photography 101.

If you don't understand exposure, then you probably don't have a full grasp of the 400D's alleged underexposure "problems" either.

 

[starstreak]

You hit it right on. I dont think the camera is broken, although I think some here on this forum may have a camera that doent work right. Mine seems ok, but I always have the EV set about one click to the right.

Correct me if I'm wrong, but EV compensation is software right? So too much EV will cause some sort of noise?

 

[Katania]

From the (admittedly) few samples I've seen comparing the two,
manual exposures on the 400D and 350D are almost identical under
the same settings. There are some minor differences, mostly due to
the different tone curves and the 400D's slightly higher dynamic
range.

As a matter of fact, one of the complaints with the 350D was that it was overexposing while that really wasn't the case. With the default presets, it was using a higher contrast setting which had the tendency to easily blow highlights.

It's possible that the default settings for the 400D are more on the conservative side. Increasing contrast might help.

 

[slydog]

I did a controlled test several times.

• Indoors, constant lighting
• Tripod, Mirror lockup, Timer
• Manual Mode
• Same Aperture, shutter speed on both XT and XTi

The XT shot was visibly brighter and the XTi histogram shifted to the left. This was consistent. It was easily reproducable.

Are you asking because you want to help, or because you don't believe?

 

[dave_s93]

I'm pretty sure the compensation is applied to the 'underexposed'
image initially after its capture before it is written to the card,
and so any existent noise in the image will therefore become
boosted.

No, no, no. That is not how exposure compensation works. On DSLR's, the EC changes either the aperture or shutterspeed (depending on your mode) to expose the image more (or less for neg compensation). It is not applied afterwards.

 

[starstreak]

I think I meant EC not EV.

 

[dave_s93]

Correct me if I'm wrong, but EV compensation is software right? So
too much EV will cause some sort of noise?

No, the EV on the camera changes either the shutterspeed or the aperture (depending on your mode). It works just like the EV setting on film cameras.

 

[JohnAugust]

I can keep posting proof until I turn blue in the face but the skeptics will continue to deny it.

Just for you, I did one last test. Note the linear tone curves.
http://forums.dpreview.com/forums/read.asp?forum=1031&message=20548088

 

[JohnAugust]

no text