Question about exposure correction in RAW

[Jon_Are]

I'm not exactly sure how to phrase this question, but I'll try:

Say I have two lenses that I am using to shoot in a relatively low-light setting (indoors during the day, for example). Lens A is a 3.5 aperture, Lens B is 1.8.

So I first shoot with the 1.8 and find that I can attain correct exposure with a shutter speed of 60. This tells me, of course, that my 3.5 will be inadequate for this shot. (I know I can bump up the ISO, but that is not relevant to my question)

Suppose I pop the 3.5 onto the body and take an identical shot using the same settings (except using 3.5 instead of 1.8). So now I have my shot and it is underexposed.

Because I shot this in RAW, I open my editor and correct the exposure in this image.

My question is this: Will my exposure-corrected 3.5 image be of the same quality as my 1.8 image which was shot with the correct exposure to begin with? If not, will the difference be obvious, or very slight?

Thanks,

Jon

 

[Big Mike]

Why not try it yourself and see?

When you increase the 'exposure' in post processing, the quality certainly gets worse. This is often where you get problems with digital noise as well.

Sure, you do have some freedom to adjust the exposure slider...but it's up to you to decide how much adjustment and quality loss is acceptable in your images.

 

[Dao]

I am not an expert on this. But this is what I had in mind. Please correct me if I am wrong.


First, I will look at the histogram. If the histogram of the underexposed shot shift too much to the left to a point that a lot of blackout shadow, you cannot correct it via any software since details are lost. As a result, the wider aperture shots will look better. How much better depends on the amount of blackout shadow.

 

[rufus5150]

It's very much dependent on the quality of your image rendering when you're underexposed. This varies camera to camera.

I've found that shooting ISO 100 with my XTi, I can underexpose a shot by about 3/4ths of a stop and sometimes a bit more and have an image that is workable some of the time. I can routinely (and have to routinely) bump an image 1/3rd of a stop or so for exposure correction. I have once gotten a workable image where I had to bump the exposure nearly 2 stops. Most of the time the results are trash bin material.

It will be noisy and you'll have to deal with that. But 1 to 3.5 is 2 stops? that's going to be rough.

 

[Moon Baby]

There is a noticeable image quality shift when adjusting raw files 1-2stops. It also depends on how much shadows are present. I have a 5D and even adjusting raws by one stop can leave some very noticeable noise.

 

[Big Mike]

It's usually best to have more exposure than you need (without clipping the highlights) and then adjust down.

Expose Right

 

[dcclark]

"Underexpose" is an imprecise word -- it just means that your exposure is darker than you would have liked.

It is possible that you may have severely underexposed, in which case the shadows in your image may have become fully black. No amount of exposure tweaking in ACR (or some other raw converter) will be able to bring back the detail which you lost in those shadows.

On the other hand, you may have just been darker than you wanted, but no shadows were lost. In this case, you can tweak the exposure and probably make things look good.

There are exact parallels for hilights and overexposure.

So, the answer is -- it depends on the scene and light which you are photographing. But I wouldn't count on it! If you underexpose (or heck, even if you expose as well as possible), there is still a good chance that you may have some clipping, especially in a high dynamic-range scene.

 

[GeneralBenson]

The digital sensor record more data in the brither tones than it does in the darker tones. So when you underexpose and try to lighten an image, you have less information, han if the image correctly exposed. That being said, at low ISO I can easily pull a full stop or more out of my Pentax K10 or K20, and At higher ISO's, can pull up to a stop.

 

[JerryPH]

As a general rule you always want to expose properly in camera. Relying on correcting for a huge error (yeah, 1-1.5 stops is considered a pretty huge error... lol), and you are going to increase the noise in the darker areas to a point that it will be easily visible to the naked eye, and not at pixel-peeper level either.


I can pull 3 stops out, it will just look like crap. In the end, just because one can do something doesn't mean this is the best way.

 

[Jon_Are]

Why not try it yourself and see?

I will. But this discussion gave me more insight into the subject than just a test would, as I anticipated.

As a general rule you always want to expose properly in camera.

Yeah, I expected as much. (and this is the bottom-line answer I was in search of)

Jon

 

[Dwig]

"Pushing" the exposure in post will yield results similar to having used a higher ISO when you take the picture. How similar will depend on a variety of factors including the flexibility of the RAW converter and the skill you have.

If you make the real test, which you should, inclulde a shot at the higher ISO as well as the two proposed in your Gedankenversuch. Shot a set of 3 shots:

1: f/2.0 @ 1/60 @ ISO 200 - adjusting lighting to make this the "correct" exposure
2: f/4.0 @ 1/60 @ ISO 200 - 2 stops under exposed

3: f/4.0 @ 1/60 @ ISO 800 - correct exposer again.

Use the same lens for all 3 to avoid lens specific differences in exposure and contrast. Also shoot RAW for all 3. Convert the RAW images for #1 and #3 using the same basic settings; the converter's defaults are a good choice. The one thing you probably want to override is the noise reduction. Set this to zero reduction for both luminance and chroma.

Now convert #2 making you own adjustments to match #1 and #3 as closely as you can. You'll probably need to boost the "exposure" and, most likely, reduce the color saturation. Don't make any changes in any noise (leave as zero) or sharpness (leave the same default used for #1 & #2).

You'll find there's no such thing as a free lunch. "Pushing" the exosure in post will, at best, match the results of using the higher ISO in camera.

 

[Jon_Are]

Use the same lens for all 3...

Well, the reason I originally asked the question was to determine if I could get comparable results using my 3.5 zoom (and adjusting exposure in PP) as opposed to switching lenses to the 1.8 prime. I was trying to get away with being lazy.

As I suspected, it seems the quality would not be as good. The key, I'm learning, is to get it (the exposure) right the first time.

Jon

 

[Dwig]

Well, the reason I originally asked the question was to determine if I could get comparable results using my 3.5 zoom (and adjusting exposure in PP)...

Fine, but if you do the test use the same lens. Only then will the test reveal the differences in using PP to "push" the exposure relative to just using a higher ISO. Switching lenses will add other differences to the image and confuse the comparison.

In general, a common "kit" zoom will be 1/2-2/3 stops slower at a marked f/3.5 than the common 50mm f/1.8. They will also have different flare levels and contrast.

 

[Garbz]

Ok asside from a few of the things that have been mentioned such as clipped shadows, and less dynamic detail in the black range rather than the white range. There is also the issue of noise.

Noise is the same regardless of lens and fast exposures have no appreciable difference either. It is always distributed around a zero point. Thus if you have noise in the shadows you amplify this noise when you bump up the exposure.
Digital exposure pushing in cameras like the D200 which have a native maximum ISO of ISO1600 but can be pushed to HI1.0 or ISO3200 show this very well, the ISO3200 shot being of horrible noise quality compared to the ISO1600, a very big difference than the ISO800-1600 change which involves more than just boosting the brightness in camera.

 

[PixFixer]

No! Just as a reminder, the digital camera fundamentally works the same as a film camera. The big difference is, changing ISO is like changing film midstream. That chip in the back of the camera is like a different kind of film. So, if you use different "f" stops and the same speed you under or over expose. If you change aperture and speed to compensate, the length of exposure will have an effect on the final image. See, same as film. How much depends on the camera. But, it won't be exactly the same. Your tolerance is what counts.
Arnie