A Test on "exposing to the right" and ISO

I though this might be useful to people out there who absolutely hate going to higher ISO's in low light conditions like I did.

I did my own test and I thought I would share the results. I was pretty surprised by them. I took 3 shots at the same shutter speed and aperture at 100, 800, and 1600 ISO. The 800 ISO image was correctly exposed so I adjusted the exposure of the 100 and 1600 ISO images in lightroom to match the 800 ISO images histogram. Take a look at the comparisons with a large crop.

The first image is 100 ISO, the second is 800 ISO and the last is 1600 ISO

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.

1600

HOSTED PHOTO
please log in to view hosted photos in full size.

Pushing that ISO 100 image three whole stops exaggerates the effect somewhat, but yes this does show why exposing to the right is important.

It also covers those "why is my camera so noisy at low ISO" questions for people who underexpose and then push it (often with auto adjusting software) in processing.

JeffreyG wrote in post #5578357
Pushing that ISO 100 image three whole stops exaggerates the effect somewhat, but yes this does show why exposing to the right is important.

It also covers those "why is my camera so noisy at low ISO" questions for people who underexpose and then push it (often with auto adjusting software) in processing.

I suppose I am exaggerating a little bit, but the main point is to show that if you expose to the right with 1600 or even 3200 ISO you can get some really nice results compared to trying to fudge it with a lower ISO.

I'm surprised the 100ISO image looks cleaner then the 1600ISO image..what camera did you use?

yogestee wrote in post #5578391
I'm surprised the 100ISO image looks cleaner then the 1600ISO image..what camera did you use?

Do you mean the 1600 ISO looks cleaner than the 100? I used a 30d

a_kraker99 wrote in post #5578402
Do you mean the 1600 ISO looks cleaner than the 100? I used a 30d

To my eyes the ISO800 and ISO1600 look about the same, probably due to pulling that 1600 back 1 stop.

The ISO100 looks a lot worse that the other two after being pushed three stops.

This really does work. I find myself especially getting careful with exposure when ISO1600 is marginal. In that case it is almost always better to shoot at ISO3200 and pull back the exposure 2/3 stop than to shoot ISO1600 and push it 1/3 stop (as an example).

I can only ass-u-me that you were trying to prove that exposure errors can be fixed in PP.

On my monitor:
ISO 100 is very noisy, expected when you boost the blacks.
ISO 800 is a little underexposed, but ok anyway.
ISO 1600 seems "washed out"

[QUOTE=chauncey;557847​9]I can only ass-u-me that you were trying to prove that exposure errors can be fixed in PP.
quote]

No, I am trying to show that you are better off using a higher ISO and exposing to the right rather than using a lower ISO and boosting the exposure later. The washed out look can be changed with a simple contrast or black level adjustment, however noise is much more of a pain to get rid of.

Here is a different example. These look about the same to me.
1st photo is 400 ISO pushed up 1 stop, 2nd photo is 1600 ISO pulled back 1 stop

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.

Interesting comparison, but pushing 100 ISO three stops might be overdoing it.

JeffreyG wrote in post #5578438
This really does work. I find myself especially getting careful with exposure when ISO1600 is marginal. In that case it is almost always better to shoot at ISO3200 and pull back the exposure 2/3 stop than to shoot ISO1600 and push it 1/3 stop (as an example).

Jeffrey, this doesn't make a lot of sense to me. Unless one has a 1D3, ISO 3200 is exposed at ISO 1600 and software boosted. IOW, it is 'exposing to the left'. To do two s/w adjustments, the second partially cancelling the first, rather than one, seems illogical - although possible quantization losses in 12 or 14 bit would probably be invisible in all but the darkest shadows. However, I know you are an experienced photographer and your observations carry some weight. Enough so that I will have to test this myself. I can say that in tests I did some time ago with the 40D, comparing ISO 1600 underexposed and pushed one stop in DPP and ISO 3200 RAW, I could see no signifant difference even at 200% zoom. Still, maybe a different converter might have given a different result.

From a neophyte that didn't realize ISO 1600 existed:

I've recently learned that you cannot boost exposure with PP, noise, and lowering exposure seems counterproductive as you've already blown out highlights.
Or are we talking about images that are without blinkies at the right to begin with?

With my landscape images and my foolhardy attempts to do HDR from one RAW image, I'm rarely able to move exposure down more than 1 1/2 stops
and don't have the room to boost it as chimping and exposing to the right is a big part of my workflow.

So, how much PP exposure alteration do you'll feel comfortable with, and how much in-camera exposure compensation is appropriate?

99, those last two images look identical to me.

Generally, as ISO increases, dynamic range decreases. If all you did to the ISO 1600 shot was reduce exposure of the RAW file then there might be something more to the lack of saturation in that image compared to the proper exposure at 800. To me the ISO 1600 picture (from the original set) looks a little softer than the ISO 800.

e-k

E-K, wouldnt the fact that most of an images dynamic range is stored in the right side of the histogram make the image look a little washed out when bringing the exposure back down? Since so much information is packed on the right hand side, when you bring the exposure down wouldnt it be normal for there to be a little less contrast?