Lightroom Exposure Adjustment vs In-Camera ISO

HeldInTheMoment

No longer a newbie, moving up!
Joined
Jul 27, 2015
Messages
297
Reaction score
33
Location
Vermont, USA
Can others edit my Photos
Photos OK to edit
Quick question, is it better to use higher ISO in the camera when taking a photo or to under expose and compensate in Lightroom by increasing the exposure setting?

Please understand, I am not saying to use low ISO in all shots to correct in Lightroom later, but in theory which would produce a better image with less noise?
 
Depends on the camera. You're shooting a Nikon D7100 in which case you may find there's so little difference as to just not matter or you could find that you prefer the results you can get by avoiding the in-camera ISO increase. In either case with your camera the difference is very slight.

My camera (Fuji) is similar and I often, just out of convenience, skip raising the ISO.

Joe
 
Most of the time underexposing a shot is bad. Exposing correctly in camera is normally preferable
 
Go to dPreview and look at their studio scene comparison test results, and look at the new Nikon D750, D810,and D7200 and compare the results of images deliberately under-exposed and then the testing sample images made by "lifting" the exposures in software by 3,4,5, and even by 6 full EV. The newest Nikons have become almost totally ISO invariant, as dPreview calls it. It is now possible to under-expose in the camera, but to "lift" or "push" the exposures quite far in software, without too much objectionable noise or banding.

The first time I saw this demonstrated, it was a sample image shot with a Pentax APS-C d-slr (Sony sensor), back in 2009 or 2010. The shooter had under-exposed a shot made at the lowest ISO, and accidentally shot using the exposure settings for an ISO of 50,000, or basically, eight full EV of under-exposure. The raw file was opened and the exposure adjusted....the resulting photo was a bit noisy, but was actually pretty decent.
 
Go to dPreview and look at their studio scene comparison test results, and look at the new Nikon D750, D810,and D7200 and compare the results of images deliberately under-exposed and then the testing sample images made by "lifting" the exposures in software by 3,4,5, and even by 6 full EV. The newest Nikons have become almost totally ISO invariant, as dPreview calls it. It is now possible to under-expose in the camera, but to "lift" or "push" the exposures quite far in software, without too much objectionable noise or banding.

The first time I saw this demonstrated, it was a sample image shot with a Pentax APS-C d-slr (Sony sensor), back in 2009 or 2010. The shooter had under-exposed a shot made at the lowest ISO, and accidentally shot using the exposure settings for an ISO of 50,000, or basically, eight full EV of under-exposure. The raw file was opened and the exposure adjusted....the resulting photo was a bit noisy, but was actually pretty decent.

Is this way only ok if you start at a low iso (iso 100), I don't know, but do you get the same result if you push 2 stops from say iso 800 as you would by selecting iso 3200 in camera?
 
Go to dPreview and look at their studio scene comparison test results, and look at the new Nikon D750, D810,and D7200 and compare the results of images deliberately under-exposed and then the testing sample images made by "lifting" the exposures in software by 3,4,5, and even by 6 full EV. The newest Nikons have become almost totally ISO invariant, as dPreview calls it. It is now possible to under-expose in the camera, but to "lift" or "push" the exposures quite far in software, without too much objectionable noise or banding.

The first time I saw this demonstrated, it was a sample image shot with a Pentax APS-C d-slr (Sony sensor), back in 2009 or 2010. The shooter had under-exposed a shot made at the lowest ISO, and accidentally shot using the exposure settings for an ISO of 50,000, or basically, eight full EV of under-exposure. The raw file was opened and the exposure adjusted....the resulting photo was a bit noisy, but was actually pretty decent.

Is this way only ok if you start at a low iso (iso 100), I don't know, but do you get the same result if you push 2 stops from say iso 800 as you would by selecting iso 3200 in camera?

Same result either way.

Joe
 
The same end result is arrived at either way...but while in the field, the preview/review images on the back of the camera are JET BLACK squares when the ISO is left at minimum and the exposure is calculated for ISO 25K or 52K...and any in-camera RAW+JPEG images are pretty much useless.
 
Go to dPreview and look at their studio scene comparison test results, and look at the new Nikon D750, D810,and D7200 and compare the results of images deliberately under-exposed and then the testing sample images made by "lifting" the exposures in software by 3,4,5, and even by 6 full EV. The newest Nikons have become almost totally ISO invariant, as dPreview calls it. It is now possible to under-expose in the camera, but to "lift" or "push" the exposures quite far in software, without too much objectionable noise or banding.

The first time I saw this demonstrated, it was a sample image shot with a Pentax APS-C d-slr (Sony sensor), back in 2009 or 2010. The shooter had under-exposed a shot made at the lowest ISO, and accidentally shot using the exposure settings for an ISO of 50,000, or basically, eight full EV of under-exposure. The raw file was opened and the exposure adjusted....the resulting photo was a bit noisy, but was actually pretty decent.

Is this way only ok if you start at a low iso (iso 100), I don't know, but do you get the same result if you push 2 stops from say iso 800 as you would by selecting iso 3200 in camera?

Derrel mentioned ISO invariant sensors/cameras. This is now a reality with quite a few camera systems. The modern Nikon cameras are of this class. Basically what they've done is cleaned up the electronic processing in the ADC so well that there's no longer any benefit from an analog signal gain -- you can just skip it.

Signal gain is still built into most cameras and when it is it's necessary to use it if you want the camera to generate a decent JPEG. In some cameras they've started to build hybrid systems. The Fuij's are hybrids. They use an analog gain for the first 3 stops of ISO increase and then just digitally scale the image thereafter.

Joe
 
Thanks all, as usual this was a great learning experience!
 
Thanks all, as usual this was a great learning experience!

One more factor that I didn't catch first time around. I was thinking "camera" and hadn't payed close attention to the fact that you also mentioned LR. The raw converter can actually play a fairly substantial roll in the end result depending on the precision (floating point) used in the scaling algorithms. Adobe has made the choice for speed versus precision and so the noise you're going to see will likely be worse from LR versus the camera signal gain whereas you could find the noise from C1 or DCraw is better than the noise from the camera.

Joe
 

Most reactions

New Topics

Back
Top