Exposures don't matter featured image

Can you shoot at low ISO and boost exposure in post and get same results as high ISO?

In 5D, 5D3, Photography by planetMitch20 Comments

Jeff Gaunt sent me this interesting set of photos the other day and I just have to get your thoughts – his theory is that you can shoot at low ISO and boost the exposure in post with photoshop instead of shooting with higher ISOs and not end up with more noise.

I’m curious to see what you think… here’s Jeff’s notes:

Exposures don’t matter

From Jeff Gaunt:

I ran an exposure test on a Canon EOS 5D Mark III recently which I thought your readers would find interesting.

Its a different test to what most people do but the result was interesting and I think it will change how people think about exposures whilst shooting night time scenes. I have written up the results of my test below.

Recently I had some time at the end of a shoot to run some exposure tests on my Canon EOS 5D Mark III.

I know many people have done exposure tests but there was test that I haven’t seen but thought was interesting to try.

I have worked on many projects that have involved filming time lapses of the night sky. There is always a balance of wanting to keep the ISO low but still get a beautiful bright exposure of the stars.

I always found in post that I wanted to brighten the shot even more, and in doing so wound up with the same amount of noise in shots no matter what the ISO was.

This lead me to the conclusion that changing the iso on the camera gives the same result as changing the exposure in post.

My test involved shooting exposures ranging from 400 to 6400 ISO in 1 stop increments and leaving all the other settings on the camera the same.

In post I took all the exposures into Photoshop, using the 6400 as my base I increased the exposure on all the other shots to match 6400.

The top half of the image shows the original shots as they were imported into Photoshop. The bottom half of the image is the result of changing the exposures in photoshop.

Click Image for Full Resolution

Click Image for Full Resolution

My conclusion is that there is nothing magical happening inside the camera at a high ISO, this means there is no benefit from running the camera at high iso when shooting at night as you can achieve the exact same result in post from a low iso.

Now I have settled on my ideal camera settings for these situations at an iso of 1600 iso and use a fast f/1.4 lens.

What’s your experience?

Is Jeff crazy? Have you seen the same results? Do you shoot high ISO only to be disappointed?

We’ve got active comments below – join the conversation and let everyone know your thoughts on high ISO and Jeff’s testing.

See more of Jeff Gaunt’s masterpieces, visit his website www.timelapsearena.com/

(cover photo credit: snap from Jeff Gaunt)



planetMitch

chief astronomer at planet5D LLC
Mitch "planetMitch" Aunger is the creator and mastermind behind planet5D.com

He's incredibly happy running planet5D and sharing so much joy of photography and filmmaking with his readers.
18 comments
James 9
James 9

I did this experiment with my old Canon 20d and found the answer was no, definitely do not adjust the exposure in post.  You will add tons of noise.  See here:


https://flic.kr/p/4sV1jE


https://flic.kr/p/4sQVuc


It may be different for the 5d Mk 3, a better test subject would be useful.

storkchen
storkchen

I did the comparison between ISO 100, and 4.5 stop underexposed, then upped the expore 4.5 stops, with ISO 2000, correctly exposed. I used a Canon G1X mark ii (RAW). The differences are dramatic, in the sense that the exposure corrected ISO 100 image has loads of very ugly (esp. chroma) noise, and the ISO 2000 image looked very clean. 

SteveOakley
SteveOakley

I"ve pretty much seen the same results when shooting stills - the noise level is about the same except at 6400 which seems to have a bit more. the differences between 640 and 3200 when shooting stills ? pretty much nothing, same level of noise when you zoom in to 200+% . its more about shutter speed / aperture and if its a good combo for what you are shooting. So I don't worry about cranking up the ISO when I need to for stills. Video is another story as it certainly suffers as you go up.

marcocolnaghi
marcocolnaghi

Just my experience:

- shooting the dark sky the difference is evident when low iso pushes to long exposure and you get undesired star trails...

- shooting people and/or free hand the difference becomes evident when you get motion blur caused by subject or camera shake 

in both situations focal length matters


In conclusion in both cases high ISO is very useful.... 




RFPhotog
RFPhotog

This is well known.  It has to do with the so-called 'ISO-less' sensor.  The signal:noise ratio drops in a linear fashion for each one stop increase in ISO.  Noise is a limiting factor in dynamic range.  For each 1 stop increase in ISO there is a corresponding 1 stop decrease in dynamic range.  DxOMark shows this for the cameras it reviews.  For the 5D MkIII, beyond about ISO 400 this relationship holds.  From 100 to 400, the dropoff isn't linear so it makes sense to increase ISO rather than push in post.  

JePe
JePe

bit difficult to say anything about 'noise', on a picture showing stars... far and in between... Difficult to tell what's noise or a star... Do the same thing again, but then on a person, with a 'normal' background.


bluemoonfilmworks
bluemoonfilmworks

That might be the case in the 5d mark III. Dont think this can be done with a Canon 7d. What do you guys think ?

SiltonBuendia
SiltonBuendia

There is more noise in the lower ISOs then the higher ones and the noise in the higher iso looks a little more pleasing and cleaner. But its impressive that the differences aren't as drastic as you would think. 


Also this works well with this kind of test because there isn't much detail in the shots and thats what shooting at higher ISOs is about it bringing out the details in low light. We also don't have a lot of color in these night shots, I'm sure if we shot something normal like a low light indoor scene and boosted the exposure of the low ISO in post we would have crazy color noise coming it. 


Not to mention the benefit of having higher shutter speeds at higher ISOs, this makes all the difference. 


And when working with video the differences between shooting at different ISOs and bringing up lower ISOs in post is way way more dramatic with most videos terrible color space and bit depth.

JoanVives
JoanVives

pretty usefull for those skeptical about the BMPC 4K, such as me. By the other hand, at high ISOs cameras tend to loose DR, as far as DXOMArks and my knowledge concerns.


TonyNorthrup
TonyNorthrup

Interesting! I've wondered about this. Thanks for the test, Jeff!

My conclusion is different, however. Just looking at his picture, ISO 400 boosted to 6400 in post seems far noisier than native 6400. Look at all those reds and greens! 

It makes sense that you can't get the same quality in post. If the raw file has 10 stops of dynamic range, and ISO 6400 would use the full range, shooting at ISO 400 uses only the bottom 6 stops of the dynamic range. 

AndyZou
AndyZou

For night photography where there is no detail to really be gained or lost in the black of space this may work because it is a really rather low dynamic range scene.  Against a black backdrop, the stars are the ONLY detail.  And it'd be a stretch to get the stars to clip without a significant boost.  For a high dynamic range scene, you're going to end up crushing the blacks and losing the detail without boosting to a proper exposure.

RFPhotog
RFPhotog

@James 9  The 20D and several older Canon models didn't exhibit the 'ISOless' sensor that is more prevalent today.

RFPhotog
RFPhotog

@marcocolnaghi  The point being Marco, that you can underexpose at the lower ISO, push in post and have the same result as if you'd shot at the higher ISO.  You're not exposing normally at the lower ISO, but rather underexposing (i.e., using the same shutter speed you would have at the higher ISO).



storkchen
storkchen

@RFPhotog @storkchenOK thanks. But what is a ISOless sensor, what does it have to do with DxO's DR results, and which cameras do have an ISOless sensor?

storkchen
storkchen

@RFPhotog @storkchenThanks for this detailed and helpful information. (I've read that pushing exposure could still increase chroma noise.)


RFPhotog
RFPhotog

@storkchen @RFPhotog  ISOless sensor is the term given to cameras where there's no benefit to increasing the ISO for better image quality vs. just using a lower ISO and pushing up exposure in editing.  

Cameras that exhibit this show basically a 1 stop decrease in dynamic range for each 1 stop increase in ISO.  So, a straight line - or very nearly.  That's what the DR charts show.  The G1X Mk II, for example, doesn't exhibit this until you get to about ISO 800.  This means that up to that point, there is benefit to increasing the ISO vs. pushing in editing.  If you looked at, say, a Nikon D610 though, you'd see the straight line down starts right at base ISO so you might just as well shoot at ISO 100 and push.  There's a small benefit to increasing ISO up to about 200, but very small.  Hover your cursor over the points on the graph and it will tell you how much DR drops for each 1 stop in ISO.