ISO Invariance Overstated - ISO settings do make a difference

Started 4 months ago | Discussions thread
ForumParentFirstPrevious
Flat view
Karl_Guttag Senior Member • Posts: 1,400
ISO Invariance Overstated - ISO settings do make a difference
6

I've been "corrected" a couple of times (one example: (example https://www.dpreview.com/forums/post/65856454) by some people that believe that setting ISO makes no difference on the R5/R6 above ISO 800 due to "ISO Invariance."

I've seen cited photonstophoto charts (R5: https://www.photonstophotos.net/Charts/PDR_Shadow.htm#Canon%20EOS%20R5) showing that noise does not vary when pushing ISO pasts IS0800. Tony Northrup has several videos on ISO invariance (from 2019: https://www.youtube.com/watch?v=QVuI89YWAsw&t=0s)

So I decided to check it out with my R5 with the RF100f2.8 macro.  I took pictures in cRAW and processed them in DPP4 and DxO PL5 with DeepPrime. In the comparison below, I shot with the same camera setting other than ISO.  For the examples below, I use IS0800 and ISO 12,8000, which are 4 stops apart. In comparing the processed images, I found them to be about 3 and 2/3rd apart (I ended up having -0.3ev on the ISO12,800 images).

My findings:

  • There is slightly more noise in boosting using ISO800 and post-processing ~+4. But as the ISO Invariant "advocates" would suggest, it is not a huge difference. I noticed more noise on some colors on the test chart than others. 
  • There seems to be a significant loss of detail in the darker areas of the chart when boosting ISO. Look at the pebble effect on the black case for the X-Rite color checker. The ISO800 boost pictures seem to blur out this area. The pebbles in the case are many pixels wide, so the loss seems pretty significant.
  • DxO with DeepPrime does a much better job reducing noise while improving resolution. 
  • With some other tests, I found that even overexposing by +1ev does not recover well with the R5. A several stop underexposed shot is "saveable" where even 1 stop overexposed does serious damage the image. 

My conclusion is that using "ISO invariance" and locking the ISO at say 800 in darkly lit situations and "saving" in while workable for many people is not the best way to shoot a photo. You will be giving up some resolution and contrast in the darker areas of the image. It is also clear that it is better to underexpose rather than overexpose.

I should also mention that if you want to try ISO invariance shooting, you should disable "exposure simulation," or the camera will not be able to focus well (focusing on the R series cameras is tied to the image in the viewfinder/LCD).  You will also lose the ability to judge photos on the back of the camera as they will all be very dark.

All in all, I think most people would be much better off NOT using ISO Invariance (and Locking the ISO). But it is a good idea to err on the lower exposure side in tricky light situations.

 Karl_Guttag's gear list:Karl_Guttag's gear list
Canon EOS R5 Canon RF 15-35mm F2.8L IS USM Canon RF 24-70mm F2.8L IS USM Canon RF 70-200mm F2.8L IS USM Canon RF 24-240mm F4-6.3 +13 more
Canon EOS R5
If you believe there are incorrect tags, please send us this post using our feedback form.
ForumParentFirstPrevious
Flat view
Post (hide subjects) Posted by
ForumParentFirstPrevious
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow