Photos with different ISO and aperture

Comments

3 comments

  • Avatar
    Götz Echtenacher

    Hi Allan,

    did you also shoot with a fixed focus?

    Because if not, then the auto focus will probably affect image geometry much more than aperture, which, if I'm not totally mistaken, doesn't change much apart from amount of light and DOF. ISO will definitely not affect geometry at all, only show a bit more noise in dark areas - on a beast like the A7R3 not even noticeable outside lab conditions. So you should be fine imho.

  • Avatar
    Allan Adams

    Fantastic.

    We shot with autofocus, but the object was big enough/far enough away that we made sure everything was in focus for each shot, so I'm not worried about DOF. 

    I was just wondering if changing ISO would create a different noise profile (and therefore not be able to match features) or if changing aperture would cause very small (unexpected) lens distortions. 

     

    Thank you!

  • Avatar
    Götz Echtenacher

    Hi Allan,

    I think to judge that better a few sample images and also to know your intended precision would be essential.

    The difference in noise between 640 and 800 is academic at best, so if your object doesn't sport loads of underexposed areas with important fine detail in it, I think it won't be a problem, RC will definitely be able to match enough features if the object is suitable otherwise.

    Do you know this site yet? https://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=sony_a7riii&attr13_1=sony_a7riii&attr13_2=sony_a7riii&attr13_3=sony_a7riii&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=400&attr16_1=800&attr16_2=1600&attr16_3=3200&attr126_0=1&attr126_1=1&attr126_2=1&attr126_3=1&normalization=full&widget=1&x=0.1812215132178669&y=0.48854685377999096 I think it only starts to show noise worth mentioning at ISO 3200.

    RC cannot calculate geometry down to a pixel, so if your variations are smaller than a few pixels, you should be fine in any case. Also, you could just run a last alignment with ungrouped images - I do that always and depending on the object and image set this can make quite a difference because RC can then fine tune the variation in image geometry caused by the auto focus.

    But as always, own experimentation is always the best - so why don't you make a few comparative studies with only a few dozen images in each set and compare them?

Please sign in to leave a comment.