Can someone explain lens cast by aperture to me

Started 4 months ago | Discussions thread
Tons o Glass 0 Class
Tons o Glass 0 Class Contributing Member • Posts: 757
Re: Can someone explain lens cast by aperture to me

fferreres wrote:

Tons o Glass 0 Class wrote:

fferreres wrote:

Tons o Glass 0 Class wrote:

Dropbox, Google Drive? Neither may be ideal but I believe they would work.

That was's the link to the two RAW files:

Thank you! The difference in the thumbnails was drastic. The camera was set to Auto for both of those exposures as opposed to being set to the same white balance.

Yes, the thumbnails was quite drastic!

Here's a mistake I did. I copied the WB from one scene to the other. But didn't pay attention, and when the mode is set to CAMERA, it only copies the mode. Thus, it did not port the Temp and Hue to the other image. When I did now moments ago, they look a lot closer, as you show, and if I brighten the wider one a bit (which is a bit darker), they look more even, but this clips a peak of the red channel, which we don't want, as it would affect color. So I lowered the other a bit instead.

OK thanks for confirming that!

Now the difference is less dramatic, although still noticeable.


I can confirm that Auto WB behaves this way for me as I stop down a lens. I think the Auto WB algorithm may get pushed hot or pushed cold depending on Axial CA and/or PF or the lack thereof.

Hank had guessed this is another post, maybe a year ago. But I couldn't follow much why, in practical terms.

When white balance is held truly constant, however, I think lenses typically do cool want to cool down a bit as we've noticed when you close them down. I've seen camera-based Auto WB push it one way, and have Capture One push it in the opposite direction though... I still have to digest Hank and Alan's discussion.

It's still a mystery how the AWB guesses so differently, but in the camera's favor, now looking at the scene much more closely:

  • The stopped down version happens almost a minute later. The sun-line moves about 4cm. You can see this in the lower part of the bright tree trunk. A branch creates a shadow there, and in the f8ish shot it's a bit higher. This caused the grass on the right side to now not get direct sunlight too.This removes a good amount of direct sunlight, and more of the scene is just whatever is reflected thru leaves and trunks. Don't ask me what it means, but probably has an effect. In addition, that lit part is warmer, and being gone, increases the blue perception.
  • Not sure if this affects it, but this particular lens seems B vs RG optimized. The PSF in the blue channel is out of focus when the green and red are in the best focus. I have described and shown examples of how the peak contrast is achieved when the tiny white spot is sharp but with a deep blue aurea around it. By f8 this halo is gone completely. This may have a direct effect into seeing more bluish things in the wide open shot in the blurred areas, maybe the AWB guessing light to be a bit more blue than it is.

What about lenses with shiny aperture blades - they can sometimes cause a loss in contrast as the aperture closes down. Depending on the scene, the white balance an algorithm or our eyes would want to see could get pushed hot or cold if warm or cool colors are bounced around by the aperture blades, respectively.

In this case, I guess it would be minimal. But veiling profile may affect it. I think I guessed this in the thread where Hank commented about AWB.

I popped these into Capture One and let it decide the white balance. It wanted to warm up the f/8-ish shot, and wanted to cool down the f/1.8 shot, but there was still a sizeable gap between the two shots in that regard.

The curves are a bit different. In the wide open one there is less relative blue. The grass that is more lit in one and not that much in another may account for some.

FWIW here's the two shots where I took C1's Auto WB on the f/8-ish shot to be "bible" and set both shots to that, then haphazardly and imperfectly pushed the f/8-ish shot such that the histograms roughly matched.

Exposure was brought down an additional (subractional?) 0.7EV from as shot. The stopped down shot undoubtedly had more contrast and the sunlight also seemed more intense here? I dunno. Maybe the camera didn't want to clip the shadows.

Hmm... which one's cooler now? This one may need to be a little warmer, actually, which is consistent with what we've observed already.

The second one is cooler. But now, I had to make a lot more effort and consider it a guess.

I see that -1.7 EV was dialed when you were shooting. Was it your intention to expose for the sunlight in doing that?

I didn't have time and had ISO 100 dialed, but the f8ish shot would be very blurry as it was already 1/50 wide open, and you see how the light moved in 47 seconds, so I didn't have time. But more importantly, if I increased the ISO, this camera doesn't have dual gain, so I would not gain much vs pushing in post.

The other factor, and this was the first to consider, see the curves. There's a peak of red to the absolute right. If I didn't expose to the right, I would have blown the not much of the highlights, but would have blown the read channel in the main subject that is the trunk and small tree. The surrounding had to look very dark, and the tree shining. I think this photo would display better in an HDR monitor with 10.000:1 contrast. The small tree was magically glowing. And the rest was really really dark. I wanted to keep that in the photo.

I see what you were going for and your edits surely make for more interesting shots. I agree with you regarding HDR - the f/8-ish scene through your Rollei looks to have more dynamic range than the A7 II could comfortably capture in a single image.

Note that the more one brighten the scene, the less and less the small tree fased into a secondary character. If you dial the exposure down, you will see the small tree as main subject.

My eyes are drawn to it first, that's for sure.

Would have been better if it wasn't there. As what really got my attention was that the small tree was the only thing with direct sunlight... plus that trunk.

Yep I think we're speaking the same language here. Those elements caught my eye.

Thanks for investigating this in more detail. In large part, it was AWB. With still some difference, but nothing that would change the character of the photo that much. There's still more red in the wide open one, consistent with Hank's comment. But much less dramatic.

You're welcome - thanks for posting the files to play with and take a look at. Curious to see what the definitive explanation will be.

 Tons o Glass 0 Class's gear list:Tons o Glass 0 Class's gear list
Canon EOS 550D Sony Alpha NEX-3N A3000 Sony a7 II YI M1 +2 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow