Looking for lens correction advice. Pictures added

[fin032]

I have took a few pictures of local bridges and even with the Lightroom lens correction profile that was posted on here applied. To me there is something still wrong. If you have a look at the outside of the images the bridges look like there leaning to me and I was wondering what other people thought.

P4P with aperture 5.6

 

[Jim Roof]

The leaning you are seeing is not distortion. It is present in the worlds most perfect lenses and will always happen when a non-shift lens is pointed either up or down.

Tilt the world's most perfect Zeiss wide angle lens up and the vertical lines will start to converge at some point off the top of the page. Tilt the lens down, as you have done here, and the convergence is off the bottom of the page.

The only way to remedy this with the actual capture is with a PC, shift, Tilt/Shift or TS-E lens (all variants of mush the same thing - the ability of the lens to physically shift on the camera body so that the image is composed by shifting the lens instead of tilting the camera up or down.

If you want a perfect capture of architecture, you will need to spend a lot of money on a platform capable of carrying a DSLR and a rather heavy shift or PC lens.

Or... you fix it in Photoshop. Select the entire canvas then use the distort transform to make it look right. Forget the 'persepctive' transform. It's goofy.

 

[Bigbells]

Please just ignore this post if you find it laughable. I thought this is a slight improvement regarding the leaning appearance of the outer bridges in the first photo, but I'm sure that Jim Roof's suggested Photoshop solution would make my effort look stupid. I am not a photographer but I wanted to give it a shot. Using Cyberlink PhotoDirector8, this is an adjustment of +5 for Keystone Vertical and -15 for Fisheye.

The second photo looks great to me as is; I have no impression at all of anything being out of whack.

 

[fin032]

The leaning you are seeing is not distortion. It is present in the worlds most perfect lenses and will always happen when a non-shift lens is pointed either up or down.

Tilt the world's most perfect Zeiss wide angle lens up and the vertical lines will start to converge at some point off the top of the page. Tilt the lens down, as you have done here, and the convergence is off the bottom of the page.

The only way to remedy this with the actual capture is with a PC, shift, Tilt/Shift or TS-E lens (all variants of mush the same thing - the ability of the lens to physically shift on the camera body so that the image is composed by shifting the lens instead of tilting the camera up or down.

If you want a perfect capture of architecture, you will need to spend a lot of money on a platform capable of carrying a DSLR and a rather heavy shift or PC lens.

Or... you fix it in Photoshop. Select the entire canvas then use the distort transform to make it look right. Forget the 'persepctive' transform. It's goofy.

Thanks Jim I will try edit it a little. The reason I was thinking it was a profile problem is the P3P has the official Adobe lens correction profile and that seems to help. No profile out for the P4P yet.

 

[fin032]

Please just ignore this post if you find it laughable. I thought this is a slight improvement regarding the leaning appearance of the outer bridges in the first photo, but I'm sure that Jim Roof's suggested Photoshop solution would make my effort look stupid. I am not a photographer but I wanted to give it a shot. Using Cyberlink PhotoDirector8, this is an adjustment of +5 for Keystone Vertical and -15 for Fisheye.

The second photo looks great to me as is; I have no impression at all of anything being out of whack.

View attachment 80968

You have made a good difference to the far end of the first bridge. I will have a mess around with the Raw file. As for the second image I was looking at the 2 towers of the bridge in the background. 1 is leaning to the left and the other leaning to the right.

 

[Jim Roof]

Thanks Jim I will try edit it a little. The reason I was thinking it was a profile problem is the P3P has the official Adobe lens correction profile and that seems to help. No profile out for the P4P yet.

The P4P has a lot of lens correction that is embedded in the DNG file that ACR does see and respond to. I know this for a fact. There is no need for a lens profile from Adobe unless you want to somehow turn OFF the built in profile. BTW, the built in lens correction profile is very good and it works REAL TIME while capturing video at 60fps in 4K.

 

[GMack]

My personal favorite for perspective correction is DxO OpticsPro 11. Once in the Customize/Develop module, you can address the tile and horizon with the box tool in the upper panel. Just move the slant lines to whatever you want straightened, horizon or items in the edges tilted and hit Apply and it's all fixed. The lens stuff seems to be DxO's forte over Adobe, imho.

Another is RawTherapee 5 (Freeware too!) which can do a lot with a DNG file and has some Transform and Lens Corrections although not as easy to work as DxO Optics 11.

 

[MaxHam]

Please just ignore this post if you find it laughable. I thought this is a slight improvement regarding the leaning appearance of the outer bridges in the first photo, but I'm sure that Jim Roof's suggested Photoshop solution would make my effort look stupid.

It's not stupid and your efforts are for sure appreciated.
Though I think you didn't improve anything - note that the horizon bulges in your sample.

The point addressed by fin032 is not a technical problem but rather a 'mental' one (and Jim Roof explained how to solve it)

Our eye basically works like any camera lens. An image is being projected.
But after light has passed the eye brain kicks in and interprets everything according to our experience.
This is why looking up/down at buildings will not give us the real feel of ratio since the brain knows buildings as being vertical.

Until we take a photo.
Again brain kicks in knowing that photos are 2-dimensional. And suddenly we notice those coinciding/falling lines.

We could say a photo has more truth to it than what we see through our eye.
But now it really gets philosophical

Excuse my German English...

 

[Jim Roof]

It's not stupid and your efforts are for sure appreciated.
Though I think you didn't improve anything - note that the horizon bulges in your sample.

The point addressed by fin032 is not a technical problem but rather a 'mental' one (and Jim Roof explained how to solve it)

Our eye basically works like any camera lens. An image is being projected.
But after light has passed the eye brain kicks in and interprets everything according to our experience.
This is why looking up/down at buildings will not give us the real feel of ratio since the brain knows buildings as being vertical.

Until we take a photo.
Again brain kicks in knowing that photos are 2-dimensional. And suddenly we notice those coinciding/falling lines.

We could say a photo has more truth to it than what we see through our eye.
But now it really gets philosophical

Excuse my German English...

Well stated. Because humans live in a world in which we move horizontally, our brains default to interpreting convergence as the product of horizontal distance.

If we were birds that spent all of our time flying up and down... then we would probably not see it that way and photographs of converging railroad tracks would be seen as odd.