Lets say there's a reference image with tones we'd like to match. In black and white it's as simple as loading that image into an image processing software and measuring the tonal values.
It's then an easy matter of taking those tonal values and processing a different image where shadow, highlight, and various mid-tones match the reference.
To prove this works, I took an image that I like that was made by Clarence White of Eugene Debs.
As can be seen, this is a Pictorialist era work.
For the exercise I'm embarking on here, there are three things to note from the above photo.
- Tonal placement
- Image softness
- Image tint
For tonal placement I measured portions of this reference image (desaturated to eliminate the tint making it easier to directly note the tonal values). The dark areas raise up off pure black to 2Bhex/43dec. The light tones are suppressed with a maximum value of ADhex/173dec. This is a rather narrow tonal range by modern visual practices/expectations and I find the old Clarence White image to be quite beautiful.
To emulate the sense of softness I used an ISCO Iscotar 50mm f/2.8, racked the focusing ring on the lens to it's closest point, then used a focusing helicoid to effectuate the final focus. The ISCO is a Cooke triplet first element focusing optic. The first element focusing introduces aberrations as it's focused on close subjects. In miniature formats this lens and the effect of the first element focusing is an approximation of how soft focus lenses behaved on larger formats.
Borrowing the tint of the reference image is as easy as opening two images in the Gimp and using the Sample Colorize tool.
As for the subject matter I guess I should apologize. He's not the best of models, but I take what I can get.
Here are the steps used to mimic the reference image of Eugene Debs.
Load the new image into image processing software - enabling demosaicing (of course), auto-select camera color management - turning OFF tone curve - which leaves the image looking very flat.
Using Channel Mix, turn the image into black and white. It looks gawd-awful, right?
Continuing with Channel Mix, emulate a late-1800's emulsion sensitivity filter - 0% red channel, 25% green channel, 75% blue channel. It still looks gawd-awful. I know, there's zero helping the model.
Find the lightest tone and move curves until that tone measures ADhex/173dec. Maybe if one squints real hard things might be looking a bit better.
Find the darkest tone and move the curve until it measures 2Bhex/43dec. OK, now we're beginning to see the full disaster the model brought to imaging.
By inspection, move the bottom of the curve until the dark tones look like the reference and add a slight vignette (if the lens hasn't already done that itself).
Tone the image by sampling the reference and... well... pretty close but no cigar. It's too sharp.
Nice, but too sharp
Comparison to reference
Using the fact that soft focus lens spherical aberration returns a depth of field that anastigmat optics don't, I slightly defocused myself by moving behind the optimal point of focus. In doing this there is still an underlying "sharpness", but diffusion that comes from spherical aberration transforms the images into something that more closely matches the Clarence White, Eugene Debs reference. In fact, if one reads Kodak's instructions on how to use the 12inch Portrait lens they made in the late 1940's it clearly states that focusing on the tip of the nose and letting the depth of field build from there will yield better results. Rather like the Heinrich Kuhn "Tiffenbuilder" - aka: Depth of Field Builder lenses of earlier era and design.
Here are a couple examples of letting the depth of field build itself.
Pretty decent sense of softness, eh?
Now 'er cook'n with gas!
OK. I'm done. Enough is enough already. Again, sorry 'bout the model. Some things just can't be helped.
No comments:
Post a Comment