Good Noise

A key characteristic I have been trying to get in digital image processing that I remember fondly from film developing is nicely organic film grain. Something like the so called “salt and pepper” look of Tri-X would be ideal.

The following shot was taken at sunset outside a dive bar down the street. The first version gives you a feel for the sunset through the color of the reflected light. The second is a black and white processed version without added “grain” simulation, but with careful control of noise reduction, sharpening and tone mapping/filtering to convert the colors to black and white. Click through to see the full-size image files.

20130407-103950.jpg
I used out of camera color rendition as a starting point, then adjusted exposure to retain highlights while bringing up the shadows a bit and enhancing midrange contrast. Then I put it through the Agfa slide film setting in Alien Skin Exposure to adjust the colors for the final version, and Topaz deNoise rather than the in-camera NR to clean up chroma noise without excessive smearing.

I also used Alien Skin Technical Pan simulation as the basis for monochrome conversion, with a blue color filter setting and a little sharpening, but with the added film grain default deactivated.

20130407-104124.jpg

In the end, on a large display, both versions look somewhat film-like to my eyes, but came out of a compact digital camera. I probably wouldn’t print these much more than 8″x8″ for close-up viewing. At that size, probably using optical RC paper for the color, and maybe a baryta fiber print for the black and white would create/recreate the look I have been searching for.

Here are some close ups comparing the “grain” from the upper left corner bokeh area in the picture above, to actual scanned Tri-X 400 push processed at ISO 800.

 

 

The first comparison image is with Tri-X 400 pushed to ISO 800 with the exposure being in bright light outside. The second is the same processing, but with an indoor image in low light. The X20 shot was in relatively low light just before sunset.

The file size/resolution of both are about the same, and if you dive too deep, pixelation gets in the way of the comparison. However, at 100%, I think what I am seeing is vaguely similar, with the big differences being the random bigger specks of black in the X20 image that don’t seem to occur in real film, maybe also some NR related smearing from the X20 (can’t get rid of it totally), and possibly some sharpening effects on both since the film scanning software also seems to leave a bit of a signature. From the X20, I tend to see a distinctive “waffle” pattern show up just before getting to the pixel level that seems to be related to post-processing.

 

Leave a comment