January 8, 2011
While many people debate whether Ansel Adams was a great photographer there is no denial he was a great printmaker. His attention to detail in the printmaking process brought a criticality to the fine art photography world and his concepts are important, even today. I’ve decided to give a quick rundown of the processing process I’ve started to use when shooting black & white film and exporting to the web. The image above is titled “Fresh Glass” and was shot in Calgary, Canada this week. I used a Mamiya RB67 Pro-S with a 90mm Sekkor C lens and developed it in Ilfosol S for Ilford’s recommended time – 15% at 68F.
This is an untouched, 2400 dpi 48 bit colour scan using an Epson V700. Of course this has been scaled down for the web. I scan my b&w using epson scan software because I like the workflow in Epsonscan for my b&w images. During the scan I use the levels and the densitometer. I scan to min/max 0/255 and I use the densitometer and the levels curve to make sure no shadows or hilights are clipped or blown out. I scan as a 48 bit colour TIFF then open it with photoshop. One of my first steps is to use use the green channel and drop the others.
The channel mixer is located in Image -> Adjustments -> Channel Mixer. The reason why I use the green channel is because most flatbed scanners (Epson V700 included) are sharpest in the green channel. The red channel is a little sharp and the blue channel is quite blurry. Open a 48 bit colour scan of b&w film and view each channel separately and you’ll see what I mean. You have to first click “Monochrome” before making your channel adjustments. After the channels are mixed I convert the image to 16 bit greyscale.
This image was my first attempt and was sharpened at the beginning and I also upped the microcontrast by applying an 80 pixel, 10% USM.
I apply curves and levels adjustments using Adjustment Layers. Adjustment Layers allow you to edit your photograph in a lossless manner. It preserves the quality of the image and makes a large difference when 4-5 adjustments are being made on a region of the photograph. I didn’t spend enough time with the image on my first attempt and it’s obvious from the final results.
For the second attempt I performed the usual channel mixing and greyscale conversion then jumped straight into a perspective correction using Filters -> Distort -> Lens Correction. I then cropped out the keystoning of the edges and started serious work on the photograph. I then began work on individual regions of the photograph. You can create a selection using the masking tool then convert the selection into an Adjustment Layer so that tweaks can be made to a specific region of the photograph. For foliage and other “natural” elements I use the brush but for the linear hard edges of buildings I use the polygonal lasso feathered 4 pixels and anti-aliased. After I have a selection I save it by going Select -> Save Selection. You can then load the selection in the future instead of having to re-do the work.
Step by step I worked through regions of the photograph, starting with the large building in the front. I darkened the black-point of the levels and increased the contrast slightly with the curves. I made other adjustments to other building then finally adjusted the foreground. This was a tough photograph because the foreground was shadowed by buildings behind me while the tall buildings were being hit by direct sunlight. There are regions with 14+ stops of dynamic range and cutting the development time down by 15% helped keep the tonality of the image.
I prefer to do my dust remove after applying USM. I perform USM in two steps. First I do a 2 pixel, 100%, 3 level sharpness adjustment. Then I will apply a “microcontrast” adjustment by unsharp masking at 80 pixels, 10% and 6 levels. This gives the details a little bit more “oomph”.
I used to go straight into dust removal but now I leave it until after the microcontrast adjustment. There are two reasons to this. The first is that dust removal takes forever and I don’t want to do it unless I feel that the image really has something there that’s worth sharing. It saves me a lot of time if I leave it until later. The second reason is that increasing the microcontrast with an unsharp mask will emphasize the dust removal process and also emphasize dust and blemishes that I missed. If I remove the dust after tweaking the sharpness and microcontrast I’ll catch more grunge.
Once the image is ready I perform a resize to my required size. If there’s a lot of linear detail that will cause a “screen door” effect in the resize I perform a bicubic softer. Otherwise I use the default bicubic resize. I follow this up with a 0.2 pixel, 80-100%, zero level unsharp mask to get a bit of the detail back lost in the resize.
The final, and important step is a level adjustment for the web. I scan 0-255 so that I use as much of the 16 bit dynamic range as possible but web browsers don’t display blacks deeper than 10 properly. I apply a final level adjustment, increasing the output black level to 10 then Export to Web as a level 80 jpeg. I don’t bother with a compression quality higher than 80 because the human eye doesn’t really notice a difference and it can double the file size in some cases. Some people are stuck behind 80 kbyte/sec connections and if I have twelve 600k files to load they’re going to close the page before they see the content.
So after about 2 hours are spent on the image I get something much more satisfying than 20-40 minutes spent. I hope there are a few people that find this helpful.