3.7 Traditional and Digital Darkroom🔗⧉

3.7.1 The traditional darkroom🔗⧉

A photograph is only half made when the shutter closes. For most of photography's history the other half happened in a darkroom, and it helps to know, both because film workflows are enjoying a revival and because nearly every tool in today's editing software is a darkroom operation in disguise.

The chemical pipeline has two stages. First the exposed film is developed: bathed in a developer that turns the exposed silver-halide grains to metallic silver, a stop bath that halts the reaction, and a fixer that dissolves away the unexposed grains so the image is no longer light-sensitive. The result is a negative, dark where the scene was bright. Then comes printing: an enlarger projects the negative onto a sheet of light-sensitive photographic paper, which is developed in its turn to give a positive print.

Printing is where the craft lives, because the enlarger is a second exposure that the printer fully controls:

• the exposure time under the enlarger sets the print's overall lightness, the global brightness knob;
• the contrast grade (the paper's grade, or a filter over the enlarger) sets how steeply scene tones map to print tones, the analog of a contrast or tone-curve control;
• dodging holds light back from a region (a hand or a paddle waved over it during the exposure) to make it lighter, while burning adds extra light to a region to make it darker: local tonal control, by hand.

This is the literal origin of the dodge and burn tools in Photoshop, of the exposure and contrast sliders everywhere, and of Ansel Adams's Zone System (the metering section's 18% gray is its Zone V), a discipline for planning, at capture time, exactly where each scene tone will land on the print. The darkroom is also the concrete meaning of the book's recurring metaphor that the print is a performance of the negative (the score and the performance, and the measurement chapter's sidebar): two printers, one negative, two different photographs.

3.7.2 The digital darkroom🔗⧉

Today that darkroom is software, and it comes in two flavors worth telling apart, because they embody two different philosophies of editing (Figure 3.7.2).

Parametric raw "developers" — the Lightroom style. Adobe Lightroom, Capture One, Apple Photos, and the open-source darktable and RawTherapee are, at heart, raw developers wrapped in a library. They are non-destructive and parametric: the original file is never altered; every edit is stored as a small list of instructions applied on top of it, re-rendered live and reversible at any moment, the digital analog of deciding how to print a negative. The toolset is mostly global adjustments: white balance, exposure, contrast, highlights / shadows / whites / blacks, the tone curve, HSL (per-color hue/saturation/luminance), clarity / texture / dehaze, sharpening and noise reduction, lens corrections (distortion, vignetting, chromatic aberration), and crop / straighten, plus local adjustments painted in with masks, graduated and radial filters, and increasingly AI-driven subject and sky masks. The other half of the job is asset management: importing, rating, keywording, and searching tens of thousands of photos. This is the everyday workflow for most photographers, applied across whole shoots at once.

Pixel editors — the Photoshop style. Adobe Photoshop (and GIMP, Affinity Photo) work the other way: they manipulate the pixels directly, with layers, selections, and masks for compositing, and a deep toolbox for retouching: the clone stamp and healing brush to remove blemishes and distractions, content-aware fill to erase objects, dodge / burn tools (there they are again), curves and levels as adjustment layers, filters, text, and montage. This is where you do pixel-level surgery, combine several frames into one, and composite, the things a parametric developer is not built for.

In practice most serious workflows use both: develop and cull in a Lightroom-style tool, then hand the few keepers to Photoshop for heavy retouching or compositing. The dividing line is global, reversible, many-photos versus local, pixel-level, one-photo.

And here is the bridge into the rest of this book: every one of those tools is an algorithm, and we are going to study how each one works. The exposure and contrast sliders are point operations and tone curves (BASIC → Point operations, Tone mapping); the histogram you push for ETTR gets its own chapter (Histograms); white balance is color constancy (Color technology); sharpening, blur, and clarity are convolution and edge-preserving filtering (Neighborhood operations; Bilateral filtering); noise reduction is denoising; raw development is demosaicking; crop and straighten are resampling; content-aware fill and seamless compositing are Poisson editing and seam optimization (EDGES MATTER); and the newest "generative fill" and AI masking are diffusion models and learned segmentation (COMPUTATIONAL TOOLS). The slider is the interface; the rest of the book is the machinery underneath it.

This chapter has walked photography end to end: exposure, modes, focus, lenses and their filters, keeping the glass clean, stabilization, the viewfinder, video, the two anatomies that book-end the field (a full-frame body where good optics do the work, and a phone where computation does), the wider taxonomy of camera types and the many cameras that measure rather than photograph, flash and lighting, and finally the post-capture darkroom, chemical and then digital. That last step is the hinge into the rest of the book: the phone is the machine the book is really about, and editing software is where its output is finished, and everything from the develop slider to content-aware fill is an algorithm we are about to study. The next parts build the pipeline (image representation, the ISP, multi-frame fusion, and the editing operations themselves) that lets a fingernail-sized sensor punch far above its physics.