Layering, color grading, texture blending, and digital painting transform captured images into something new. This category documents my evolution with post-processing tools from basic Lightroom adjustments to complex multi-layer Photoshop compositions. It is the largest collection on the site — 115 images — and the one that raises the most questions, because a fair number of them aren't photographs in any strict sense. They are creative expressions that treat a photograph as raw material.
The photo is only the first layer
Almost everything here begins as a single ordinary frame. A sky I shot over Shenzhen Bay, a wall of weathered shutters in Huaqiangbei, a portrait made with the Canon EOS 5D Mark II and the Canon EF 85mm f/1.8 USM — none of these is precious on its own. What makes an image worth building is a quality I can pull out of it: a texture, a color cast, a shape that suggests a scene that was never there.
From there I work in Lightroom Classic first, almost always the same way. I straighten and crop, correct white balance, set exposure and contrast, then pull Highlights down and push Shadows up to flatten the tonal range the way a canvas wants. Only after the base grade feels calm do I open the file in Photoshop; if I start compositing on a badly exposed frame, every layer above it inherits the problem.
Layers, masks, and blend modes
A composite is patience expressed as a layer stack. A new element — a moon, a flock of birds, a second sky — arrives as its own layer with a black layer mask, and I paint white only where I want it to appear. Blend modes do the invisible work: Screen to lighten and merge a sky, Multiply to deepen a shadow, Soft Light to seat a texture into the grain without it looking pasted on.
I work at full resolution and non-destructively, which means a finished piece can easily run to thirty or forty layers. Adjustment layers for every tonal move, masks instead of erasing, and a plain-English name on each layer even when it stays hidden. Three months later I can't remember why a layer exists, so "sky_darken_left" beats "Layer 27 copy copy" every time I return to revise a print.
The texture layer is my favorite habit. I keep a small library of my own images — concrete from a Chongqing street, dry grass from the English countryside, cloud detail from Moraine Lake — and I lay one over a finished composite at low Opacity, usually somewhere between 12 and 25 percent. It unifies elements from different sources so they read as one photograph rather than a clippings collage.
I use gradient maps and curves adjustment layers for the final tone. Cool teal in the shadows and warm orange in the highlights, applied through a gradient map, ties a composite together faster than repainting individual areas. Layer masks let me pull that grade back on skin alone, so a face doesn't get dragged into the same stylized color as the background behind it.
Mistakes I made on purpose
My early files in this category are loud. I pushed Clarity until edges glowed, replaced skies without matching the direction of the existing light, and stacked so many textures that the underlying photograph vanished. Those Victoria Peak composites look like posters because I treated every slider as something to max out. Keeping them visible is deliberate — they mark the distance between chasing an effect and controlling one.
The fix turned out to be subtraction. Now I build a composite, walk away from the screen for an hour, and come back asking what I can remove. Nine times out of ten the answer is a layer I loved making but that the image doesn't need. A subtle curve beats another texture; a single strong idea beats four weak ones competing for the same patch of sky.
One habit has saved more images than any technique: I export a small flattened JPEG and look at it on my phone, away from the editing monitor. Zoomed to 200% on a calibrated screen I can prove a mask is perfect; at thumbnail size on a phone I find out whether the picture is actually worth looking at. The phone test catches the composites that were clever rather than good, and there have been a lot of those.
Photoshop doesn't make an image interesting. It amplifies what was already interesting and hides what wasn't — which turns out to be a useful mirror for how I shoot.
Somewhere between photograph and illustration
Not every image here aspires to look real. Some are digital paintings built on top of a photo — brush strokes laid over a hillside in Corsica, a sky repainted into something the weather never produced. Those pieces accept that they are illustrations and stop pretending otherwise, and I find they age better than the composites that tried to pass for straight photographs.
Snapseed on my phone handles the quick experiments when I'm traveling with the Sony Alpha A6000 and no laptop. It isn't where the serious pieces are made, and the layer control it offers is far thinner than Photoshop's, but it's a fast way to test a color idea before I commit to it on the desktop. When a file is finished, whether barely touched or heavily built, there's one test I apply: if a viewer asks "where was this taken?" rather than "how was this made?", the image has stayed a photograph. If they ask how it was made, I've either pushed too far or pushed in a direction that made the picture less honest than it should have been. Both outcomes are informative, and both end up in this collection.