With the risk of this writing coming off as pretentious, many consider photography a form of art meant to tell stories, evoke emotion, and/or be aesthetically pleasing. The processing of raw sensor data, which comes from photons detected on a digital sensor after passing through multiple pieces of glass, into a photo that conveys the intention of the artist is a very subjective process.

My goal in photography is to share what I remembered to have experienced with my own eyes. Which means my photos are as close to what the real scene looked like as possible with some subjective emphasis added. And because a static 2D image is not the same as the dynamic 3D world, conveying a scene involves photo editing.

Photo editing is unavoidable and involves anything from demozaicing raw sensor data to focus stacking to replacing objects with generative AI. Here, I outline what I do and do not do, and why.

Noise, dust spots, lens corrections, and vignette corrections are great. There are many imperfections related to how photos are captured. Lenses distort the perspective and often have some dirt on them, sensors need enough signal to noise, etc. I think it’s fine to clean up most of these issues. Vignette, the darkening of the exposure towards the outer edges of the photo, is a lens artefact, but actually can have an artistic purpose in focusing the scene. For some scenes I even add a bit of vignette, to make the corners of a photo less distracting.

HDR, focus stacking, panoramas, cropping are great, too. Sometimes, technical limitations require you to combine multiple photos into one. Whether it’s due to the dynamic range of light being too great (imagine a mountain scene with shades rocks and bright direct sunlight in overhead clouds), too shallow depth of field (macro photography and other intimate scenes), or a very wide object (a shelf cloud that spans your entire field of view), all require multiple photos to merge into one. This, too, I think it completely fine. I typically avoid it though, as it can be a painful and long process behind the computer to get right, especially when you made a mistake in the field while capturing the shot.

Coming home and realising you framed the scene incorrectly can sometimes be fixed with a crop, which I also have no problem with. I shoot everything in standard 4:3 ratio and decide later whether it should be 1:1, 16:9, or whatever else.

Colour, contrast, and exposure is the most difficult. The first part of the editing process is to start in a neutral image, apply lens corrections, remove default sharpening, and enable a bit of denoising. Then, it’s a matter of setting the exposure, contrast, white balance, tint, and saturation to a point where it as closely matches how I remember the true scene.

Some scenes have a lot of drama in the sky, or somehow have intense and vibrant light. Photos at home can look bleak by comparison. In those cases, I add some local contrast to the clouds, which can emphasize their texture and weight in a scene. Or I play a bit with (local or exposure-dependant) saturation. This is very difficult to get ‘right’. For some photos I need to take a few days break and take a fresh look, only to realise I should start over. Many of my favorite photos have required very minimal editing, which I think says something important.

Generative fill or object removal in rare exceptions. Removing objects defeat the purpose of photography for me. Yes, a little tree branch that breaks up a composition can be annoying. Or a piece of brick that betrays humans exist in a scene that is otherwise 100% natural. But it is very hard to draw the line of what is acceptable to remove, and so while I have done the occasional clone-stamping, I simply don’t use any of it anymore. The idea is that I pay attention while taking the photo, and if there are distracting objects in the scene, then maybe it wasn’t as good of a scene to begin with.

Generative AI or ‘content-aware’ filling in photos is obviously even worse in that regard. I principially do not use it. The one exception is when I have a panorama shot with a distorted perspective that leaves a few corners empty (no photo data). It can be fixed by cropping the image, but sometimes that can crop out the subject or ruin the composition. I’ve used some ‘content-aware’ filling in those cases. This works fine when it’s just some grass. When it takes too long to get right, then, too bad, I throw out the photo…