A batch workflow for settings watermarks

One watermark takes thirty seconds. Thirty of them is a different job: the order drifts, the margins stop matching, and you can no longer tell which file in the export folder is the original. This is about the route from memory card to published post, not which button to press.

Hands-on version: Adding camera watermarks to a whole set — the three-step flow with interface screenshots, a before/after pair and the failure cases people actually hit.

1. Select first, watermark second

The commonest waste is doing it backwards: watermarking forty frames, then cutting to eight. Pick your keepers first.

Then check three things:

• Are these the original files? A JPEG straight off the card, or one exported from a RAW, carries full EXIF. A copy saved out of a messaging app has already lost it; see the watermark guide.
• Is the editing done? Grade and crop before the watermark goes on. The strip is pixels, and a crop afterwards will cut straight through it.
• Do the frames share a tonal range? Night shots and snow shots won't both look right in one white border. Split them into two batches.

2. A naming rule

DSC_0421.JPG tells you nothing inside a set, and file numbers collide as soon as you mix bodies or cards. This scheme is hard to give up once you've used it:

Together: 20260902-westlake-003-mark.jpg. Mark adds the template name to the original filename when it exports, so rename the originals first and the suffix comes along for free. Keep originals and exports in separate folders and never let an export overwrite a source file. A watermark can't be taken off; lose the original and you're reshooting.

Use hyphens, not spaces: a space turns into %20 the moment the file becomes a link. And stick to plain ASCII so the name survives the trip between camera, phone, cloud drive and someone else's computer.

3. One style per set

Across a set, consistency beats any single frame. In a grid of nine, nine slightly different border widths look sloppy before anyone notices how good one of them is.

1. Lock the style and fields first. Strip or border, which three or four items; then use them for everything. See choosing a watermark style.
2. Work the margin out from the short edge when the set mixes orientations. A fixed pixel value makes vertical frames look pinched.
3. If a field is missing on some frames, drop it from all of them. Frames shot on an adapted lens have no lens model, so switch the lens line off for the whole set rather than leaving gaps.

4. Export size and format

• Export JPEG. A photo saved as PNG is several times larger with no visible gain, and platforms transcode whatever you upload anyway.
• 1600–2560 pixels on the long edge is plenty for social. Anything bigger gets squeezed harder at the other end and takes longer to upload. Keep a full-resolution copy for the archive or for print.
• Quality 85–92. Above that the file grows fast for differences you can't see.
• Convert to sRGB. Adobe RGB or Display P3 files look washed out in apps and browsers that ignore colour profiles.

Set size and quality when you export from your editor, with metadata included, before the files go into Mark. Mark has no size or quality setting: it renders at the original's full resolution and writes JPEG at quality 100.

The web version takes several photos at once. You switch between them on the filmstrip, each one keeps the same template and settings, and nothing is uploaded. Downloading is still one file at a time; a zip download for the web version is being worked on. For a few dozen files, the desktop, Android and command-line versions batch a whole folder in one go with the same rendering engine.

Blurred-background fill is the slow part on big files. Timed with the v0.2.31 command-line build on an i7-12700H, one 24 MP frame padded out to 4:5 took 3.9 s with black fill and 106.4 s with blur. For a big batch, switch to black or shrink the long edge to 2000–3000 pixels first.

5. What happens after you post

This is where the effort pays off or gets wasted.

• Normal sending recompresses and strips EXIF. The watermark survives because it's part of the image.
• Grid previews crop to squares. A corner tag can lose half of itself in the preview, so use a bottom strip or a full border for a set. Or set Mark's ratio to 1:1, which extends the canvas instead of cropping the photo.
• Sending "the original" sends the EXIF too, GPS included if it's there; see what is hidden in a photo. Mark's exports already have GPS removed and the settings kept, and with the settings printed on the frame there's little reason to send originals at all.
• Archive the unmarked file. The watermarked copy is only for publishing.

6. FAQ