← Blog

Mixing photos and screenshots in one combined PDF — DPI surprises

CombinePDF accepts both PDFs and images: photos from a phone, screenshots, scans, even SVG diagrams. Each one becomes a page in the output. The surprise that catches people: those pages don't all look equally crisp. A high-megapixel photo and a low-resolution screenshot, both fitted to A4, render at very different effective pixel densities. The PDF is correct; the visual result is uneven.

Why this happens

An image's "size on the page" is computed from pixel count divided by DPI. To fit any image onto A4, CombinePDF chooses a DPI just high enough that the longer axis matches A4's longer axis (within margins). The chosen DPI varies wildly between sources:

A 4032 × 3024 phone photo on A4 portrait → ~488 DPI. The 4032 source pixels stretch across A4's full width.
A 1920 × 1080 desktop screenshot on landscape A4 → ~164 DPI. The 1920 source pixels stretch across A4's longer side (11.69 in), leaving white margins above and below.
A 1024 × 768 older screenshot on portrait A4 → ~124 DPI. The pixels are stretched thin.
A 3300 × 2550 scan of a Letter page at 300 DPI → exactly 300 DPI on Letter (best case: original page at original size).

The on-screen impression at 100% zoom is similar across all four — they all fill the page rectangle. The on-print impression diverges sharply: 488 DPI is excellent print quality; 124 DPI is visibly pixelated on a laser printer at typical reading distance.

What the PDF stores

For each image-derived page, the PDF contains:

An Image XObject with the encoded pixel data.
A page content stream with a cm (concatenate matrix) operator that scales the unit-square image XObject to the page rectangle.

What "changes" between low-DPI and high-DPI images is the matrix — a different rectangle on the page for the same pixel grid.

One thing to know about CombinePDF specifically: it re-encodes images as part of building each page. JPEG inputs are re-encoded at quality 85 when the source quality is higher, and palette/grayscale PNG inputs are run through a lossy quantizer that reduces palette size. CombinePDF prioritizes a smaller merged file over archival-grade fidelity; the original byte stream of a high-quality JPEG or a precise PNG palette is not preserved as-is. For a combine workflow where source fidelity is paramount, convert each image to a PDF with a tool that preserves bytes (such as JPG2PDF) and then combine the resulting PDFs.

What this means for printing

On a 300 DPI laser printer, a 488 DPI source page is downsampled by the print driver to 300 DPI; details are crisp because the printer had more source data than it needed. A 124 DPI source page is upsampled to 300 DPI; the printer had to invent the missing data, smoothing edges. Side by side in the same combined PDF, the difference is obvious: photo pages look like photos, screenshot pages look fuzzy.

If your combined PDF will be printed and visual consistency matters, plan around this:

Match source resolutions ahead of time. Resize all images to the same target pixel count for the same target page size before uploading.
Don't mix dramatically different sources. A document of phone photos prints cleanly; a document of screenshots prints cleanly; a document mixing both has the inconsistency.
Print at "fit to page" rather than 100% scale. If you accept that the printer will downsample everything to its native resolution, the inconsistency reduces to "everything prints at ≤300 DPI" rather than "this is sharp, this is fuzzy".

On-screen the gap is smaller

Most modern monitors are 100–250 PPI. At 100% zoom, a 124 DPI image renders close to its native pixel grid; a 488 DPI image is downsampled to the screen. Both look reasonable; the high-DPI source has invisible "extra" detail that emerges only when zooming in.

Zoom in to 400% and the gap shows: the photo stays crisp because the source had the resolution to support it; the screenshot becomes blurry because the source had no more detail to reveal. This is unavoidable — pixels are pixels, no PDF magic creates them.

What CombinePDF doesn't do automatically

The tool fits each image to A4 individually, choosing per-image DPI. It does not:

Pick a uniform DPI across all pages and pad/center smaller images to match.
Upsample low-resolution sources to match high-resolution ones.
Preserve original JPEG quality above 85 — high-quality sources are re-encoded down to 85.

If you need uniform print-quality across mixed sources, render all images to PDF first in a photo editor that lets you pick DPI explicitly (Photoshop's "Save As PDF" dialog, Lightroom's print module, or any image editor with a "Save as PDF" option that exposes a resolution setting). Then combine those PDFs with CombinePDF — at that point all your inputs are PDFs of consistent dimensions, and the combine step won't touch the embedded images.