Images to PDF

How Image-to-PDF Works

A PDF is a structured document with pages, fonts, and embedded objects. Images go into pages as XObjects.

Page Sizes (in points: 1 inch = 72 pt)

PDF uses points internally. The render size is determined when printing or displaying.

Image Encoding in PDF

The PDF spec supports JPEG, PNG-style (Flate-compressed), JBIG2, JPEG2000, and CCITT (fax) as image encodings. In practice:

• JPEG photos → embed as-is (no recompression). Best size.
• PNG screenshots → embedded via Flate (PNG-style). Lossless.
• Black-and-white scans → CCITT Group 4 is tiny (PDF/A for archive standard uses this).
• WebP, AVIF, HEIC → must be transcoded to JPEG/PNG first (most PDF libraries can't embed them directly).

A well-built image-to-PDF tool detects format and embeds optimally. Naive tools force re-encode to JPEG quality 75, fine for photos, terrible for diagrams.

Margins and Layout

The image is scaled (preserving aspect ratio) to fit (page - 2×margin) on each axis. Either width or height hits the limit first; the other dimension is what the image's aspect dictates.

Multi-Page PDFs

Each image becomes one page (default), or N-up for contact sheets:

For photo contact sheets: 4×3 = 12 thumbnails per page is standard.

Page Order and Reordering

UI typically shows a draggable thumbnail list. Internally, the order is just the order of addPage calls. Reordering = recomputing the array, re-rendering pages.

File Size Math

A 5MB JPEG embedded raw: PDF is ~5MB + small overhead. 10 such images: ~50MB PDF.

Optimization:

• Downscale large images before embedding (4000px → 1600px halves dimensions, quarters bytes).
• JPEG re-encode at lower quality (q70 instead of source q95): 30-50% savings, slight visual loss.
• Convert PNGs of photos to JPEGs (PNG-of-photo is huge; JPEG is 5-10× smaller).
• CCITT Group 4 for B&W scans: 100× smaller than color JPEG.

For email-friendly PDFs: target <5MB total. For archival: keep originals.

PDF Metadata

Helps with file organization and search. Stripped automatically by some sharing platforms.

Common Workflows

1. Scanning a multi-page document: phone-scan each page → reorder → set page size to A4 → margin 0.5 in → export.
2. Photo album: select photos → set page size to 8×8 in (square) → orient landscape for landscape photos, portrait for portraits → zero margin → export.
3. Receipt collection for taxes: scan receipts → fit each to its own page → metadata "2024 Receipts" → export to send to accountant.
4. Slideshow handout: export slides as PNGs → images-to-PDF with 2-up landscape → handout-ready.
5. Tickets/boarding passes: combine into one PDF for travel → keep originals as backup.

Common Issues

• Wrong rotation: phone photos have EXIF orientation; some PDF builders ignore it, putting portraits sideways. Pre-rotate before adding to PDF.
• Massive files: embedding 4000×3000 JPEGs at full size when the target is screen viewing (where 1200px is enough). Downscale first.
• Watermarks/timestamps on photos: get baked in. Edit photo first, then PDF.
• Selectable text impossible: pure image PDF has no text layer. Run OCR after if you need search.

Alternatives

For more complex needs:

• ImageMagick convert *.jpg out.pdf: command-line batch, supports OCR via Tesseract integration.
• Adobe Acrobat: paid, full control, OCR built in.
• Print-to-PDF: open images, print, choose "Save as PDF", works for one-off but slow for batches.
• Print-to-PDF from HTML: build an HTML page with images, print to PDF, most flexible layout.

For pure image-to-PDF: the browser tool is fastest and most private.

Privacy Note

Image bytes are loaded into memory, embedded into the PDF in browser, and the result is offered for download. Nothing is sent. This matters because the typical use case (scanning IDs, receipts, contracts, business documents) involves exactly the kind of content you wouldn't want sitting on a third-party server's log. Confidence-test: open browser devtools network tab, watch it stay quiet while you generate the PDF.