PDF to Images

PDF.js handles the rasterization by parsing the page content stream and rendering each drawing operation onto a canvas scaled to the target DPI. Text is rendered from embedded fonts where available, falling back to substitution fonts when a font is referenced but not embedded, which is why PDFs that rely on system fonts sometimes look slightly different when rasterized in the browser versus opened in a desktop viewer. Images embedded in the PDF are rendered at whatever resolution they were stored at, then resampled to match the output canvas, so raster detail is bounded by the lower of "embedded resolution" and "output DPI".

Format choice matters more than most people expect. PNG is the right default for pages with text, diagrams, or line art because it is lossless: the pixels on the output match what the rasterizer produced exactly. JPEG re-encodes those pixels with DCT quantization, which softens text edges (JPEG's 8x8 blocks were not designed for the sharp high-frequency content in rasterized text) and can introduce visible ringing around high-contrast edges. For pages that are mostly photos, say, a PDF of a magazine spread, JPEG produces dramatically smaller files at equivalent perceived quality. For pages of code screenshots or technical diagrams, stick with PNG.

When you need rasterized images and multi-page output, consider the delivery format. Exporting all pages as separate PNGs in a ZIP gives maximum flexibility. Exporting to a single image strip (one tall image containing all pages stacked vertically) is useful for embedding in blog posts. Generating thumbnails at 100-200 pixel dimensions is the right call for document library previews where the actual PDF is one click away. Rasterizing to 300 DPI for print is the specific case where file size is secondary to output quality, do not rasterize to 150 DPI and then try to print at 300 DPI, because upscaling will not add detail that was not captured at rasterization time.