Image Formats for PDF: JPEG, PNG, TIFF, and More

← Back to Blog

Why Image Format Choice Matters in PDF

Images are typically the dominant contributor to PDF file size. A print-quality brochure with a handful of photographs embedded at 300 dpi using uncompressed TIFF data can easily exceed 100 MB; the same document with appropriately compressed images may be under 5 MB. The choice of compression method directly determines the trade-off between file size, visual quality, and rendering fidelity. Getting this right matters both for distribution (download speed, storage cost) and for end use (print output, screen readability, archival longevity). The PDF specification natively supports a specific set of image compression filters — not every image format used in desktop workflows can be embedded directly.

JPEG — Lossy Compression for Photographs

JPEG (DCT in PDF terminology, using the FlateDecode or DCTDecode filter) is the most widely used image compression type in PDF documents, particularly for photographs and continuous-tone colour images. JPEG uses Discrete Cosine Transform to divide the image into 8×8 pixel blocks, applying quantisation that discards high-frequency detail less perceptible to the human eye. The result is a significantly smaller file at the cost of some image information being permanently discarded.

In PDF, JPEG is ideal for photographs, product images, and any content with smooth tonal gradations. It is a poor choice for images containing sharp edges, flat areas of solid colour, text, or diagrams — the blockiness and colour fringing introduced by JPEG compression are clearly visible at these boundaries. Quality settings typically range from 1 to 12 in Acrobat (equivalent to roughly 10–100%), with quality 6–8 providing a good balance for most photographic content in PDFs intended for screen viewing, and quality 10–12 for print output where quality preservation is paramount.

PNG — Lossless Compression for Graphics and Screenshots

PNG uses lossless DEFLATE (ZIP) compression, meaning no image data is discarded. In PDF, this corresponds to the FlateDecode filter applied to an uncompressed image stream. PNG preserves every pixel exactly as created, making it the correct choice for screenshots, diagrams, logos, illustrations with flat colour areas, and any image where text or sharp edges are present. PNG also supports a full alpha transparency channel, enabling images with irregular outlines or soft edges to composite cleanly over any PDF background.

The trade-off is file size: a PNG-encoded photograph is substantially larger than an equivalent JPEG at medium quality, with no visible quality benefit at screen resolutions. For photographic content in PDFs, PNG is the wrong choice. For graphics-originated content — anything created in a vector application or screen-captured — it is usually the right one.

TIFF — Lossless Archival and Prepress

TIFF (Tagged Image File Format) is the professional standard for archival image storage and prepress workflows. It supports multiple compression schemes internally (uncompressed, LZW, ZIP, CCITT, and JPEG), bit depths up to 32-bit per channel, multiple colour spaces including CMYK, and embedded ICC profiles. When images are placed from TIFF files into PDF-generating applications such as InDesign or Acrobat Distiller, the application’s PDF output settings determine how the image is compressed in the resulting PDF — the TIFF format itself is not the compression stored in the PDF stream.

TIFF is the preferred source format for professional image editing and production: it preserves all data across multiple save operations, unlike JPEG which re-applies lossy compression each time. For archival PDF standards such as PDF/A, images with lossless compression are preferable for long-term fidelity. In practice, PDF/A-compliant tools often embed images as JPEG or JPEG 2000 with quality settings chosen to balance fidelity against file size.

JBIG2 — Bilevel Compression for Scanned Text

JBIG2 is a compression standard specifically designed for bilevel (black-and-white, 1-bit) images, with particular efficiency for scanned document pages containing primarily text. JBIG2 works by identifying recurring patterns of pixels — effectively recognising that many characters on a page are repeated — and encoding them by reference rather than storing each instance independently. For scanned text documents, JBIG2 achieves dramatically better compression than any general-purpose algorithm: ratios of 10:1 or better compared to uncompressed bilevel data are typical, and JBIG2 consistently outperforms CCITT Group 4 for text-heavy content.

JBIG2 was introduced in PDF 1.4 and is used extensively in Acrobat when scanning documents to PDF with the “Black & White” or “Grayscale” colour mode. It supports both lossy mode (where visually similar character shapes can be substituted, improving compression at some risk to absolute pixel accuracy) and lossless mode. For archival applications, lossless JBIG2 is strongly preferred. JBIG2 is not suitable for greyscale or colour images.

CCITT Group 3 and Group 4 — Fax-Compatible Bilevel

CCITT Group 3 and Group 4 are older bilevel compression standards originating from fax transmission protocols. Group 3 uses run-length encoding with Huffman coding; Group 4 is a more efficient two-dimensional extension. Both produce lossless compression of 1-bit images. In PDF, CCITT compression (the CCITTFaxDecode filter) provides broad compatibility — it predates JBIG2 and is supported by every conforming PDF reader. Group 4 in particular is widely used in document scanning workflows and scanned-document PDFs, especially in older systems and those requiring compatibility with fax infrastructure. For new documents, JBIG2 lossless provides better compression for equivalent quality, but CCITT Group 4 remains prevalent in legacy and compatibility-focused applications.

JPEG 2000 — Better Compression, PDF 1.5 and Later

JPEG 2000 (the JPXDecode filter in PDF) uses wavelet-based compression rather than JPEG’s block DCT approach, avoiding the characteristic blockiness of JPEG artefacts. It supports both lossy and lossless modes within a single format, handles transparency natively, and achieves better compression-to-quality ratios than standard JPEG at equivalent settings — typically 20–40% smaller at the same visual quality. JPEG 2000 also supports progressive rendering and multiple resolution levels within a single file, which can benefit streaming and tiled image display.

JPEG 2000 is available in PDFs from version 1.5 onward and is used by Acrobat when the “JPEG 2000” option is selected in PDF Optimizer. Adoption has been limited by the higher computational cost of encoding and decoding compared to standard JPEG, and by the prevalence of existing JPEG-based tooling. For new archival or high-quality distribution workflows where PDF 1.5 compatibility is acceptable, JPEG 2000 merits consideration.

WebP and AVIF — Not Natively Supported by PDF

WebP and AVIF are modern web image formats that offer excellent compression efficiency for screen delivery. However, neither is a natively supported image compression type in the PDF specification. Images in these formats must be decoded and re-encoded to a supported type (JPEG, JPEG 2000, or lossless DEFLATE-compressed bitmap) during PDF creation. Acrobat and PDF-generating applications handle this conversion automatically when placing WebP or AVIF source images.

How Acrobat Handles Image Compression in PDF Optimizer

Acrobat’s PDF Optimizer provides separate compression settings for colour images, greyscale images, and monochrome (bilevel) images. For colour and greyscale, the available options are JPEG, JPEG 2000, and ZIP (lossless). For monochrome images, the options are JBIG2 (lossy and lossless), CCITT Group 4, CCITT Group 3, ZIP, and Run Length Encoding. The Audit Space Usage tool in PDF Optimizer reports how much of the file’s total size each image type currently occupies, allowing targeted optimisation decisions.