Some PDFs are bloated because of redundant internal data and inefficient encoding. Standard compression tools handle those well by restructuring the file and removing unnecessary objects. Other PDFs are bloated for a different reason entirely: they are packed with high-resolution images that consume far more space than their display size requires. Standard compression helps these files too, but targeted image-specific optimization helps dramatically more. Converting an image-heavy PDF into a lightweight version means identifying where the image data is excessive and reducing it without making the document look degraded or unprofessional.
The core insight behind image-heavy PDF optimization is simple but not obvious. Most images embedded in PDFs are stored at resolutions far exceeding what the eye can perceive at normal viewing distances. A photograph embedded at 300 DPI and displayed at a quarter of a standard page has roughly four times the pixel data it needs for screen viewing. The extra pixels consume storage space and transfer bandwidth without contributing anything visible. Reducing the stored resolution to match the actual display size is where the real file size savings come from.
According to a 2025 analysis by the PDF optimization company Callas Software, images account for an average of 76 percent of the total file size in image-heavy PDFs, with the majority of those images stored at resolutions two to four times higher than their on-page display size requires (Callas Software, "PDF Composition Analysis Report," 2025). Targeting images directly produces far greater size reduction than targeting text content and structural data combined.
Step One: Assessing the Compression Potential
Before applying any compression, check what resolution the embedded images are actually stored at. Open the PDF and zoom in on a representative image to about 300 percent. If the image remains sharp and detailed at that zoom level, it is stored at high resolution and has significant compression potential. If it pixelates quickly and details become blurry, the image is already at or near screen resolution, and aggressive downscaling would produce visible quality loss.
This visual assessment takes seconds per image and tells you more than any file analysis tool. A PDF full of images that stay crisp at 400 percent zoom can probably be reduced by 70 to 80 percent. A PDF where images start to pixelate at 200 percent zoom has limited remaining compression potential and should be treated more conservatively. The PDF Compression strategy should match the actual resolution headroom in the file, not an arbitrary preset.
Browser-based compression tools typically offer an image downscaling setting measured in DPI. Setting the target resolution to 150 DPI is appropriate for documents viewed primarily on screens. Setting it to 200 DPI provides a safety margin if printing is expected. Setting it to 300 DPI preserves print-quality images while still removing resolution above what a standard printer can reproduce. The tool downscales each image to the target, which can reduce file size by 50 to 80 percent for image-dominant PDFs.
Step Two: Matching Compression Type to Image Content
Photographs compress most efficiently with JPEG compression, which uses lossy encoding algorithms specifically designed for natural images with smooth gradients and subtle color variations. Charts, logos, diagrams, and graphics with sharp edges and flat color areas compress better with lossless methods like ZIP or LZW compression. Applying JPEG compression to a chart or a logo introduces visible artifacts around text edges and sharp boundaries that make the document look sloppy. Applying lossless compression to a photograph produces almost no size reduction because the algorithm cannot find the repeated patterns it relies on.
The right approach is to match the compression method to each image's content type. WukongPDF's Reduce PDF Size tools apply appropriate compression to photographs and graphics automatically, detecting the image type and selecting the optimal compression algorithm. For maximum control over the output, advanced settings allow you to choose the compression method per image or per image category.
Step Three: Stripping Unnecessary Image Ancillary Data
Every image embedded in a PDF carries data beyond the pixels themselves. EXIF metadata from the camera that captured the photo, including GPS coordinates, camera model, and capture settings. ICC color profiles that describe how the image's colors should be interpreted. Embedded thumbnail images that were generated by the source application for quick preview and serve no purpose inside the PDF. All of this ancillary data increases file size without contributing anything to how the image looks when displayed or printed.
Stripping this metadata is a separate operation from resolution downscaling. A compression tool that removes metadata in addition to downscaling resolution produces a smaller file than a tool that only does one or the other. The PDF to Image pipeline approach strips everything by converting each page to a clean image and reassembling, but this is a destructive method that also removes text layers and searchability.
When to Use the Full Rasterize-and-Rebuild Approach
For extreme cases where file size is the absolute priority and searchability is not required, the most aggressive approach is to convert each PDF page to an image at the exact target resolution, then reassemble those images into a new PDF. This method discards all text layers, vector graphics, fonts, metadata, and structural data. What remains is a clean, consistently-sized collection of page images at precisely the resolution you specified. The output file size is predictable and minimal.
This method is appropriate for scanned documents that already have no text layer, presentation handouts where the PDF exists only for viewing, and archival copies where file size matters more than editability. It is not appropriate for documents that need to remain searchable or editable. WukongPDF provides both targeted compression for files that need to preserve their text layers and the full optimization pipeline for files where size reduction is the only priority.
Testing Compression Results Before Committing
The difference between acceptable and unacceptable image compression is visible, but it is visible at specific zoom levels and on specific content types. A photograph compressed at 150 DPI may look perfect at normal viewing zoom and show visible softening only at 200 percent magnification. A chart with small axis labels compressed at the same setting may become illegible. Before compressing a large batch of image-heavy PDFs, test your settings on one representative file. Open the compressed output and review it at the zoom level your recipients will use. If the text is sharp, the images are clear, and no compression artifacts are visible at normal viewing distance, the settings are correct.
Documenting compression settings for recurring document types ensures consistency. When you find settings that work, note the target resolution, the compression method for photographs versus graphics, and whether metadata stripping was enabled. When the same document type arrives next month, apply the same settings and get the same results. This documentation is particularly valuable in team environments where different people process similar documents. A shared reference of compression settings eliminates guesswork. A browser-based compression platform applies consistent settings across sessions. The Reduce PDF Size settings you used last month produce the same results this month.