PdfPecker
Docs 1.20.0: PDF to image
Download PDF Changelog

This is the manual for PdfPecker 1.20.0. Read it for the latest version, 1.25.0.

PDF to image

PdfToImage and PdfRenderer render PDF pages to PNG, JPEG, WebP or TIFF (or raw RGBA pixels). This works for any PDF, not only PdfPecker's own output, and also uses no third-party packages: the PDF parser, font engines, JPEG and fax decoders, rasterizer and the PNG, JPEG, WebP and TIFF encoders are all part of the library.

// Every page as PNG bytes (rendered in parallel), or one page (0-based index)
byte[][] pages = PdfToImage.Convert(pdfBytes, new ImageOptions { Dpi = 150 });
byte[] first = PdfToImage.ConvertPage(pdfBytes, 0);

// report.pdf -> out/report-1.png, out/report-2.png, ...
PdfToImage.ConvertFile("report.pdf", "out");

// JPEG or lossless WebP: report.pdf -> out/report-1.jpg, ... (or .webp)
PdfToImage.ConvertFile("report.pdf", "out", new ImageOptions { Format = PdfImageFormat.Jpeg, Quality = 90 });
byte[][] webps = PdfToImage.Convert(pdfBytes, new ImageOptions { Format = PdfImageFormat.WebP });

// Open once, render many pages; RenderedImage also exposes the RGBA pixels
var renderer = PdfRenderer.Open(pdfBytes, password: null);
RenderedImage image = renderer.RenderPage(2, new ImageOptions { Width = 800 });
image.SavePng("page3.png");
File.WriteAllBytes("page3.jpg", image.ToJpeg(quality: 85));
File.WriteAllBytes("page3.webp", image.ToWebP());

// The whole PDF as one multi-page TIFF, here black and white for fax or archiving
byte[] tiff = PdfToImage.ConvertToTiff(pdfBytes, new ImageOptions { Dpi = 200, TiffColorMode = TiffColorMode.BlackAndWhite });
PdfToImage.ConvertFileToTiff("report.pdf", "report.tif");

HTML to images

HtmlToImage converts HTML straight to one image per page, for previews, thumbnails and e-mails. It is the shortcut for HtmlToPdf.Convert followed by PdfToImage.Convert, takes the same PdfOptions and ImageOptions, and gives the same images. Security and Signature are ignored, since they change nothing on the page.

byte[][] pages = HtmlToImage.Convert(html, new PdfOptions { PageSize = PageSize.A4 }, new ImageOptions { Dpi = 96 });
byte[][] jpegs = await HtmlToImage.ConvertAsync(html, imageOptions: new ImageOptions { Format = PdfImageFormat.Jpeg });

Output formats

Output formats (ImageOptions.Format, default Png; RenderedImage.Encode()/Save() use it too):

  • PNG: lossless, RGB for opaque pages and RGBA otherwise.
  • JPEG: baseline JFIF with the standard (IJG) quantisation tables scaled by Quality (1 to 100, default 85) and Huffman tables optimised per image. Below quality 90 colour is subsampled 4:2:0; from 90 it keeps full resolution (4:4:4), which keeps coloured text sharp. JPEG has no transparency, so a Transparent render is flattened onto BackgroundColor (white by default). At quality 90 a 150 DPI text page decodes at about 45 dB PSNR.
  • WebP: always lossless (VP8L), alpha included, decoding to exactly the PNG's pixels and usually 50% to 70% smaller. Pages with up to 256 colours use a palette (packed 2, 4 or 8 pixels per byte when small); others use subtract-green and per-tile prediction, then LZ77, a colour cache and canonical prefix codes. Here Quality is the compression effort (default 29): higher spends more time searching for a smaller file and never changes a pixel; 85, the default before 1.19.0, makes files about 1% smaller and takes nearly twice as long. Lossy WebP is not written. WebP is limited to 16,384 pixels per side (JPEG to 65,535); larger images throw NotSupportedException.
  • TIFF: lossless, one page per file with Format = PdfImageFormat.Tiff (.tif from ConvertFile, or RenderedImage.ToTiff()), or every page in one multi-page file with ConvertToTiff, ConvertToTiffAsync, ConvertFileToTiff and PdfRenderer.RenderAllToTiff. See TIFF.

TIFF

ImageOptions.TiffColorMode picks Color (the default: RGB, or RGBA when the render is transparent), Grayscale or BlackAndWhite. TiffCompression defaults to CCITT Group 4 for black and white and LZW otherwise; None, Lzw, Deflate, PackBits and CcittGroup4 (always black and white) can be set, and all of them are lossless. For black and white, TiffThreshold (1 to 255, default 128) is the brightness below which a pixel becomes black. Greyscale and black and white flatten transparency onto BackgroundColor, and Dpi is written into the file so it prints at the right size. Files are classic TIFF, so output over 4 GB throws NotSupportedException (BigTIFF is not written).

// Greyscale archive copy with Deflate
byte[] archive = PdfToImage.ConvertToTiff(pdfBytes, new ImageOptions
{
    Dpi = 300,
    TiffColorMode = TiffColorMode.Grayscale,
    TiffCompression = TiffCompression.Deflate,
});

Options

ImageOptions sets Dpi (default 150) or a target Width/Height in pixels, BackgroundColor or Transparent, RenderAnnotations, Password, MaxDegreeOfParallelism, PngCompression, Format, Quality (85 for JPEG and 29 for WebP until it is set; an ArgumentOutOfRangeException outside 1 to 100) and the TIFF options. A PdfRenderer is thread-safe and shares decoded fonts and images between pages. Encrypted files open with the user or owner password (PdfPasswordException otherwise); unreadable files throw InvalidDataException, as does a page that needs more than 20 million drawing operations (nested forms can multiply a few bytes into billions). Pass a cancellation token to bound the time a page may take.

What the renderer supports

  • Files: cross-reference tables and streams, object streams, incremental updates, damaged files (rebuilt by scanning), RC4 and AES-128/256 encryption, all standard filters (Flate, LZW, ASCII85, ASCIIHex, RunLength, predictors).
  • Graphics: paths, fills (non-zero and even-odd), strokes with joins, caps and dashes, clipping, anti-aliasing, all blend modes, constant alpha, soft masks, transparency groups, axial, radial, function and mesh shadings (types 1 to 7), tiling patterns, form XObjects, optional content (layers) and annotation appearances.
  • Text: embedded TrueType, OpenType, CFF and Type 1 fonts, Type 3 fonts, CID fonts with CMaps, vertical text, all text rendering modes. The 14 standard fonts and non-embedded fonts are drawn with built-in Liberation and DejaVu-derived faces stretched to the document's widths.
  • Images: every bit depth and colour space (Gray, RGB, CMYK, Lab, ICC-based, Indexed, Separation, DeviceN), baseline and progressive JPEG, JPEG 2000 (JPXDecode, including embedded alpha), JBIG2 (generic, text, symbol, refinement and halftone regions, with JBIG2Globals), CCITT fax, stencil masks, colour-key masks and soft masks.

For what the renderer does not support yet, see Limitations.