Page Orientation and Rotation in PDFs

Orientation vs. Rotation

These are not the same property:

• Page orientation: the page layout. Portrait means height is greater than width; Landscape means width is greater than height.
• Rotation angle: how the content sits relative to the natural reading direction, for example 90, 180, and so on.

Option 1: Read Rotation from Metadata

PageRotation returns the rotation value embedded in the PDF file, so it tells you how the content is programmatically rotated. For well-formed PDFs this is reliable and consistent.

var pdf = PdfDocument.FromFile("sample.pdf");
for (int i = 0; i < pdf.PageCount; i++)
{
    var page = pdf.Pages[i];
    var orientation = page.Width > page.Height ? "Landscape" : "Portrait";
    var rotation = page.PageRotation;
    Console.WriteLine($"Page {i + 1}: {orientation}, Rotation: {rotation} degrees");
}

var pdf = PdfDocument.FromFile("sample.pdf");
for (int i = 0; i < pdf.PageCount; i++)
{
    var page = pdf.Pages[i];
    var orientation = page.Width > page.Height ? "Landscape" : "Portrait";
    var rotation = page.PageRotation;
    Console.WriteLine($"Page {i + 1}: {orientation}, Rotation: {rotation} degrees");
}

Imports System

Dim pdf = PdfDocument.FromFile("sample.pdf")
For i As Integer = 0 To pdf.PageCount - 1
    Dim page = pdf.Pages(i)
    Dim orientation = If(page.Width > page.Height, "Landscape", "Portrait")
    Dim rotation = page.PageRotation
    Console.WriteLine($"Page {i + 1}: {orientation}, Rotation: {rotation} degrees")
Next

You get both facts in one pass: orientation from the width-to-height comparison, and the rotation angle straight from the metadata.

Best for: structured PDFs with consistent, valid metadata.

Option 2: Detect Orientation with OCR

When the metadata is missing or untrustworthy, analyze the page visually instead. DetectPageOrientation() examines the layout of each page and infers the likely reading direction, with no reliance on metadata.

using var input = new OcrInput();
input.LoadPdf("sample.pdf");
var results = input.DetectPageOrientation();
foreach (var result in results)
{
    Console.WriteLine($"Page {result.PageNumber + 1}: Rotation: {result.RotationAngle} degrees");
}

using var input = new OcrInput();
input.LoadPdf("sample.pdf");
var results = input.DetectPageOrientation();
foreach (var result in results)
{
    Console.WriteLine($"Page {result.PageNumber + 1}: Rotation: {result.RotationAngle} degrees");
}

Imports System

Using input As New OcrInput()
    input.LoadPdf("sample.pdf")
    Dim results = input.DetectPageOrientation()
    For Each result In results
        Console.WriteLine($"Page {result.PageNumber + 1}: Rotation: {result.RotationAngle} degrees")
    Next
End Using

This is the right choice for scanned documents, image-based PDFs, and any file lacking proper metadata.

Accuracy here tracks text density and clarity. Sparse or misaligned text can produce a wrong result, so restrict OCR to a Rectangle over a denser region of text to steady the detection.

Recommendation

• Digital PDFs: prefer PdfDocument.PageRotation when the metadata is correct.
• Scanned or image-based PDFs: use OcrInput.DetectPageOrientation() where metadata is missing or unreliable.

For critical workflows, run both methods and apply your own logic to reconcile any discrepancy. If you are seeing inconsistent rotation results, check the specific page layout and validate one method against the other.