mozilla/pdf.js · error · JpxError

No size marker found in JPX stream

Error message

No size marker found in JPX stream

What it means

Thrown by JpxImage.parseImageProperties() after scanning the entire stream byte-by-byte without finding the SIZ marker (0xFF51). The SIZ marker carries image/tile dimensions and component count; without it the image properties cannot be determined.

Source

Thrown at src/core/jpx.js:116

      // Image and tile size (SIZ)
      if (code === 0xff51) {
        stream.skip(4);
        const Xsiz = stream.getInt32() >>> 0; // Byte 4
        const Ysiz = stream.getInt32() >>> 0; // Byte 8
        const XOsiz = stream.getInt32() >>> 0; // Byte 12
        const YOsiz = stream.getInt32() >>> 0; // Byte 16
        stream.skip(16);
        const Csiz = stream.getUint16(); // Byte 36
        return {
          width: Xsiz - XOsiz,
          height: Ysiz - YOsiz,
          // Results are always returned as `Uint8ClampedArray`s.
          bitsPerComponent: 8,
          componentsCount: Csiz,
        };
      }
    }
    throw new JpxError("No size marker found in JPX stream");
  }
}

export { JpxError, JpxImage };

View on GitHub (pinned to 5903d58d58)

Solutions

1. Verify the stream begins with a JP2 signature (\x00\x00\x00\x0cjP ) or a J2K codestream SOC (0xFF4F) followed by SIZ (0xFF51).
2. Confirm /Filter is actually JPXDecode and that earlier filters (e.g. FlateDecode) were applied.
3. Check the stream isn't truncated — ensure /Length matches available bytes.
4. Catch JpxError and render a placeholder.

// Confirm a JP2/J2K header and SIZ marker exist
function looksLikeJpx(data) {
  const isJp2 = data.length >= 12 && data[0]===0 && data[1]===0 && data[2]===0 && data[3]===0x0c;
  const isJ2k = data.length >= 2 && data[0] === 0xff && data[1] === 0x4f;
  return isJp2 || isJ2k;
}

try { const props = JpxImage.parseImageProperties(stream); }
catch (e) { if (e.name === 'JpxError' && /size marker/.test(e.message)) { /* not JPX */ } else throw e; }

When it happens

Trigger: parseImageProperties reads bytes until newByte < 0 (end of stream) without code === 0xFF51. Means the data has no JPEG2000 SIZ segment — it's not JPEG2000, is truncated before SIZ, or is corrupt.

Common situations: A PDF image XObject with /Filter /JPXDecode whose stream isn't actually JP2/J2K (mislabeled), a truncated JPX stream, or random data routed to the JPX decoder. The SIZ marker is mandatory and near the start of every valid codestream.

Related errors

• ${errorMessages}
• JpegImage.parse - unknown marker: ${fileMarker.toString(16)}
• JpegImage.parse - no frame data found.
• Unknown error
• Invalid data format, must be a TypedArray.

AI-assisted analysis of mozilla/pdf.js@5903d58d58 (2026-08-13). Data as JSON: /api/errors/ddb403e74771472e. Report an issue: GitHub.