When a PDF stores a name object using PDFDocEncoding / Latin‑1 bytes, GemBox.Pdf decodes the name as UTF‑8 unconditionally. Bytes that are not valid UTF‑8 are replaced with the Unicode replacement character U+FFFD, and the original byte is then unrecoverable through the public API. This affects both the raw PdfName.ToString() and the typed name‑valued properties such as PdfCheckBoxField.ExportValue and PdfField.Value.

PdfString falls back to PDFDocEncoding when the bytes are not UTF‑16BE, and exposes its raw bytes (Length, Bytes[] indexer, enumerator). PdfName offers neither, so there is no lossless way to read such a name.

Sample file: G0561_2025_10_14_Verdienstausfall_Antrag.pdf. The checkbox group AW_1 has three options whose export values / appearance‑state names contain German umlauts. In the file these are stored as PDF name objects with hex‑escaped PDFDocEncoding bytes (#E4 = byte 0xE4 = ä):

2613 0 obj