Faulty Parsing for PDF Version 1.2 and 1.3

### Bug
When parsing a PDF document of the versions 1.2 or 1.3 the content is not extracted. The markdown result for these files is sometimes just a long concatenation of "\n"-symbols or some random symbol chain (e.g. ✢ ✗ ✜ ✚ ✘ ✴ ✓ ☛ ✒ ( ✕ ✯) instead of actual raw text.

While I can't share most of these documents, I down-rated one I can share, which causes the same behavior.

[Fußball-Bundesliga 2022_23 – Wikipedia-PDF-1-2.pdf](https://github.com/user-attachments/files/20727999/Fussball-Bundesliga.2022_23.Wikipedia-PDF-1-2.pdf)

### Steps to reproduce
Steps to reproduce
`from docling.document_converter import DocumentConverter

source = "elfa.docx" # document per local path or URL
converter = DocumentConverter()
result = converter.convert(source)
print(result.document.export_to_markdown()) # output: "## Docling Technical Report[...]"`

Docling version
Docling version: 2.33.0
Docling Core version: 2.33.0
Docling IBM Models version: 3.4.4
Docling Parse version: 4.0.1
Python: cpython-311 (3.11.11)
Platform: Linux-5.15.0-94-generic-x86_64-with-glibc2.35

Python version
Python 3.11.11

Start with the supplied PDF and the DocumentConverter example, then inspect the conversion path used by converter.convert() and result.document.export_to_markdown(). Done means PDFs using versions 1.2 and 1.3 produce readable extracted text rather than newline or symbol chains, including the attached Bundesliga PDF.

Written by the indexing model from the issue text.