Add Optional Charset (like ASCII) Option in Watermark Extraction

[mhellmeier]

🚀 Feature Request

Current Problem

The robustness problem is described in the upstream issue #20.

Proposed Solution

Since the library can currently embed any byte-encoded sequence inside the cover text, most use cases use plain text encoded in ASCII (letters, numbers, maybe some special characters like "!?) in the watermark. This is only a small part of the Unicode standard.

An option during the watermark extraction can be set to, e.g., only try to extract a watermark encoded in ASCII. In this case, all watermarks containing non-ASCII characters can be identified as broken watermarks.

Additional Context

./.