Suggestion: Change file open method in generation.py (text2speech)

[smallbutfine]

line 492:
with open(file_path, encoding="utf-8", mode="r", errors='ignore') as f:

I added errors='ignore' because I had problems in batch conversion, when I used splitted files, that were written by a perl script, no matter what, even with explicit UTF-8 encoding, that always throw a decode error with "wrong start code". I had no problem with any other program reading these text files. Python still thinks, chars are malformed.
The addition prevent the error and made huge batches run thru. (book conversion)