Is there only one OCR engine in SE?

[mzso]

I looked a bit in the configuration, but I didn't find any config to change it.
What's included makes some pretty fundamental errors, perplexingly with rather high resolution blu-ray PGS subtitles.
For example it regularly produces "I111" from "I..." Which is very much dissimilar.

[darnn]

There are quite a few, actually. You don't change them in the configuration but in the OCR window itself:

[mzso]

@niksedk
Any particular reason the OCR engines are so restricted, and even the normal mode default (binary image compare) is missing?

[niksedk]

I think that OCR often requires some review and in the case of BinaryOCR user input for unknown words + pixels-is-space.
Tip: Inspect compare matches to improve wrong letters for BinaryOCR and nOCR.

BinaryOCR will be included in the next SE 5 beta in "Batch convert".

[mzso]

@niksedk
But doesn't all need review? I tried nOCR manually yesterday and it needed interaction. So I'm guessing in batch this is skipped for all available OCR methods?

[mzso]

@niksedk
I tried Binary OCR in the latest beta, but it stalled at 1% on the first file.

[mzso]

@niksedk Tried the newer b21, it stalls at 0% with binary OCR.

PS: I figured out that only "-batchconvertui" (which I used until now for a year or two) stopped working for some reason, "/batchconvertui" still works. However all the processing options in batch convert were reset every time so far. Which is a pain.