Mkv Subtitles passthru extracting issue

[Coolnodje]

I’ve bumped into a few situation where subs embedded in an Mkv would become unreadable in the resulting mp4 file.

I suspect the original encoding would be specific because these were east asian script.

I don’t think Subler is supposd to support encoding other than UTF-8, or is it planned/enabled?

Is there any tool I could use to check the subtitle encoding in an MKV file?

‌

FYI: VLC would qualify these subs as ‘BD PGS subtitles (bdpg)’

---

I know understand better Subler’s possibilities at this level.
The passthru option upon importing Mkv is absolutly perfect. There’s then no point using OCR.

But it seems there’s a font sizing problem: the resulting MP4 with passed-thru VobSub disaply them in a very big font.
The attached pics illustrate the problem for jap. subs, but note that it’s exactly the same issue for regular English subs.

In this particular case both Jap. and Eng. VobSub are DVD Subtitles (Spu), as displayed with VLC.

HTH.

‌

[Damiano Galassi]

PGS subtitles are not text, are images. Subler will perform OCR on it and try to get some text out. For the best result, you have to download the language data file for the languages you want (in the Preferences window) and set the track language.

[Coolnodje]

It’s amazing man, I didn’t realize Subler was doing OCR, haven’t had to open prefs up to now.
Subler is working wonderfully and is super useful for me atm.

How did you get the OCR done in such a small piece of software?
Are you using Google Vision API or some services? Or were you able to integrate something like Tesseract to perform it offline?

[Damiano Galassi]

Subler uses Tesseract