Subtitles not coming up.

[lokito50]

Hi all,
Once again another thread about subtitles. I have sifted thru the previous threads and still running into the same issue. What I want is spanish subtitles to come thru UMS, I am using an LG smart TV thru DLNA and the attached settings.
What is happening is the following: the only way the subs are coming up is when I play the "resume movie" but I want them to work since I play the beginning of the movie. I disabled the resume function in UMS and then the subs don't work. The subs are in an .srt file that I named exactly the same as the movie file so in the movie folder there are two files: movie.mp4/mkv/avi and movie.srt. Also there aren't any subs embedded in the movie file itself. Any help would be appreciated it, thank you for looking.

[Nadahar]

Take a look at the "Audio settings" tab. There you can specify the audio/subtitle combinations you want. If the audio is English, use Spanish subtitles etc. I bet it's set to not show subtitles for the audio language you use.

[lokito50]

I changed it to this and still not working. The only way they work is when I enable the resume function and play the resumed file. It's very annoying because I want to get it working since the beginning. Thanks

[Nadahar]

I think you should take a look at this: https://github.com/UniversalMediaServer ... -Subtitles

You need to take into account that the audio language, the subtitles language or both isn't detected. Thus, something like this is probably what you want:

Code: Select all

spa,off;*,spa;*,und

[Rejpal]

Try to add proper ISO language code to the filename - ie. "movie.mp4" and "movie.spa.srt". If "spa" isn't working, try "es" (ISO 639-2 Language Codes, https://www.loc.gov/standards/iso639-2/ ... e_list.php)
UMS should be able to autodetect the subtitles language but the detection is not perfect (seems more like a guesswork to me), this way, with the language code, you'll tell what language exactly it is.

[Nadahar]

@rejpal You're absolutely correct, it works much better if the language code (2 or 3 letter, which one doesn't matter) is added to the file name. I don't know why people doesn't do this by default, as there's no reliable way to determine the language automatically that I know of. If we sent every file for "analysis" at Google or similar, I guess the auto-detection would improve, but that's not a viable solution for many reasons.

UMS uses ICU4J for both character set and language detection. It uses n-gram based statistical detection, which is in short calculating a "score" for the languages supported by the detector based on "signatures" of typical properties of the given language. The language with the highest score is then returned. This leaves a lot to be desired both when it comes to precision and when it comes to which languages it can detect.

Subtitles are also often short, incomplete sentences which can make the statistical analyses even less accurate. If we had a proper text of a decent length, the detection would probably be better. In any case, there's nothing much we can do to improve this, I doubt that we would be able to make a better detection ourselves

[lokito50]

Alright guys I'll try it out next time. Thanks for the help.