MP writing TV subtitles as image not text (1 Viewer)

[doveman]

It seems that when recording TV (DVB-T UK Freeview) the recordings are stored in the .ts file as images, rather than text. This means that MCE Buddy will be unable to extract them and convert them into a .srt file when compressing the .ts to an .avi or .mp4, so the subtitles will be lost.

Is it possible to change this so that the subtitles are stored in the .ts as text, as done by MCE and NPVR?

 

[mm1352000]

Hi doveman

Think of the TS written to the HDD as a slimmed down version of the full stream from the tuner. Like taking 3 wires (video, audio, subtitles) from a bunch of 50 (all the other irrelevant PIDs). We don't change the content *in any way*. As such, the subtitle format is entirely determined by the broadcaster. We can do nothing about this.

mm

 

[doveman]

Hi mm

OK, thanks for explaining. I think the people I spoke to who get the subtitles as text in the .ts with NPVR are probably in the US then and it must be broadcast like that over there.

 

[mm1352000]

Yes, that is possible. You'd have to find out more about the actual format that your subtitles are provided in. Sorry, and good luck with that.

mm

 

[SimonP]

UK Freeview does indeed use image-based subtitles, which are known as DVB-SUBs. In some cases the recording also contains teletext subtitles, which will convert to .srt easily, but normally you'll have to run the .ts file through something like 'Project X' which can de-mux the file to its video, audio and subtitle components, then you run the subtitle file through an Optical Character Recognition (OCR) program such as SubRip to read the images and output them as .srt.

Have a go and if you want, I can let you have a copy of my setup which does it in moments.

 

[doveman]

I was discussing this with foxbenw in another thread https://forum.team-mediaportal.com/threads/do-i-need-handrake-as-well-as-videoredo.107655/page-2 and he also suggested Project X, although I'm hoping that MCE Buddy might be able to incorporate some of the tools to automate the process, as Project X is rather cumbersome just to demux the subs.

Does Project X not do the OCR then? It's just that it produced four files, including a .sub and .idx which I thought were text rather than images. Even if not, it seems they can be muxed back into the video, so perhaps the OCR step isn't necessary.

 

[SimonP]

I don't know MCE Buddy (as I used MediaPortal) but once you have Project X set up it's dead easy to use and I drive the whole process using a batch file to make it even easier.

For my purposes Project X just de-muxes the file and controls the output, it doesn't do OCR, and .sub is the DVB-SUB image-based subtitle file that I referred to before. If you can mux the .subs and your player plays them then all well and good. Otherwise, here is my subtitle directory all set up and ready to go with a README file explaining how to use it. This setup rips the subtitles to simple mono .srt files as that is the standard, and it ignores the video and audio output for speed and because I'm only interested in the subtitles but play around with the presettings if you want it to do something different.

https://www.rapidshare.com/files/1523770478/SUBTITLES.zip

 

[doveman]

Thanks SimonP, that'll be a great help. I'll have a play around with it later