Does anyone know how to extract subtitles from .ts?

[netmask56]

I hope the programmers of ProjectX take the information that their project could be a bit more userfriendly.

As it is a free program and written in java I am only grateful that it was written and provided free of charge as it is the only program that will extract subtitles from MPEG2 transmission stream files. I once described it as user toxic but if you are willing to spend a bit of time experimenting with the various options and seeing what they do you can learn a lot in a short period of time. It's the non standard menu layout that confuses if you are only used to Windows of MAC - I wrote a Wiki entry on how to set it up for Australian conditions to save time for users you were confused by the myriad of options and the default settings that come with it presumably for German broadcast conditions. Once you set the options then it's really straight forward. When I get a new program the first thing I want to do is experiment with ALL the options to find out for myself what they do. So my advise to anyone using ProjectX for the first time, select a short TS file and try all the options and see what happens - you can always make a copy of the X.ini file that sets all the options as a backup.

[Clumpco]

For those wanting a simple and fast solution to extract Teletext subs from .ts files, I have written my own programme (that is a GUI for Pirayen's command line DVBTextSub) - you can install it from this page.

If you already have .Net 3.5 installed you just need to click the "launch" link, otherwise hit the "Install" button to download and install .Net 3.5.

It is somewhat quicker and simpler than ProjectX.

PS. VideoRedo V4 (new HD version) now adjusts subtitle timecodes when editing.

[netmask56]

Thanks for that - I have posted your link on the Beyonwiz forum (Australia) Just tried it on a local program (page 801) and works quite quickly. http://www.beyonwiz.com.au/phpbb2/viewtopic.php?p=92401#92401

[Clumpco]

Thanks for that - I have posted your link on the Beyonwiz forum (Australia) Just tried it on a local program (page 801) and works quite quickly. http://www.beyonwiz.com.au/phpbb2/viewtopic.php?p=92401#92401

Glad it works for you!

In your beyonwiz post you say it is java, unfortunately this isn't so - it's .Net, so you might have got non-Windows users interested for nothing.

BTW: The colour output option works seamlessly on all players that it has been tried on.

[netmask56]

Error corrected....

[petr_]

I've created tiny tool that extracts teletext subtitles from TS to SRT file. You can find it in this thread in section Subtitles.

[Geo-geo]

Hi everybody,

For graphics based subtitles you can also use ProjectX (even for HD, though it cannot decode H264), but with adequate configurration.
I have found the solution here (thanks to the author) :
https://forum.videohelp.com/threads/333392-Extract-subtitles-from-ts
Two parameters (Menu "Preferences / Subtitles) are particularly important :
- IRD color model for DVB subtitles ; the default value is 4 colours which gives awfull results. Choose a model depending of the boradcaster (In France, for the terrestrial DVB, I have chosen "UKFreeView" which gives excellent résults)
- parameter "Additional export as VobSub" must be checked.
The latter gives you a .sub/.idx file that can be used with most players in conjunction with your video file or multiplexed in .mkv files.
The problem is that, often, your subtitles will be out of sync. This can be easily fixed with "SubResync" (included in VSfilter package). Furthermore, SubResync gives you the possibility to extract a .srt file (text based) by OCR.

[elsmandino]

Hello there,

This thread seems to help with exactly what I am after.

I have a load of digital broadcast files form which I want to cut out all the adverts (and the padding from the being and the end), correct any errors and preserve the original subtitles.

I know that Project X can do all of this but not really got to grips with how to use it.

I would be really grateful for some instructions plus advice on what settling to use.

The resulting file does not necessarily need to be a .ts files, just as long as the original video, sound and subtitles are there - an MKV, for example, would be fine.

Thanks very much.

[netmask56]

Have a look at my guide (under my sig) and just substitute the correct page number for Australia's page 801. ProjectX in demux mode is the best program for doing this as it cleans up any transmission errors as well as extracting the native tracks.

[elsmandino]

Hello Netmask56.

I shall give your guide a whirl tonight.

Just to double check - if I follow the instructions, will I end up with an identical file but with all the adverts gone?

[jairovital]

I'm also in task to extract some text subs from TS Mpeg.
I've already tried everything without success.

This its MediaInfo:

General
ID                                       : 1 (0x1)
Complete name                            : B:\test\Remember.the.Night.1940.720p.ts
Format                                   : MPEG-TS
File size                                : 2.10 GiB
Duration                                 : 1h 33mn
Overall bit rate mode                    : Variable
Overall bit rate                         : 3 201 Kbps
Maximum Overall bit rate                 : 35.5 Mbps

Video
ID                                       : 4113 (0x1011)
Menu ID                                  : 1 (0x1)
Format                                   : AVC
Format/Info                              : Advanced Video Codec
Format profile                           : High 10@L3.1
Format settings, CABAC                   : Yes
Format settings, ReFrames                : 4 frames
Codec ID                                 : 27
Duration                                 : 1h 33mn
Bit rate mode                            : Variable
Maximum bit rate                         : 40.0 Mbps
Width                                    : 972 pixels
Height                                   : 720 pixels
Display aspect ratio                     : 4:3
Frame rate                               : 23.976 fps
Color space                              : YUV
Chroma subsampling                       : 4:2:0
Bit depth                                : 10 bits
Scan type                                : Progressive
Language                                 : English

Audio #1
ID                                       : 4352 (0x1100)
Menu ID                                  : 1 (0x1)
Format                                   : AAC
Format/Info                              : Advanced Audio Codec
Format version                           : Version 2
Format profile                           : HE-AAC / LC
Muxing mode                              : ADTS
Codec ID                                 : 15
Duration                                 : 1h 33mn
Bit rate mode                            : Variable
Channel(s)                               : 2 channels
Channel positions                        : Front: L R
Sampling rate                            : 48.0 KHz / 24.0 KHz
Compression mode                         : Lossy
Language                                 : English

Audio #2
ID                                       : 4353 (0x1101)
Menu ID                                  : 1 (0x1)
Format                                   : AAC
Format/Info                              : Advanced Audio Codec
Format version                           : Version 2
Format profile                           : HE-AAC / LC
Muxing mode                              : ADTS
Codec ID                                 : 15
Duration                                 : 1h 33mn
Bit rate mode                            : Variable
Channel(s)                               : 2 channels
Channel positions                        : Front: L R
Sampling rate                            : 48.0 KHz / 24.0 KHz
Compression mode                         : Lossy
Language                                 : Portuguese

Text
ID                                       : 4608 (0x1200)
Menu ID                                  : 1 (0x1)
Format                                   : PGS
Codec ID                                 : 144
Duration                                 : 1h 32mn
Delay relative to video                  : 13s 972ms
Language                                 : Portuguese

• MKVToolnix won't accept it, once it isn't MKV file.
• CCExtractor process it, but outputs a zero SRT file.
• SubtitleEdit also won't accept to import as a MKV.

• eac3to says:

TS, 1 video track, 2 audio tracks, 1 subtitle track, 1:33:42, 24p /1.001
1: h264/AVC, 972x720 24p /1.001, 10 bits
2: AAC, English, 2.0 channels, 24kHz
3: AAC, Portuguese, 2.0 channels, 24kHz
4: Subtitle (PGS), Portuguese
The source file doesn't contain any subtitle tracks in this format.  <ERROR>

• ProjectX is the better hope to solve this, but it shows this error screen:

• XMedia Recode can see the subtitle and offer two ways to handle it:

Convert -> DivX XSUB codec
and
Render -> DivX XSUB codec

Convert mode goes to crash the program at the beginning of the task.
Render mode gives this crap result:

• VLC shows correctly the subtitle, but outputs nothing, if Convert/Save process task is tried.

I know there is a valid subtitle stream inside TS file.
So, how to grab and put my hands on it?
(I mean, SRT format or other text format).

[Bonnevue]

If it's srt subtitles you want to try ccextractor.
After that you can merge the ts file with te srt file in MKVMerge
If it's dvb subtitles you want, you will need ffmpeg. I Wrote my own little programm for that.
Anyway I can help you with that, just send me a private message.

[jairovital]

Hi, Bonnevue.

I found out that subtitle inside that TS file is not a text stream, but a sup one.
Searching more, I found tsMuxerR and it could strip out the subtitle as a SUP file.

Now, I'm involved to OCR that SUP to convert it to SRT.

I like SubRip, but it reads only SUB files, not SUP ones.
Instead, I'm using SupRip, a small program that deals with SUP files.
But it is not so smart than SubRip and I have to type too much.

I tried to convert SUP to SUB and I use BDSup2Sub, but generated images are useless and not recognized by SubRip.

Anyone has a suggestion to OCR SUP files?
Or to convert in a way to SubRip recognizes the file?
I use BDSup2Sub with default parameters, but SUB and IDX was not recognized by SubRip.
In fact, it reads the file, but images are completely messed.

[Simon_Haddad]

I'm also in task to extract some text subs from TS Mpeg.
I've already tried everything without success.

I live in the US, and receive TS over the air, but it's MPEG-2 with AC-3 audio and EIA-608 subtitles (closed captions).

Still, I'd love to make an attempt. Could you upload a 10-15 second clip somewhere so I (and probably others) could try?

[netmask56]

It's such a pity that ProjectX won't handle H264 formats and nobody has attempted to rewrite ProjectX to be more universal, however I believe VideoRedo can handle your file - I don't have it so can't help there also TS Doctor latest version should be able to do it. If you an upload the first 30 seconds or so some of us can have a play with it.

[KarMa]

Hi, Bonnevue.
Anyone has a suggestion to OCR SUP files?
Or to convert in a way to SubRip recognizes the file?
I use BDSup2Sub with default parameters, but SUB and IDX was not recognized by SubRip.
In fact, it reads the file, but images are completely messed.

Subtitle Edit is the best because it uses tesseract. Install it and drag and drop the file. In the OCR window that pops up, choose tesseract OCR and click to download the latest files for that language.

Tesseract will still make a few mistakes, so you can tell it to ask you for all unknown words. Words it can not find in its dictionary. Also click on the italics box if the subs have italics, or else it will mess up.

[PeterDuke]

I used Subtitle Edit to extract teletext (text, not image) subtitles from MPEG2 TS files. I removed unwanted start, end and advertisements (commercials) first using VideoReDo.

[PeterDuke]

--

[ProWo]

I live in the US, and receive TS over the air, but it's MPEG-2 with AC-3 audio and EIA-608 subtitles (closed captions).

EIA-608 subtitles can be extracted by clever FFmpeg-GUI.