Automatic subtitle generation for sound in video files

[yaswanth]

I selected AUTOMATIC SUBTITLE GENERATION FOR SOUND IN VIDEO FILES as my major project and when I searched on internet I found some documents explaining about this topic and the end result is not satisfactory so I want to know is it possible using java? Result doesn't have to be accurate but is it possible to get 80% success rate if audio is clean and 50% for a normal video? or have I done mistake by selecting this project?

[KevinWorkman]

You're asking a question that's way too vague to really answer. What approach are you planning on taking? What have you done so far? What have other people done?

If you really have no idea how to proceed, then perhaps you did bite off more than you can chew.

[JosAH]

I smell speech to text transformation and that means speech recognition (a nasty subject); I don't know if it can be done where actors speak (or shout, or whisper) simultaneously; the other way around, i.e. text to speech, is much easier. Maybe Google knows more ...

kind regards,

Jos

[yaswanth]

I wanted to approach this project in three steps.

1. Audio extraction from videos
2. voice to text using speech recognition
3. generate subtitle

Thus far I didn't got much further in this project sorry for the absence of reply.

though I got the basic knowledge in approaching those modules.

I am thinking of using Xuggler to extract audio and by using Sphinx-4 which is an application completely written in java which helps us in speech recognition module and I am thinking to generate subtitle by recording time in the video where the voice has started and stopped by using acoustic principles.

Now I need help with extracting audio from video using Xuggler or any other optimum solution.

[broskully]

I am also doing my project in automatic subtitle generation in videos. i have done audio extraction ,speech recognition with sphinx 4 but am having problem generating subtitles.Any java code to implement this and how to go able it

[Tolls]

If it's recognising the speech and turning it into text, then where's the problem?

[yaswanth]

If it's recognizing the speech and turning it into text, then where's the problem?

Ya. I am able to convert speech to text using Sphinx 4(barely though). But the real problem is with syncing the text with audio.
Generating .srt(subtitle) file is easy but syncing is the real problem.

[yaswanth]

I am also doing my project in automatic subtitle generation in videos. i have done audio extraction ,speech recognition with sphinx 4 but am having problem generating subtitles.Any java code to implement this and how to go able it

Are you able to sync audio with text? If yes tell me how to do it.

generating subtitle is easy with this

FileWriter fileWritter = new FileWriter(file,true);
            	        BufferedWriter bufferWritter = new BufferedWriter(fileWritter);

I am studying JMF for audio and text syncing.

[broskully]

how to convert speech to text using Sphinx 4 in subtitle generation using sphinx4 and and how to create the subtitle file using a java

[yaswanth]

how to convert speech to text using Sphinx 4 in subtitle generation using sphinx4 and and how to create the subtitle file using a java

To convert from speech to text use Sphinx 4. Download sphinx 4 from here

to get the overall idea about how to use sphinx 4 download source version.In this version complete program code is available.

watch out video tutorial here for how to execute sphinx 4 program.

Out of curiosity may I ask you why you are interested in this project? general or project purpose?

[broskully]

I am doing my project work in automatic generation of subtitles in videos

[broskully]

Thanks but I am also doing my project in automatic subtitle generation in videos. i have done audio extraction ,speech recognition with sphinx 4 but am having problem generating subtitles.Any java code to implement this and how to go able it

[yaswanth]

Thanks but I am also doing my project in automatic subtitle generation in videos. i have done audio extraction ,speech recognition with sphinx 4 but am having problem generating subtitles.Any java code to implement this and how to go able it

Are you able to sync audio with text generated? If not you need to study more about JMF and come up with your own code.
if you just want to write generated text into .srt file then I did it like this.

File file =new File("F:\\subtitle.srt");
if(!file.exists())
{
  file.createNewFile();
 }
//true = append file

            FileWriter fileWritter = new FileWriter(file,true);
            BufferedWriter bufferWritter = new BufferedWriter(fileWritter);
            bufferWritter.newLine();
            bufferWritter.write(data);// here "data" is my generated text

[broskully]

Thanks
But what if you don't have a subtitle and want to generate one from scratch using sphinx4 and java

[yaswanth]

Thanks
But what if you don't have a subtitle and want to generate one from scratch using sphinx4 and java

I just told you that. the above code generates new subtitle file.