How to create your own "transcript" from a Chinese video

[lingshu8]

Hi All,

I wanted to share a method that I use for extracting the Chinese hard subtitles from AVI, FLV and RMVB videos. Most of the Chinese videos available on Youku, Emule, Youtube et. al. have "hard subs" which means that the subtitles are "burned" directly onto the image and therefore cannot be extracted or opened separately. I use an excellent program called AVISubDetector (freeware available here: http://www.videohelp.com/tools/AVISubDetector). You can also download a guide here: http://tinyurl.com/ylcerzv. Here's a brief overview of how to use AVISubDetector to make "transcripts" out of hard-subbed videos (the guide goes into more detail):

1. Load an AVI file with Chinese subtitles into AVISubDetector (if the original file is RMVB or FLV you'll need to convert it to AVI first using a program such as AVS Video Converter).

2. Under "Settings," crop the image so that it just includes the subtitle area at the bottom of the frame.

3. Under "Project" select an output folder.

4. Click "Start (Full)"

The program will automatically detect the subtitles, crop them, and save them as a series of small BMP image files in a subfolder labeled "SubPic." It's that simple. The subtitle BMPs are also automatically numbered. You can easily batch-insert them into a word-processing document or combine them into a PDF file which will basically give you a "transcript" of the video. You can also drag & drop individual subtitles into ANKI to make flashcards (unfortunately, since the BMPs are image files you cannot paste them directly into a dictionary program like Wenlin). Personally, I use another free program called PdfLrf to combine the subtitle images into a Sony Reader LRF document which I then load in my Sony Reader (a Kindle-type device -- reading computer screens for a long time makes my eyes ache).

Anyway, I hope some of you find this useful. I've attached a sample which actually came from a fairly low-resolution source (RMVB of 家有儿女 downloaded off EMule)

[liuzhou]

Thanks for that. I can see that being useful.

[roddy]

Have you tried then running the ,bmp files through OCR to get a soft-copy - that would be very handy indeed.

[renzhe]

OCR would probably have trouble with this kind of input, but I can imagine that even just re-typing this stuff is easier than having to constantly pause, rewind, and all that other stuff that you need to transcribe an episode using the standard method.

[AxelManbow]

That's very interesting. I might use that.

[Guest realmayo]

Thanks very much, I will use this.

[roddy]

Also, if you could keep track of the timings you could then use it to create soft subtitle files - either of the original or translations. Does it do that? That would also be handy.

[abcdefg]

Looks helpful. I'll give it a try. Thanks.

[Guest realmayo]

Well I've failed on this right from the start: although I believe I may have loaded an .avi file into the programme, I've got no idea how to get it to show me a frame, so I can mark out the boundaries for the subtitles.

[pom45i]

Yeah, you can use this technic to keep the timing information. I am using a similar method to only save the timing then upload it on viikii or create the softsub directly by adding the translations. http://www.viikii.net/my/page/Pom45I#TOPIC/527.

[lingshu8]

@realmayo

Well I've failed on this right from the start: although I believe I may have loaded an .avi file into the programme, I've got no idea how to get it to show me a frame, so I can mark out the boundaries for the subtitles.

Did you download the guide? Here's the full URL:

http://www.wxfs.org/guides/HowtoRiptheTimingandEnglishSubsFromanAVIfileUsingAviSubDetector.pdf

Especially look at p.5 of the guide where it describes how to use the "Settings" tab, which is where the "Crop Settings" are. You have to slide the Top and Bottom crop settings until they just cover the subtitle region.

[arreke]

What shall I do to make AVISubDetector work properly ?

 

 

I thought it might be the space distance that is too small, so I changed it, still no effect...

[CH09]

Thank you

[jiangsen]

Anything comparable to sub detector on a Mac?