Guide: How to remove

text for the hearing impaired and lyrics

in DVD subtitles

 

By CoNS.

 

The guide is written for DVDSubEdit v1.3 and above.

All thanks go to Jeanl for this marvelous application.

 

Questions, comments, bug reports and suggestions for improvements

regarding the guide are most welcome in the doom9 forum thread.

 

 

 

Introduction

 

Many DVD subtitle streams, and especially English subtitles, include text for the hearing impaired (also called “subtitles for the deaf and hard of hearing” or simply “SDH”). Example:

 

 

While these subtitles obviously can be very useful for hearing impaired persons, they are often annoying to people who don’t have a hearing handicap.

 

Lyrics are a similar kind of “explanatory” or “informative” subtitles that often occur in DVD subtitle streams. Example:

 

 

People who are not hearing impaired would often want to get rid of such subtitles with lyrics, too, as they tend to distract people while watching the movie.

 

The scope of this guide is to show you how to remove DVD subtitles for the hearing impaired and subtitles with lyrics quite easily using the freeware program DVDSubEdit by Jeanl.

 

If you are already familiar with the way DVDSubEdit works, this is the suggested procedure in the guide, in short:

 

       a.   Run OCR.

       b.   Setup and run the automatic removal of text for the hearing impaired.

       c.   Filter subpics with these characters: “[]()#:”

       d.   Browse through the filtered subpics and select any unwanted text and erase it.

 

If you are new to DVD editing in general, and have a hard time understanding the user interface of DVDSubEdit, I highly recommend reading the user manual for the application before proceeding with the guide.

 

 

 

More examples of subtitles for the hearing impaired

 

Text for the hearing impaired and lyrics may appear in many different forms and variations. Unfortunately there is no official standard for these kinds of subtitles in DVD subtitle streams.

 

Usually they are not even placed in separate subpics, but mixed with “normal” subtitle text with the dialogue in the movie. Example:

 

 

Like in the above example, the text for the hearing impaired is often wrapped in square brackets or parentheses.

 

A different way of displaying text for the hearing impaired is when the name of the person speaking is placed in front of the “normal” text in the subpic along with colon, for example like this:

 

 

Lyrics and titles of songs often appear with one or more music notes, and may be combined with text for the hearing impaired, like in this example:

 

 

Music notes may also just indicate that someone is singing in the movie, without displaying the lyrics, like here:

 

 

In all of the above examples, DVDSubEdit is able to detect and remove the subpic (or part of the subpic) containing lyrics or text for the hearing impaired, following the steps described below.

 

 

 

Step 1: Open the DVD and select a subpic stream

 

Make sure you have the DVD on your hard disk as IFO, BUP and VOB files.

 

Choose “Open full domain” from the File menu.

 

Browse to the folder containing the DVD files and select any of the VOB files from the main movie VTS.

 

(Alternatively you can choose “Open VOB file(s)” in the File menu and mark and select all of the VOB files from the main movie VTS. Or simply drag-and-drop all of the VOB files from the main movie VTS to the application window.)

 

Tip: The main movie VTS is usually the largest set of VTS files in the folder. You may want to use the standard “Show details” option in the Windows “Open file” dialog to view the file sizes.

 

After loading the DVD you can either keep the default subpic selection, which is “All Streams”, or you can select a specific subpic stream to edit.

 

In the example, I’m selecting an English subpic stream. “WS” indicates that the stream is to be used when viewing the movie in WideScreen format (“LB” means LetterBox format).

 

 

 

Step 2: Run OCR

 

Select “Run OCR” in the OCR section of the main window (or hit CTRL+SHIFT+T) to convert all the bitmap subpics in the selected stream to text.

 

This is done via the Optical Character Recognition (OCR) function integrated in the application.

 

When the OCR function is done, you’ll see the OCR output of the current subpic in the OCR output field.

 

In the example, the two “#” characters represent music notes in the bitmap.

 

 

 

Step 3: Setup the automatic removal function

 

Choose “Preferences” from the File menu to display the preferences window.

 

The options in the preferences window are pretty much self-explanatory.

 

I recommend enabling the first three of the four removal options in the “Hearing-impaired text removal” section.

 

 

Notice that if the removal function leads to unintended results, you can always undo and change the removal options accordingly. However, it can be difficult to discover such unintended results until actually watching the movie.

 

I do not recommend enabling the last one of the removal options. In my experience there’s a higher risk of having “normal” text unintendedly removed with this option. For example, if the option is checked, it will unintendedly remove “One: “ in the lower line in this subpic:

 

 

The last two options in the “Hearing-impaired text removal” section (re-align and re-center) are rather self-explanatory, too. Usually you would want to leave both of these options on, but you should always browse through a few of the subtitles in the selected stream to see if the original subpics are aligned at the bottom and are centered by default, and adjust the two options accordingly.

 

Click “OK” to close the preferences window.

 

 

 

Step 4: Run the automatic removal function

 

To execute the automatic removal of the unwanted text and lyrics, choose “Hide hearing impaired text on all selected subpics” from the OCR menu.

 

 

The function changes the bitmap subpics and automatically performs a new OCR scan of the affected subpics so that the OCR output reflects the changes.

 

After the automatic removal function has finished, you can compare the changed subtitles with the original subtitles.

 

First, use the “Only show modified subpics” option to filter the subpics that have been changed by the automatic removal function.

 

Then, check the option “Show original” in the OCR section in the main window to see the original OCR output of the modified subpics.

 

 

 

 

The limitations of the automatic removal function

 

Overall, the OCR function and the function to automatically remove lyrics and hearing impaired text are very intelligent and robust.

 

However, they have certain limitations, due to the many, many possible variations of the look and syntax of DVD subtitles. There are some quite common instances where the automatic removal function does not remove unwanted lyrics or text for the hearing impaired.

 

For example subpics with lyrics that only have one music note, like this:

 

 

In the above subpic the syntax is “# lyrics”, i.e. the lyrics are not “wrapped” in music notes (syntax “# lyrics #”). The automatic removal function would not know where to stop the removal in these subpics, and thus they are not affected when running the function.

 

Another type of problematic subpics is when one of the characters used for “wrapping” lyrics or text for the hearing impaired is somehow not recognized in the OCR scan. As an example, “[” in “[ text ]” is sometimes recognized as “f”, and “]” as “J” etc.

 

Faulty OCR outputs mostly occur in subpics with italic text, and especially when italic text and normally formatted text is mixed together in the same subpic. Like here:

 

 

Furthermore, if the fourth removal option in the setup window is unchecked, you will also have the “Man: Now way!” type of subpics left after running the automatic removal function, if such subtitles were present in the original subtitle stream.

 

 

 

Step 5: Filter subpic with remaining unwanted text

 

DVDSubEdit has some tools which can be used for manual removal of unwanted subpics and parts of subpics that the automatic removal function didn’t remove in step 4.

 

Any occurrences of the above mentioned types of subpics with remaining unwanted text can be displayed by using the “Select subpics with these chars” function.

 

By typing in the square bracket characters “[“ and “]”, and the paranthesis characters “(“ and “)”, the program will filter subpics where the OCR function has not recognized one of the two characters used for “wrapping” text for the hearing impaired.

 

By typing in “#” the program will display any remaining subpics containing lyrics, including subpics with the syntax “# lyrics”.

 

And finally by typing in “:” you will be able to see any “Man: Now way!” type of subtitles for the hearing impaired.

 

You can now browse the filtered subpics by using the slider, the arrow buttons or the arrow keys on your keyboard.

 

Notice that both the “:” and “#” characters in some cases may be used as a part of “normal” subtitles, which you may want to keep untouched.

 

Also notice that this filtering method cannot help you locate subtitles with remaining parts of lyrics or text for the hearing impaired in all cases. For example, lyrics are sometimes displayed in “normal” text subtitles without any music notes or other indication that they are lyrics, making them impossible to detect in an automatic process.

 

Furthermore, in some rare cases with very special font styles and font sizes the OCR function will not be able to recognize any one of the two square brackets or parentheses “wrapping” the text for the hearing impaired. The same can happen with colons and music notes, too.

In these cases you can only be sure to detect the subpics with the unwanted text or lyrics by browsing through all of the subpics one by one.

 

Tip: If the key characters “[“, “]”, “(“, “)”, “#” or “:” are consequently erroneously recognized in a specific subtitle stream, you can try and use the “Select subpics with these chars” function to filter the subpics with the faulty OCR output for these characters!

 

 

 

Step 6: Manually hide or erase unwanted text

 

After locating the remaining subpics which contain unwanted text or lyrics in step 5, you can discard it using two different tools in DVDSubEdit.

 

If the entire subpic is unwanted, like the example above with “# Hey, Macarena”, you can hide the subpic by clicking “Hide subpic” in the main window (or hit CTRL+H).

 

This function will set the transparency to “clear” for all four pixel types that are used in the subpic. The subtitles in the subpic will still remain in the subpic stream, but they are now invisible.

 

If the unwanted text or lyrics only make up a part of a subpic and you want to keep the rest of the subpic, you can use the manual erase function instead.

 

In the example, the OCR function has erroneously recognized “[“ from the bitmap subpic as “f”.

 

To erase the text that reads “[Sebastian] “ in the bitmap subpic, select the corresponding text in the OCR output field and click “Erase selected text” (or use the Delete key or backspace key) to erase it.

 

Tip: You can place the cursor in the OCR output field and use the arrow keys to move the cursor, and the backspace and Delete keys (or CTRL+X) to remove text.

 

Notice that when erasing text, the remaining text in the subpic is re-centered and re-aligned, if the two last options in the “Hearing-impaired text removal” sectioin the preferences window are enabled (see step 3 above).

 

Repeat the procedure described in this step for each subpic with unwanted text or lyrics you find when browsing the filtered subpics.

 

 

 

Step 7: Save and test

 

When you’re done editing the subtitles, you can save the changes by choosing “Save all modifications” in the File menu (or hit CTRL+S).

 

Warning: The program will overwrite the original files, and no backup is made.

 

This means that there’s no undo function to revert the changes after saving.

 

 

Tip: You may want to manually back up your files before saving the modifications.

 

Also, when editing DVDs it’s always recommend to test the result in a software player before burning the disc. Or burn to a rewritable disc and test in your standalone player.