Last updated:

Quick and Easy Subtitling from a Transcript using Reaper and Subtitle Workshop

ReaperSubtitles
This article is part 2 of 3 in the series Le Sous-Titrage Facile avec Reaper
  1. A Detailed Guide to Subtitling with Reaper
  2. Quick and Easy Subtitling from a Transcript using Reaper and Subtitle Workshop
  3. Better Subtitling with SWS v2.6.0 and Custom EEL Scripts
This articles is dedicated to REAPER v4 and is, since August 2015 v5 release, considered as obsolete. For a more modern workflow please refer to this article :

REAPER v5, Text Items and Scripts: Subtitling, Faster Than Ever! / ExtremRaym

In my article A Detailed Guide to Subtitling with Reaper, I described different ways to writte subtitles easily and quickly, with the DAW Cockos Reaper.

Because my first article is one of the most viewed on my website, I decided to push the thing even further, and to show you a practical exemple, in reel time, of the Subtitle a one audio track project of section my previous article, a very important chapter with too low information… until now ! So, I made a screencapture of me subtitling a whole video, based on the transcript, in order to show you how I did, clicks after clicks !
To put my hand to work give me a lot of ideas to improve the workflow I described at this time (via mouse modifiers, custom actions – under Reaper, via Regex under Notepad ++ etc…). It takes me 4 whole different videos to establish an optimal workflow.
The fourth video is the one you will see in this article.

In order to combine the pleasant and the useful (or the useful to the useful), I decided to subtitle videos that really speak to me, because I find their subject and their message interresting.

The videos came from the YouTube -not enough famous but brilliant- show Extra Credits, an awesome show speaking about game video theories and concepts (game design, scenario and interactivity, choices…), with 5 to 10 minutes episodes, illustrated with original drawings and photos from the net.

What is convenient, it’s that there is a team dedicated to the Extra Credits episodes translation, which provide transcripts that need to be synchronized and translated.

The video that will be subtitled in this tutoriel is : Extra Credits: Energy Systems.

Here is the result on the online collaborative subtitles service Amara :

It is pretty nice for 40 minutes of work, isn’t it ? 😀

Enough said, let’s go serious.

Tutorial Details

  • Duration : Video duration multiplied by a number from 6 up to 8
  • Difficulty : advanced
  • Requirements : Basic knowledge of Reaper, use of my Reaper Config file X-Raym.ReaperConfigZip, have a transcript from the media to subtitle
  • OS : Windows
  • Softwares :

Reminders

Subtitling is 3 major steps :

  1. Transcript, in other words, the writting of all words pronounced in the media, without any time-code information.
  2. Synchronisation, in other words, the timing placement of the sentences from the transcript, together with the media to be subtitled
  3. Translation, the conversion of the subtitles in an other language, which can be made with an already synchronized subtitle file

The steps 1 and 3 doesn’t require any particular information, we will focus on the steps 2, which appear the most difficult and less accessible to a lot of people.

Video

I decided to not speak during the video, in order to be fully focused. That said, small unforseen events happend, but nothing that will disturb you.
It would be very hard to make a 40 minutes video long without making any misclick !

The pie chart below show you the time proportion of each steps of the process.
[gdoc key=”https://docs.google.com/spreadsheets/d/1i7sbM7fwedMwWyVu1Z2cQtVH_NUIzRb_47s7UYzyMbc/edit#gid=451416573″ chart=”Pie”]

As you can see, more of 50% of our time is spend in Reaper, which proves it’s importance in this workflow.

Explaination

Preparation (facultative)

Timecode : 0s
The video came from YouTube.
In order to download it, I use the great Firefox plugin YouTube Center (wich is awesome for a lot of other reasons).

Synchro

It is precisely where Reaper will save you a lot of time compare to a classic subtitle software. With a precise display of the waveform, you will be able to define begining and end of subtitles quickly, without having to listen to the audio track. More, Reaper is really ergonomic and allow you to zoom and move in the project very easily.

Timecode : 17s
I start by opening my “Video” project template, and I set the frame rate to match the frame rate of the video.
Snaps and grid are enable on Frame mode. Synch with frame is very important in fiction and documentary, if you don’t want your subtitles to disturb the viewer attention when there is sequence change, but less with audio commentaries or conferences. Still, I think that working with snap on frame is a good reflex to have, but to really synch with shots cuts of the video requires extra patience and time.
I set the project duration to something a little superior to the duration of the video, in order to have a minimal zoom level which corresponds to what I need.

After the video importation, I set the stereo audio track to mix mono (Left + Right), in order to see the waveforms much clearly.

You just have to define regions that will allow subtitling, considering the flow of speech. I made regions start from a begining of speech and ends at pauses/breathing, in avoiding too shorts regions.
In order to define regions quickly, I use two custom actions set on the mouse :

  1. Shift + Left clik on the bottom half of the object => Remove Time Selection, Move Edit Cusror to Mouse Cursor, Time Selection: Set Start Point
  2. Shift + Middle click => Time Selection: Set End Point, Insert Region from Time Selection, Remove Time Selection

Don’t worry if the subtitles appear to long, they will be split in a subtitle editor.
Often, your region will simply follow the punctuation of the transcript, which correspond to breathings, semanticly good points to split subtitles.

Notice that I often keep the region a bit more long than the actual flow speech, in order to be sure that the viewer will have time to reed.

After you finished, check if there is no forgotten sentances.
You can execute the action Renumber all markers in timeline order if you add other regions. You will not be able to execute this action after that moment, cause it may invert subtitles.

Paste Dialog Lines

From the transcript, I will copy/paste dialog lines in the regions subtitles of the Notes Window, avaible with the SWS extensions.

Timecode : 8min 53s
I use DeskPins, as I already show you in anther article, in order to keep OpenOffice at the top of the other windows.

Timecode : 9min 50s
On a toolbar, I set the Stop Playback at the End of the Time Selection mod, and I use two buttons with custom actions in order to Go to the Next Region, Insert Region from Time Selection, then Play.
This allow me to play a region (and this region only) in order to see that is inside, which is extremy useful in this worflow.

Timecode : 10min 10s
Copy/Paste work can begin.

Timecode : 22min 50s
Before exporting, check if every region has its subtitle.
There was a small mistake in my case, that I could fix.
Note : as i already said, don’t execute the Renumber all markers in timeline order at this point, because as you can see, it makes some subtitles inversion. I learn that during the video !

Breaking Lines

Timecode : 25min 05s
I use Notepad ++ in order to break lines very quickly, in minimising the number of clicks.
I created a synthax highlighting color UDL file for SRT files, which is very convenient.

Timecode : 25min 11s
I set the display of a line (column limit) at 45 characters in order to have a rough idea of where I have to break lines.
Timecode : 25min 22s
The breaking lines work can start.
Of course, you will have to evaluate the right balance between semantic places and line lenght for finding the right place to break the line.

Timecode : 35min 23s
I end up by using two Regex expression in the Find/Replace window in order to delete unwanted spaces at the begining or at the end of a line.

  • Replace ^\s by nothing.
  • Replace \s\r by \r\n (it is possible that you have to execute it several times).

Timecode : 35min 42s
For the moment, Subtitle Workshop doesn’t handle the encoding format UTF-8, as in SRT files exported by Reaper.
Changing the encoding format of a document can create problems. In order to avoid them, I select all the text, I cut it, I change the enconding format to ANSI and I paste it back.
Your file is ready for Subtitle Workshop.

Timecode : 35min 57s
During my screencapture test sessions, Subtitle Workshop has trouble when the SRT file was loaded with a video. To avoid that, I rename my SRT file with a different name that the video name. It is possible that you don’t have to do this.

Lines Division

As you notices, some subtitles have a too much lines. We will use Subtitle Workshop, which provides lines division tool based on line breaks and set the duration of the new subtitle created in function of an optimal charachter/second speed.

Timecode : 36min 04s
I execute the Information and Errors tool : blue subtitles are those with too many lines.

Timecode : 36min 10s
You just have to select thoses too long subtitles individually and press Ctrl + Maj + D to launch the Divide Lines window.

Some other errors can stay (too many characters in a line, duration too small…), you will have to evaluate if you care about it or not. Some will really need correction (merge, division), some don’t.

Timecode : 39min 00s
I save the document.
And this is it ! Your subtitles are ready !

Amara import (facultative)

Timecode : 39min 30s
Amara is a online collaborative subtitles service. You add a video from another website, you upload subtitles and HOP everyone can make a translation !

Note : If you writte your transcript in a word processor software as Microsoft Office, it is possible that some characters (accent, quotes etc…) are transformed in a more elegant version. So, ' can become , "..." can become “...” or « ... » etc…
The problem is that these special characters are not equally supported by the platform you will choose for uploading your subtitles, which is the case for YouTube if your file is in ANSI.
You then have two possibilities : replace all the character which can cause problem, with a succession of Find/Replace in Notepad ++, or in converting the SRT file back to UTF-8, which is a far more faster and better solution, you will need to check if it’s good.
Upload to Amara doesn’t require such conversion.

Timecode : 40min 15s
I add my signature to the SRT file. 😛

Other Examples

Here is other videos I subtitled with these methods.
They speak mostly about Games and Education, a subject that interest me a lot, which ideas that need to be spread.

Here is the Tangential Learning Video :

I also made the Power Creep video.

and Gamify Education :

Conclusion

As you can see, Reaper and its SWS Extensions are perfectly integrated in subtitling tasks, with the right shortcuts, mouse modifiers, toolbars and custom actions, providing powerful tools as waveform display, macros, etc… and having a really good ergonmic feelling rarely find in a subtitle software.
Reaper definitely allows to subtitle media easily ! 🙂

If you find that the Reaper workflow is good, check my Feature Requests for Cockos Reaper SWS Extension SWS, in order to increase to power of Reaper about subtitles !

Thanks to all,

And good subtitling !

Reaper Forum Thread : A Detailed Guide to Subtitling with Reaper

This article is part 2 of 3 in the series Le Sous-Titrage Facile avec Reaper
  1. A Detailed Guide to Subtitling with Reaper
  2. Quick and Easy Subtitling from a Transcript using Reaper and Subtitle Workshop
  3. Better Subtitling with SWS v2.6.0 and Custom EEL Scripts