subtitle-me v1.0.8

subtitle-me

Subtitle generator for videos using IBM Watson Speech to Text

This is a program written in Node.js used to generate a SubRip .srt file from an .mp4 video by using the IBM Watson Speech to Text service on IBM Cloud.

English is currently the only supported language

Prerequistes

You need to have Node.js installed on your machine. If you don't have Node.js, then you can download it from nodejs.org. We use the latest Node LTS version, Node 10.

In order to be able to use this program, make sure you have ffmpeg installed on your system (including all necessary encoding libraries such as libmp3lame or libx264).

This project relies on the fluent-ffmpeg package, which requires that you have a ffmpeg version greater than 0.9. The fluent-ffmpeg package will call ffmpeg and ffprobe so you need to have these in your PATH or set in the FFMPEG_PATH environment variable and the FFPROBE_PATH environment variable. This program will automatically convert your file to a .mp3 file so you must have the libmp3lame codec installed on your system.

You must also create an IBM Cloud account and create service instances for Watson Speech to Text.

Usage

Set the IBM_CLOUD_API_KEY environment variable in your terminal to your Speech to Text API key from IBM Cloud.

Use npx to run this program without cloning the repository and installing dependencies. It ships with NPM so no need to install.

MacOS and Linux

IBM_CLOUD_API_KEY=your-api-key npx subtitle-me

Windows

set IBM_CLOUD_API_KEY=your-api-key npx subtitle-me

You can indicate whether or not sentence casing should be performed. By default subtitle-me will capitalize the first letter in the first word of a subtitle and add a period to the end of the subtitle.

Once the command has executed, it will create a file named the same as the video filename except with the .srt extension. A raw speech events file will also be created and has the same name as the video file except it will append on _events.json. Additionally, an .mp3 file will be created that contains the extracted audio from the video file. It will be named the same as the video filename except that it will end in .mp3.