User:Thomascharlesbridges/CaptionHub

CaptionHub

CaptionHub is a commercial cross-platform caption and subtitle editing program. It is browser-based, and has built-in tools for quality control and validation. It makes extensive use of speech recognition to assist and speed up the process of captioning and subtitling.

Browser-based
CaptionHub is a web application that works in most modern web browsers. It has a rich HTML 5 timeline editor and collaborative workflow tools, allowing individuals and teams to create original captions, and to manage the subtitle translation process.

Being based in the cloud is allows large teams to work in multiple time zones. It also simplifies the process of version control: everything is only ever in one place, and changes are logged. Multiple people can log onto the same project and work on it alongside other users.

Quality control
Captions and subtitles are often created to a specification. For example, many broadcasters will stipulate a minimum duration that captions must appear on screen. CaptionHub continually analyses captions, and warns the user if any specification limits have been breached.

Subtitle validation
For subtitle translation, users can be assigned viewing rights or editing rights for a particular video. At any one time, many people can have viewing rights, but only one person can have editing rights. Viewing rights can be transferred between users, and the approval process can be customised. Email serves as a notification system and a snapshot history provides a record of the changes each user has made. This can be used for educational or audit purposes.

Speech recognition
CaptionHub makes use of speech recognition to analyse the audio, and return an original language transcript. This transcript is then processed by CaptionHub in order to create automatically generated captions, timed alongside the video. Additionally, if a transcript is supplied, then CaptionHub can use speech recognition to align that transcript with the video. In both instances, CaptionHub attempts to split up captions in the same way that a human would. For example, captions won't end on a conjunction.

Features

 * Input / output formats: WebVTT, SRT, TTML, SMI, SCC and others
 * Integrated burnt-in subtitle creation