NSI is a project to search through many caption files (SRT) at once for Non-Speech Information
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.



NSI is an acronym for Non-Speech Information, a generic term for any information conveyed by captions that isn't a pure transcription. This is an application that stores caption files (.SRT), and makes it easier to search through those files for NSI. The primary uses of this application are to find Non-Speech ((ALARM RINGING)), to find Speaker IDs (TOM: hello world), and to find musical notes, but full text search is supported as well.

This tool was written for my father's use during his research for his book, Reading Sounds

This tool is not very fast, and would be unbearably slow with more than ~100 files. It has a "database" -- literally a folder with SRT files in it. It's not secure, and it's hastily written. That said, it works OK.

Compiled releases can be found here. The executables must be launched from a command line, and won't work via double-clicking. (cmd.exe/terminal.app)


Index page shows a list of caption files

Clicking on one file shows a page to search in just that one file

Clicking on the search button at the top lets you select potentially many files.

The results from a search are split into 3 categories. Non-Speech, Speaker IDs, and Other.