feature request: various output formats #39
Description
Right now the only possible output format is *.txt. One can put the *.json-Extension in the filename, but the result is a "JSON"-File without the inner structure of a whisper-json-file and without the additional information, that usually comes with whisper-json-files.
Whisper offers a multitude of different output formats, of which especially the JSON-Format is very useful: If the parameter for word segmentation is set in Whisper (which could be the default when using JSON-Output), the JSON-File contains the transcribed chunks plus a finer segmentation per word. The important thing is that the file contains time stamps for the transcribed words/chunks.
The transcript can then easily be imported into transcription software like "Partitur-Editor" of Exmaralda for alignment with the audio and an easy correction of the transcription.
So the request would be to be able to chose between output formats (and - if JSON is chosen - to include the segmentation per word as default).