Skip to content

Latest commit

 

History

History
89 lines (55 loc) · 5.97 KB

websocket-clients.md

File metadata and controls

89 lines (55 loc) · 5.97 KB
Raw

Websocket Clients

All Interactions with the Application besides the Speech recognition and OSC is done using Websockets.

Currently, all Websocket Clients are provided as HTML files. They should run on every relatively modern Browser that supports Websockets. (see support table on caniuse.com)

You have build a nice Client and would like to share?, let me know so i can add it to this list.

Using the Websocket Clients

  • Simple websocket_clients/simple

    Is the simplest implementation.

    Mostly intended to show how to use it without much additional logic around it, to make it easier to build your own Overlay.

    Supports only the ws_server URL argument.

  • Streaming Overlay 01 websocket_clients/streaming-overlay-01

    Has a lot more around the actual websocket implementation. Including a Styling that should work well for Streaming Overlays.

  • Streaming Overlay 02 websocket_clients/streaming-overlay-02

    A rewritten implementation that looks more like traditional subtitles.

  • Websocket Remote websocket_clients/websocket-remote

    Is the most complete Websocket implementation with additional Configuration Options that can be changed without restarting the application.

    Clicking on any Transcription Box feeds the Transcription and/or Translation into the textual translation Fields to Translate it into another language, edit it or send again over OSC.

All Clients Support some options that can be set without changing the HTML

using Parameters added to the URL in the form:

index.html?ws_server=ws://127.0.0.1:5000&no_scroll=1&auto_hide_message=30

(first option is added with a ?, all following with a &)

All possible Configuration URL arguments

Arguments Default Value Description
ws_server ws://127.0.0.1:5000 Sets the Websocket Server IP and Port to connect to.
no_scroll None Removes the Scrollbar if set.
bottom_align None align to bottom of page (also removes scrolling)
no_loader None Disables the audio processing indicator if set.
auto_hide_message 0 Sets the Seconds after which Transcriptions are hidden, but not removed. (Can be shown again by clicking in the lower part where the Boxes appear) 0 = Disabled
auto_rm_message 0 Sets the Seconds after which Transcriptions are removed. (makes it impossible to access them again) 0 = Disabled
show_transl_results websocket-remote=both
streaming-overlay=auto		 Sets what is shown as translation / transcription result. Can be:
both = shows both whisper and text translations.
whisper_only = shows only Whisper results.
auto = shows text translation if available, otherwise Whisper result.

Use-Cases

Parallel processing of A.I. Transcriptions

Since it is possible to run multiple AI's concurrently, it is possible to combine multiple Websocket Clients.

For Example, you can create a HTML Page with I-Frames that show the different Websocket Clients.

Following example shows on the left the local remote control webpage and on the right the streaming overlay page connecting to a secondary PC.

(but could also be the same PC using a different websocket port).

<!DOCTYPE html>
<html>
   
   <head>
      <title>Websocket Client Frames</title>
   </head>
   
   <frameset cols = "60%,40%">
      <frame name = "left" src = "audioWhisper/websocket_clients/websocket-remote/index.html" />
      <frame name = "right" src = "audioWhisper/websocket_clients/streaming-overlay-01/index.html?ws_server=ws://192.168.2.136:5000" />
      
      <noframes>
         <body>Your browser does not support frames.</body>
      </noframes>
   </frameset>
   
</html>