Improved support for the a11y technology Speech-To-Text

Blastoise186
Discord has been great wtih setting up a11y features and improvements so far. One thing that would really help is improved Speech-To-Text Support for programs like JAWS and Windows Speech Recognition - this is NOT the same as using Cortana in Windows 10. At the moment they don't play well with Discord, and it can be frustrating for some users. I found out for myself when I did an experiment a while ago using Windows Speech Recognition with Discord Canary. While it was usable, the experience was by no means as elegant as I would have liked it to be, not least because Windows Speech Recognition had a ton of trouble trying to pinpoint what to click on whenever the "Show Numbers" command was used. It would be great if Discord could put in some neat tricks that would allow the apps to play nice with a11y Speech-To-Text programs.
22

Kommentarer

3 kommentarer

Datum Röster
  • bastet_of_orion
    Speech-To-Text is super important, and it seems to be a bit of a blind spot with Discord currently.
    2
  • DoubleD-1397

    Why no support for the hearing impaired? I invited a clan-mate to get on discord, and he said he was deaf, so he couldn't.  Man, what a bummer that was!  I felt like a total idiot.

    Does anyone know if Dragon would work with Discord?

    1
  • noahg2

    I dont know but i did attempt to write an extremely basic version in python

    import pyaudio
    from pocketsphinx import Pocketsphinx, Decoder
    from os import environ, path

    decoder = Pocketsphinx(hmm="model/en-us", lm="model/en-us.lm.bin", dict="model/cmudict-en-us.dict")
    p = pyaudio.PyAudio()
    stream = p.open(format=pyaudio.paInt16, channels=1, rate=16000, input=True, frames_per_buffer=512, input_device_index=1,
    as_loopback=True)
    stream.start_stream()

    in_speech_bf = False
    decoder.start_utt()
    while True:
    buf = stream.read(512)
    decoder.process_raw(buf, False, False)

    if decoder.get_in_speech() != in_speech_bf:
    in_speech_bf = decoder.get_in_speech()
    if not in_speech_bf:
    decoder.end_utt()
    print('Result:', decoder.hyp().hypstr)
    decoder.start_utt()
    0

Du måste logga in om du vill lämna en kommentar.