#speech on #CH32V003 #microcontrollers #electronics #hardware using #Talkie
on 02025-11-05"Talkie" is an #LPC #speech library for #Arduino #electronics #hardware
on 02025-11-05discussion of #KittenTTS #text-to-speech #speech #neural-networks
on 02025-08-06in-browser version of #KittenTTS free-sofware #text-to-speech #speech
on 02025-08-06another in-browser version of #SAM proprietary #text-to-speech #speech software for the Commodore 64.
on 02025-08-06"SAM", Software Automatic Mouth, compiled with Emscripten. proprietary #text-to-speech #speech software for the Commodore 64.
on 02025-08-06This is the source of the proprietary #DECTalk #text-to-speech #speech software at webspeak.terminal.ink.
on 02025-08-06#DECTalk #text-to-speech #speech software packaged for the web. Proprietary. It doesn’t handle Unicode well, but it’s pretty decent as Stephen Hawking voices go.
on 02025-08-06is a new Apache2-licensed #text-to-speech #neural-networks model that’s only 25MB, but using espeak-ng via phonemizer. #speech
on 02025-08-06"LiveCaptions" #speech recognition application that apparently isn’t just a wrapper around Whisper but is supposedly a little more accurate (with English only) for live audio. Instead it uses something on the CPU called #aprilasr by Fangjun Kuang. #neural-networks
on 02025-05-28how to do text to #speech with Mozilla’s "Common Voice" system in JS. wss = window.speechSynthesis; s = new SpeechSynthesisUtterance('My lovely lady lumps'); s.voice = wss.getVoices()[100]; wss.speak(s) works. To specify specifically English voices, wss = window.speechSynthesis; s = new SpeechSynthesisUtterance('My lovely lady lumps'); for (let v of wss.getVoices()) { if (v.lang.substr(0,2) !== 'en') continue; s.voice = v; wss.speak(s);}
#speech synthesis in #MicroPython using DECtalk (?) on a #Raspberry-Pi #RP2040 and a so-called SYN6988, or in #Emscripten. Or categorizing birds with BirdNET-Pi #neural-networks.
on 02023-07-13David Rowe’s #Codec-2 was extended in 02018 to provide usable #speech at 450 bits per second
on 02021-02-09David Rowe’s "LPCNet" decoder for #Codec-2 can transmit very realistic #speech at 1700 bits per second using #WaveNet
on 02021-02-09