#speechd — Public Fediverse posts

I'm using Text-To-Speech (TTS) on my Desktop (Calender Events, Firefox Read Aloud) and with Home Assistant. However the quality isn't the best and I wanted a better option.

I've started a new project for this called crane-wyoming which is in alpha/beta stage as it is not feature complete but works well for TTS.

https://codeberg.org/cryptomilk/crane-wyoming

#homeassistant #speech #speechd #speechdispatcher #tts #inference #llm #voice #asr #qwen #voxtral #ai #rust

I'm using Text-To-Speech (TTS) on my Desktop (Calender Events, Firefox Read Aloud) and with Home Assistant. However the quality isn't the best and I wanted a better option.

I've started a new project for this called crane-wyoming which is in alpha/beta stage as it is not feature complete but works well for TTS.

https://codeberg.org/cryptomilk/crane-wyoming

#homeassistant #speech #speechd #speechdispatcher #tts #inference #llm #voice #asr #qwen #voxtral #ai #rust

I'm using Text-To-Speech (TTS) on my Desktop (Calender Events, Firefox Read Aloud) and with Home Assistant. However the quality isn't the best and I wanted a better option.

I've started a new project for this called crane-wyoming which is in alpha/beta stage as it is not feature complete but works well for TTS.

https://codeberg.org/cryptomilk/crane-wyoming

#homeassistant #speech #speechd #speechdispatcher #tts #inference #llm #voice #asr #qwen #voxtral #ai #rust

I'm using Text-To-Speech (TTS) on my Desktop (Calender Events, Firefox Read Aloud) and with Home Assistant. However the quality isn't the best and I wanted a better option.

I've started a new project for this called crane-wyoming which is in alpha/beta stage as it is not feature complete but works well for TTS.

https://codeberg.org/cryptomilk/crane-wyoming

#homeassistant #speech #speechd #speechdispatcher #tts #inference #llm #voice #asr #qwen #voxtral #ai #rust

I'm using Text-To-Speech (TTS) on my Desktop (Calender Events, Firefox Read Aloud) and with Home Assistant. However the quality isn't the best and I wanted a better option.

I've started a new project for this called crane-wyoming which is in alpha/beta stage as it is not feature complete but works well for TTS.

https://codeberg.org/cryptomilk/crane-wyoming

#homeassistant #speech #speechd #speechdispatcher #tts #inference #llm #voice #asr #qwen #voxtral #ai #rust

Now I need to go figure out why speechd-el isn’t interacting with Evil properly. It reads lines rather strangely, and moving by word or character is quite strange in that it reads the whole line as you move by characters. I wonder if there is advice somewhere I need to disable. #emacs #speechd-el

This was done, at least to a certain extent, in #Emacs with #Emacspeak and #speechd-el, starting way back in the 90s. And Emacs is, with a few caveats, a very nice editor to use. But I don't want every app to have its own, custom assistive technology built in. That's a huge duplication of effort, and they for certain won't all have feature parity. Plus, two screen readers talking at the same time is highly irritating. 😀 We need a standard way for applications to expose more detailed semantic, application specific info to ATs.

This was done, at least to a certain extent, in #Emacs with #Emacspeak and #speechd-el, starting way back in the 90s. And Emacs is, with a few caveats, a very nice editor to use. But I don't want every app to have its own, custom assistive technology built in. That's a huge duplication of effort, and they for certain won't all have feature parity. Plus, two screen readers talking at the same time is highly irritating. 😀 We need a standard way for applications to expose more detailed semantic, application specific info to ATs.

This was done, at least to a certain extent, in #Emacs with #Emacspeak and #speechd-el, starting way back in the 90s. And Emacs is, with a few caveats, a very nice editor to use. But I don't want every app to have its own, custom assistive technology built in. That's a huge duplication of effort, and they for certain won't all have feature parity. Plus, two screen readers talking at the same time is highly irritating. 😀 We need a standard way for applications to expose more detailed semantic, application specific info to ATs.

I'm coming to a terrifying conclusion...

The only time I've ever been satisfied with a text editor was VS Code on Windows. On Linux, it's too laggy (probably #Orca's fault). #Vim is great, but not very #accessible. #Emacs is also fantastic and very accessible, but #emacspeak uses the old, unmaintained espeak synth instead of espeak-ng, I haven't played much with #speechd-el yet, and Emacs takes ages to set up, plus I dislike Elisp and don't really have a clue how to start learning it. 1/2

I'm coming to a terrifying conclusion...

The only time I've ever been satisfied with a text editor was VS Code on Windows. On Linux, it's too laggy (probably #Orca's fault). #Vim is great, but not very #accessible. #Emacs is also fantastic and very accessible, but #emacspeak uses the old, unmaintained espeak synth instead of espeak-ng, I haven't played much with #speechd-el yet, and Emacs takes ages to set up, plus I dislike Elisp and don't really have a clue how to start learning it. 1/2