Self-hosted fully working ST:TNG voice computer interface

Submitted ⁨⁨2⁩ ⁨weeks⁩ ago⁩ by ⁨troed@fedia.io⁩ to ⁨startrek@startrek.website⁩

https://video.troed.se/w/5FqDa2jVqCcQWxetJEivnw

This has been my dream since I was a teen - and now the tech is real. This is a fork of an existing speech-to-speech project by Huggingface, with a lot of additions to add the right chimes, the voice* and the overall wakeword architecture.

I've had it running for a day and ... I like it. I'll most likely wire up a version of this throughout the house so the whole family can live the trekkie life :)

*) Cloning Barrett's voice ... I know. It's just little old me showing that we now have the tech, not some commercial venture. It's done with the utmost of respect for her craft throughout all those years!

source

Comments

Sort:hotnew top

troed@fedia.io ⁨2⁩ ⁨weeks⁩ ago
Oh thanks all for the nice comments! My fork is at https://github.com/troed/speech-to-speech but beware it's not kept in any other state than what I've last pushed :) Also you'll have to supply your own chime wavs and voice sample etc.

source
- troed@fedia.io ⁨2⁩ ⁨weeks⁩ ago
  I've added example scripts for client, server and MCP configs. It should be much easier to replicate with those now.
  
  source
theyllneverfindmehere@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
This is cool as heck. Can you share the details at some point?

source
homesweethomeMrL@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
It needs some punch-card-shuffling noises to bring the TOS vibes. Great work!

source
WhoIzDisIz@lemmy.today ⁨2⁩ ⁨weeks⁩ ago
I’d be tempted to have it switch voices for different purposes. Robert Picardo’s EMH for medicinal queries, Kate Mulgrew for “what should I do” queries, alternate between James Doohan & LeVar Burton for technical stuff, etc. Might throw in some Lower Decks characters for fun, too.

source
- sj_zero@lotide.fbxl.net ⁨2⁩ ⁨weeks⁩ ago
  "I'm thinking of investing in this crypto scam, should I do it?"
  
  Micheal Dorn: "you should not. Crypto scammers are without honor."
  
  source
  - HubertManne@piefed.social ⁨2⁩ ⁨weeks⁩ ago
    quarks voice. crypto can be very lucrative.
    
    source
- lemmylommy@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
  And Michael for those teary eyed, pathos laden, dull speeches.
  
  source
Einskjaldi@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
Majel was fine with her voice being used for computer so I don’t think there’s any problem there.

source
- Seimhe@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
  Assumption is not a form of consent
  
  source
  - IcedRaktajino@startrek.website ⁨2⁩ ⁨weeks⁩ ago
    
    the so-called “First Lady of Star Trek” recorded an entire library of phonetic sounds, allowing her voice to be used in future products outside of Star Trek and, quite possibly, as the computer voice in Star Trek: Discovery.
    
    source
  - fonix232@fedia.io ⁨2⁩ ⁨weeks⁩ ago
    I think this topic is actually ripe for debate.
    
    On one hand we know that Majel specifically recorded phonemes so her voice can be used for Star Trek purposes after her passing. While not a direct consent of "do whatever you want with my voice", the intent of similar purposes as OP's has been clearly present.
    
    There's also Paramount's refusal to use her voice in new Trek, aside from a few snippets here and there (mostly Picard for the throwback value). Which kinda goes against the spirit of Trek.
    
    On the other hand, the entire area of voice cloning is gray as fuck. It can very easily be used for nefarious purposes, yes - let that be corporate greed or fraud or propaganda. But if I make my own voice system, do I really hurt anyone by using their voice? I am not profiting off of it, I am not sharing the end result, it lives on my computer and that's it.
    
    But, let's say I want to do the same. There's this TV show from some 20 years ago called Eureka. In it is one of the earliest representations of a smart house with an AI governor, called SARAH. SARAH is voiced by one of the recurring cast members, in fact it's an ongoing joke that the house is voiced by this guy "talking like a girl", and that he's trying to get Sarah Michelle Gellar to do the proper, final voice...
    
    Now, the voice itself is pretty unique, but it isn't the voice of the actor, it's been heavily digitally edited. It's also a pretty unique voice but not something 99.9% of people would recognise. I make my clone from my own BR disk collection, with a set of automated and AI tools to get the right audio, clean it up, split by talker, and process all 77 episodes, manually selecting the right samples, manually tagging and labelling each; and finally, fine-tuning my own model for my own use. It will never leave my home, I won't sell or share it, and technically, with me owning the BR collection, this would fall under fair use.
    
    Which is still quite different than, say, a corporate entity trying to sell more product by using your voice.
    
    source
  - Blue_Morpho@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
    It’s not different than pirating an mp3.
    
    source
Olgratin_Magmatoe@startrek.website ⁨2⁩ ⁨weeks⁩ ago
That is cool as fuck. Nice job

source
Blue_Morpho@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
Fantastic! I built my own lcars control panel for a wall in my house using Html/JavaScript and always dreamed of finishing it with voice control.

source
- christopher@lemmy.ca ⁨2⁩ ⁨weeks⁩ ago
  I have a Home Assistant LCARS theme on my dashboard right now. Still learning how to make a nice looking dashboard with it.
  
  source
hopesdead@startrek.website ⁨2⁩ ⁨weeks⁩ ago
Circa 2009-2010 there was a desktop overlay theme for Windows XP that I think was called System47. It was super basic, a LCARS overlay that let you access basic apps and info. It included a voice command which you could open apps or perform certain functions.

source
Blue_Morpho@lemmy.world ⁨2⁩ ⁨weeks⁩ ago
Is there a guide on how to duplicate this? I saw a voice to LLM back to voice tutorial a few weeks ago but it was much slower and the developer talked about the fixes needed that you must have implemented!

source
charonn0@startrek.website ⁨2⁩ ⁨weeks⁩ ago
A keyboard? How quaint.

source
danielquinn@lemmy.ca ⁨2⁩ ⁨weeks⁩ ago
What do you use for voice synthesis?

source
- troed@fedia.io ⁨2⁩ ⁨weeks⁩ ago
  Qwen3-TTS does excellent cloning and is very fast. On my workstation (5060Ti) it renders the audio at around 2.5x realtime. I've added an example with my parameters now.
  
  source
danielquinn@lemmy.ca ⁨2⁩ ⁨weeks⁩ ago
OMG I love this. Thank you for sharing!

source
Kirk@startrek.website ⁨2⁩ ⁨weeks⁩ ago
Holy shit I love LLMs now

source
data1701d@startrek.website ⁨1⁩ ⁨week⁩ ago
A long time ago, I tried to tune a Piper voice using Majel Barrett’s speech from her reading of the novel “Q In Law”, but I didn’t get very far - just a small amount of training data collected.

I agree; for reasons other people have discussed in detail here, I think Majel Barrett is probably one of the least ethically ambiguous actors you can do this with.

source