[HN Gopher] Creates hyper-realistic voice clones from just 3 sec...
___________________________________________________________________
Creates hyper-realistic voice clones from just 3 seconds of audio
Author : blacktechnology
Score : 21 points
Date : 2025-01-10 18:16 UTC (4 hours ago)
(HTM) web link (anyvoice.net)
(TXT) w3m dump (anyvoice.net)
| ge96 wrote:
| 3 seconds? That's crazy
|
| "Huuhhhhhhhhhhh"
|
| I wonder what their "fox jump" sentence is
| sailfast wrote:
| Default for me was: "What a beautiful day it is today, with
| bright sunshine and gentle breeze. Let's talk about the future
| of artificial intelligence."
|
| That said, I'm not going to be submitting a sample because
| [reasons]
| mk_stjames wrote:
| A "Panphonic Poem" is what may do well here. As in...
| The pleasure of Shawn's company Is what I most enjoy.
| He put a tack on Ms. Yancey's chair When she called him a
| horrible boy. At the end of the month he was flinging two
| kittens Across the width of the room. I count on
| his schemes to show me a way now Of getting away from my
| gloom.
|
| As discussed here:
|
| https://literalminded.wordpress.com/2006/05/05/a-panphonic-p...
|
| And recited very famously, in part and slightly modified, here:
|
| https://www.youtube.com/watch?v=CgX4uJSj00Y
| bugglebeetle wrote:
| Sure, just let me submit my voice for cloning to a closed
| sourced, online service of unknown provenance. What could ever go
| wrong?
| dvh wrote:
| That's why you submit politician's voice instead
| HanClinto wrote:
| Yeah, but they have you read a specific text, so not as much
| of an option if you use the primary demo.
|
| Seems like a heck of a nice way to gather a training set! :)
| unsnap_biceps wrote:
| The "upload audio" feature doesn't require any specific
| text.
| lubujackson wrote:
| Cue reference to "Sneakers"...
| superkuh wrote:
| I submitted an 8 second clip of speech and the resulting
| synthesized speech did not sound like the same voice. Too bad.
| infogulch wrote:
| I hope you have a nice voice, I'll be listening to it try to
| sell me an extended car warranty for the next 3 months.
| xnx wrote:
| What model is this using? I've had good results with e2-ft-tts
| running locally via Pinokio. You can also run it online for free
| https://huggingface.co/spaces/mrfakename/E2-F5-TTS
| krainboltgreene wrote:
| Getting a 500 from the HTTP API and also there's an `debugger` in
| the javascript.
| mxuribe wrote:
| Immediately, i thought that cybersecurity is now ruined for the
| distant future. Imagine if you will, a starship captain ready
| with a plot to overcome the evil plaguing their crew...and all
| they need to do is over-ride the starship computer's safety
| controls with the captain';s own voice override
| authorization...but, alas, early in 2025 a tech company developed
| the means by which said evil entity could re-override the
| captain's voice auth....and block the captain's plan...thereby
| dooming the entire crew of the starship.
|
| This is why we can not have nice things; not now nor in the far
| off future! All of our uniqueness will be more easily duplicated.
| Thankfully, i won';t upload any of my voice recordings, and i
| will continue to walk around in my faraday cage suit. /s
| montag wrote:
| Yep, this is a real Star Trek TNG episode, S4 E3 "Brothers"
| croemer wrote:
| Getting error: Failed to generate voice
| HeatrayEnjoyer wrote:
| I am hitting this error as well. I was additionally unable to
| create an account. Seems beta?
| xqcgrek2 wrote:
| Has anyone tried multiple iterations? That is, upload a real
| voice, get its synthesized version, upload synthesized version 1
| to get synthesized version 2, rinse and repeat...
| abeppu wrote:
| Perhaps Alvin Lucifer reading his "I am sitting in a room" text
| would be ideal.
| clueless wrote:
| anybody try this and have a good result?
| gamblor956 wrote:
| This was a great way for them to collect a lot of free voice data
| to train their model.
| inerte wrote:
| Every time there's a voice recognition post here someone
| comments about acquiring data. Why is this method better than
| having access to all of the video and podcasts sites on the
| internet?
___________________________________________________________________
(page generated 2025-01-10 23:00 UTC)