←back to thread

Show HN: Dia, an open-weights TTS model for generating realistic dialogue

(github.com)
652 points toebee | 10 comments | 21 Apr 25 17:07 UTC | HN request time: 0.793s | source | bottom
Show context
tyrauber ◴[21 Apr 25 18:34 UTC] No.43755003[source]▶
Hey, do yourself a favor and listen to the fun example:

> [S1] Oh fire! Oh my goodness! What's the procedure? What to we do people? The smoke could be coming through an air duct!

Seriously impressive. Wish I could direct link the audio.

Kudos to the Dia team.

replies(5): >>43755963 #>>43756347 #>>43756931 #>>43757403 #>>43763123 #
1. jinay ◴[21 Apr 25 20:13 UTC] No.43755963[source]▶
For anyone who wants to listen, it's on this page: https://yummy-fir-7a4.notion.site/dia
replies(4): >>43756236 #>>43756357 #>>43761303 #>>43763439 #
2. mrandish ◴[21 Apr 25 20:41 UTC] No.43756236[source]▶
Wow. Thanks for posting the direct link to examples. Those sound incredibly good and would be impressive for a frontier lab. For two people over a few months, it's spectacular.
3. DoctorOW ◴[21 Apr 25 20:54 UTC] No.43756357[source]▶
A little overacted, it reminds me of the voice acting in those flash cartoons you'd see in the early days of YouTube. That's not to say it isn't good work, it still sounds remarkably human. Just silly humans :)
replies(3): >>43757460 #>>43760252 #>>43760682 #
4. selimthegrim ◴[21 Apr 25 23:03 UTC] No.43757460[source]▶
Reminded me of the Fenslerfilm G.I. Joe sketch where the kids have something on the stove burning
replies(1): >>43757575 #
5. wisemang ◴[21 Apr 25 23:23 UTC] No.43757575{3}[source]▶
Stop all the downloading!
6. Cthulhu_ ◴[22 Apr 25 09:00 UTC] No.43760252[source]▶
"flash cartoons in the early days of Youtube" Wouldn't those be straight from Newgrounds?
replies(1): >>43777892 #
7. 3by7 ◴[22 Apr 25 10:36 UTC] No.43760682[source]▶
Overacted and silly humans indeed: https://www.youtube.com/watch?v=gO8N3L_aERg
8. dostick ◴[22 Apr 25 12:24 UTC] No.43761303[source]▶
This is an instant classic. Sesame comparison examples all sound like clueless rich people from The White Lotus.
9. intalentive ◴[22 Apr 25 15:37 UTC] No.43763439[source]▶
Sounds great. One of the female examples has convincing uptalk. There must be a way to manipulate the latent space to control uptalk, vocal fry, smoker’s voice, lispiness, etc.
10. DoctorOW ◴[23 Apr 25 23:44 UTC] No.43777892{3}[source]▶
Thank you! I couldn't remember the name Newgrounds for some reason!!