Eleven v3 • Hacker News Board

Re: Eleven v3

#51

Post by SoftTalker » Thu, Jun 05, 2025, 8:41 PM UTC

ricketycricket wrote: post #28

From the example: "Oh no, I'm really sorry to hear you're having trouble with your new device. That sounds frustrating." Being patronized by a machine when you just want help is going to feel absolutely terrible. Not looking forward to this future.

Yeah it's irritating enough when humans do it, it's so transparently insincere. Just help me with my problem.

I guess I am just old now but I hate talking to computers, I never use Siri or any other voice interfaces, and I don't want computers talking to me as if they are human. Maybe if it were like Star Trek and the computer just said "Working..." and then gave me the answer it would be tolerable. Just please cut out all the conversation.

Re: Eleven v3

#52

Post by lharries » Thu, Jun 05, 2025, 8:41 PM UTC

jurgenaut23 wrote: post #34

French is atrocious. It sounds like beginner-level english speakers trying to decipher a text without understanding it.

Can you try with this voice? https://elevenlabs.io/app/voice-library?voiceId=xTZlmU8dKXdy...

Voice selection matters more for this model

Re: Eleven v3

#53

Post by lharries » Thu, Jun 05, 2025, 8:47 PM UTC

gozzoo wrote: post #35

GrayShade wrote: post #21

Earlier quoted context omitted.

I'm not sure what you mean. I chose Romanian from the language selector and tried Matilda, Alice and Laura. Laura actually sounds like an English TTS trying to pronounce Romanian.

Exactly the same thing with Bulgarian voices.

It should work well for the logged in voices here: https://elevenlabs.io/app/voice-library?language=ro

We are in the process of updated the homepage voices for the new languages

Re: Eleven v3

#54

Post by carlosjobim » Thu, Jun 05, 2025, 8:47 PM UTC

lharries wrote: post #43

carlosjobim wrote: post #26

Their non-English (automated?) localization of the front page is ridiculously badly translated.

Which language isn't good and I'll get that fixed asap?

You need native or at least fluent speakers to help you, to get the expressions right. For example Swedish is written like a word-for-word translation from English.

Re: Eleven v3

#55

Post by jsemrau » Thu, Jun 05, 2025, 8:51 PM UTC

taf2 wrote: post #37

echelon wrote: post #29

Earlier quoted context omitted.

They're also still too expensive, and that's creating a lot of opportunity for other players. Even though ElevenLabs remains the quality leader, the others aren't that far behind. There are even a bunch of good TTS models being released as fully open source, especially by cutting-edge Chinese labs and companies. Perhaps in a bid to cut off the legs of American AI companies or to commoditize their compliment. Whatever…

What would say are some of the best open source TTS - chatterbox maybe?

I had good results with Nemo + xTTS_v2

https://docs.nvidia.com/nemo-framework/user-guide/latest/nem...

https://huggingface.co/coqui/XTTS-v2

Re: Eleven v3

#56

Post by zamadatix » Thu, Jun 05, 2025, 8:53 PM UTC

lharries wrote: post #39

zamadatix wrote: post #16

The (American English) voices are absolutely amazing but the tags for laughs still feel more like an "inserted dedicated laugh section" than a "laugh at this point in speaking" type thing. I.e. it can't seem to reliably know when to giggle while saying a word, "just" giggle leading up to a word.

If you edit the text so that laugh makes sense in the context it should be much more natural like this one: https://x.com/elevenlabsio/status/1930689782331412811

The first laugh in that " Hey, Dr. Von Fusion" is a dedicated laugh section, which the model does extremely well, but it works because that's a natural place to laugh before actually speaking the following words. Skip ahead to "...robot chuckle. Jessica: I know right!" and you get an awkwardly time/toned light chuckle completely separated from the "I know" you'd naturally continue saying while making that chuckle.

You can always rewrite the text to avoid times where one would naturally laugh through the next couple of following words but that's just attempting to avoid the problem and do a different kind of laugh instead.

Re: Eleven v3

#57

Post by brian_herman » Thu, Jun 05, 2025, 9:04 PM UTC

Unfortunately voice actors will be replaced by someThing like this hopefully they will find someThing else To do

Re: Eleven v3

#58

Post by lharries » Thu, Jun 05, 2025, 9:08 PM UTC

politelemon wrote: post #50

louisjoejordan wrote: post #46

quick note that that voice selection matters a lot with our new v3 model, especially voice language! We have a curated list of v3 voices in the library, but feel free to try others to find what works. Make sure language voice language match.

Unfortunately many of the foreign language generation sounds unnatural, with a strong American accent. I've tried the Spanish, Galician, Tagalog, German. I did try the curated samples.

Can you choose a voice that's native in that language in the voice library: https://elevenlabs.io/app/voice-library?language=es

Re: Eleven v3

#59

Post by maxglute » Thu, Jun 05, 2025, 9:09 PM UTC

What's the state of open source tts? I'm a heavy TTS user, anything that can run at 3x-4x speed off enthusiast hardware?

Re: Eleven v3

#60

Post by christophilus » Thu, Jun 05, 2025, 9:11 PM UTC

We’re using elevenlabs in a new prototype, and it gets confused by its own voice which my mic picks up. Unless I wear headphones, it thinks I’m talking, and it gets into a loop.

I hope this release fixes that bug!