Closed NickDiamondsIsForever closed 1 year ago
Exactly what I noticed as well- makes the audio sounds significantly worse. Hoping for this to get fixed!
When I first started using it, this wasn't happening, so i was assuming it's just a glitch? Has it always happened for you?
Yes, for me it hasn't been any different way.
It seems the " Don't use commas/semicolons/etc. to break down replies into sentences" option might fix this!
rats, i tried that but it didn't work :( i'm gonna keep digging around, let me know if you figure it out on your end! would be nice if ElevenLabs was reachable with this
oh my god no it works!!
Thank you so much for help. Truly
Just wondering what settings do you use for Stability and Similarity? Try different settings such as 0 for both, or Stability=0.1 and Similarity=1 ?
I've found that it greatly impacts the result - I think that's just how it works at ElevenLabs.
hey. It was set to 0 for stability and 1 for similarity. i tried moving some things around, as you suggested and it seems to change things slightly, but nothing quite fixes it. If i type the text directly into ElevenLabs it's very consistent and smooth but every sentence does tend to get chopped up into different "samples". Not sure how to correct this...
I have the problem that, despite my settings in German, he suddenly chooses the English language mode to answer, even though it is set differently. But the history also says that he used the English model. but it should be multilingual.
I'm currently experiencing an issue with the extension. I've entered my API from Eleven Labs and selected a voice. After saving it, when I input a question via voice in ChatGPT, the corresponding answer is not being read aloud. However, when I deactivate Eleven Labs, the answer is read aloud in the default voice. Does anyone have any suggestions on what I can do to make this work with the Eleven Labs voice? Thank you.
In the second browser I get the following error message after api input "chat.openai.com says [3] ElevenLabs API error: TypeError: Failed to execute 'decode' on 'TextDecoder': The provided value is not of type '(ArrayBuffer or ArrayBufferView)'." Can anyone help me with this?
Hi @kamilo558 I will be making an update soon to address the ElevenLabs problem. Thanks for your patience!
hello, I have attempted to fix this in v2.7.0 as of this message, v2.7.0 is currently pending approval on Chrome/Edge stores, but you can install it manually to try. the manual installation method is described on the homepage.
Please let me know if v2.7.0 fixes this issue? For me everything works fine after hours of testing - both with browser text-to-speech and ElevenLabs
What I did to get it working:
The entire chrome extension seems to be down...Not making sounds, and appears to be a problem for everyone. Apparently there is an update that was released today but it's not showing on my chrome browser. Do you know how to fix?
Anyone know if the extension was removed from the chrome store and
if it will be returning?
The current version on the Chrome store was not working properly due to an update from OpenAI which disabled the ElevenLabs functionality. As a result I was getting review-bombed with 1 star reviews from frustrated users (understandably). I decided to take it down momentarily, until the newest version is approved.
Meanwhile those who still want to use the extension can install it manually (v2.7.1 as of today) installation instructions are on the project home page.
apologies, i didn't realize you actually wrote this extension! people are ingrates. thanks for making this!
v2.7.1 is now available on the Chrome store and fixes this. I am closing the Issue, feel free to reopen in case you are still having trouble (with v2.7.1)
Hi, When using the TalkToChat extension with ElevenLabs, I am having a serious issue with the audio playback. It's coming out clipped and choppy as different voices/samples. When I look at the "history" on ElevenLabs, i find that, instead of one chunk of text/sample, each word is being separated as its own "generated sample".
I'm wondering why this is happening and if you can help me resolve this issue. When I use the extension within ChatGPT, i'd love for it to play back as one smooth consistent sample, the way it does within ElevenLabs. thanks n