ChatGPT
ChatGPT copied to clipboard
[Bug]v0.12.0中的SetSpeechLanguage显示NoData
Version
0.12.0
Bug description
如图,在最新版本中的Set Speech Language一栏中显示NoData
OS
Windows11
Environment
No response
Have the same problem here
Same issue here too
Me 2
Same issue
+1
It will be fixed in the next version, someone has contributed PR. https://github.com/lencx/ChatGPT/pull/572
The current text-to-speech is really strange. Can Microsoft or Google's speech synthesis API be added in subsequent versions to improve the user experience?
The current text-to-speech is really strange. Can Microsoft or Google's speech synthesis API be added in subsequent versions to improve the user experience?
APIs are never free, do you buy APIs?
Hi, until the new version with the bugfix is released, you can use this workaround:
- In the "Preferences" menu click on "Go to config - Ctrl+Shift+G",
- This will open the file explorer of your OS. Open the .chatgpt folder, and then open the chat.conf.json file.
- Edit the chat.conf.jon file by setting { "speech_lang" : "" } using one of the following values:
Microsoft Hortense - French (France) Microsoft Mark - English (United States) Microsoft Zira - English (United States) Microsoft David - English (United States) Microsoft Julie - French (France) Microsoft Paul - French (France)
- Restart the app, the voice you selected will then just work on the chat windows, on click on "Say" button...
(This workaround does not solve the list issue, but it just allows you to select the voice : it removes the need for the voices list select)
As an improvement to this feature, I suggest allowing the user to change the voice in the chat window for each message (overriding the default voice configured globally).
I usually use different languages when using Chat GPT, and it would be great if I could change the synthesis voice...
It would be even better if the language of the message was automatically detected and a voice corresponding to that language was automatically selected for the speech synthesis.
Microsoft David - English (United States)
Although it does show a selected language, it still don't have any data to get the language from, so this didn't work.
Microsoft David - English (United States)
Although it does show a selected language, it still don't have any data to get the language from, so this didn't work.
my workaround does not fix the issue of "No data" on the language list... But it actually changes the voice without the need of this list... so, if you corretly followed the steps above, the voice will just work in the chat windows, when you click on the "say" button.
Microsoft David - English (United States)
Although it does show a selected language, it still don't have any data to get the language from, so this didn't work.
my workaround does not fix the issue of "No data" on the language list... But it actually changes the voice without the need of this list... so, if you corretly followed the steps above, the voice will just work in the chat windows, when you click on the "say" button.
but still, I don't get any sound even if Iv'e edited chat.conf.json
. Can the issue be that i'm not on windows?
but still, I don't get any sound even if Iv'e edited chat.conf.json. Can the issue be that i'm not on windows?
Same question. I'm on Debian and cannot seem to get the "Say" function to work.
Is it possible to add a second text box where I can define a program using placeholders for text-to-speech MP3 file generation? I would love to see this feature as it would resolve all issues with Linux and provide me with the flexibility to use any text-to-speech engine of my choice.
So something like:
/home/user/tools/aws-polly.sh --message "{message}" --lang"de-DE" --output "{output}"
Another example following @scolastico's, using gTTS:
gtts-cli -f ./message.txt --output ./message.mp3
The current text-to-speech is really strange. Can Microsoft or Google's speech synthesis API be added in subsequent versions to improve the user experience?
APIs are never free, do you buy APIs?
If I remember correctly, Mozilla has an open and free text to speech project that is (as I understand) able to run locally on a machine with low resource requirements. I am new to this project so if this is already in use here, I apologize. Nonetheless, here is a Github Link to the project. Thanks to you and all the contributors for enhancing ChatGPT and bringing it to the Desktop (especially Linux). https://github.com/mozilla/TTS
It will be fixed in the next version, someone has contributed PR. #572
When will be the next version?
It will be fixed in the next version, someone has contributed PR. #572
Hello, I did what @melbarra88 suggested, but it doesn't work. And after I clicked that video button, the terminal told me:
It looks like this module isn't loaded correctly.
My system is Debian testing, and I'm using AppImage.
It will be fixed in the next version, someone has contributed PR. #572
Hello, I did what @melbarra88 suggested, but it doesn't work. And after I clicked that video button, the terminal told me:
It looks like this module isn't loaded correctly. My system is Debian testing, and I'm using AppImage.
I wrote a inject script:
if ('speechSynthesis' in window) {
// `speechSynthesis` is supported
const synth = window.speechSynthesis;
var voice = synth.getVoices();
consle.log(voice);
} else {
// `speechSynthesis` is not supported
console.log('Speech synthesis not supported');
}
And the console result is: