#I am working on a project to which I uploaded a pronunciation dictionary. The dictionary did not se
1 messages · Page 1 of 1 (latest)
To manage pronunciation dictionaries in your project, you can follow these steps:
-
Access Project Settings: Go to the project where you've uploaded the pronunciation dictionaries.
-
Locate the Dictionaries: In the project settings, find the section where the pronunciation dictionaries are listed.
-
Remove the Dictionaries: You should have the option to delete or remove the dictionaries from this section. Look for a delete or remove button next to each dictionary entry.
-
Re-upload if Necessary: Once you've removed the unwanted dictionaries, you can re-upload the correct one if needed.
If the pronunciation issues persist, ensure that your .PLS file is correctly formatted and that the entries are case-sensitive. You might also want to check if the phonemes or aliases are correctly specified for the words you are targeting.
That is not working. There is a circle with three dots next to the file name, but it doesn't do anything. I can not delete the files or anything else.
It seems like you're experiencing an issue with removing pronunciation dictionary files. Unfortunately, the provided content does not include specific instructions for deleting files directly from the user interface. I recommend checking the Help Center for more detailed guidance or updates on this functionality. If the issue persists, you might want to reach out to support for further assistance.
Hello. I am working on a new project that involves both spanish and english. Do I need to tell the Multilingual V2 model when I switch between spanish and english, or will it adjust automatically?
The Multilingual V2 model will automatically adjust based on the text you provide. You don't need to explicitly tell it when you switch between Spanish and English. Just ensure that the text is clearly written in the language you want it to be spoken, and the model will handle the rest. If you encounter any issues, feel free to reach out for further assistance.
I am confused about the best way to structure my project. I have a collection of 12 voiceovers, some very long and some shorter. I want to keep them together in one project. Does that mean I should divide my content in 12 chapters, one for each voiceover?
Yes, dividing your content into 12 chapters, one for each voiceover, is a good approach. This will help you manage your project more effectively, allowing you to organize and generate voiceovers for each section individually. Additionally, it provides flexibility to adjust settings or regenerate specific chapters without affecting the entire project.
Within a single chapter, what does a line break mean to the model? Will it be rendered as a longer pause than a period might produce? Or does it result in multiple audio files at generation?
In a single chapter, a line break does not inherently create a longer pause than a period. It also does not result in multiple audio files upon generation. The model treats line breaks as a way to organize text, but they do not affect the audio output in terms of pauses or file separation. If you need specific pauses, consider using the <break time="x.xs" /> syntax for precise control over pause duration.
can I put the <break time...> syntax on its own line between lines of spoken text?
Yes, you can place the <break time="x s" /> syntax on its own line between lines of spoken text. This will create a natural pause in the speech, and the AI will recognize and implement the break as specified. Just ensure that the break time is within the allowed range, up to 3 seconds, to maintain stability in the audio output.
When I divide the text into idependent lines with line breaks, the generation step generate the entire set, as if it were all in one big paragraph? Or do I need to select all of the lines to do that?
When you divide text into independent lines with line breaks, the voice-over system interprets each line as a separate sentence. To ensure the entire set is generated as one continuous paragraph, avoid placing line breaks between words. If you want to generate the entire set together, you should select all the lines before generating the audio.
But, when I generate it as an audio file, will I get one file, or many?
When you generate audio for a chapter in your project, it will produce a single audio file for the entire chapter. If you have divided your text into multiple lines or paragraphs within that chapter, they will be combined into one audio file during the generation process.
But I have to generate it to hear is spoke as a continuous audio segment? It would appear there is no way for me to preview it as continuous audio prior to generating it?
Currently, there is no option to preview text-to-speech outputs without using credits. Each audio generation will deduct credits, as the servers need to generate the audio. To manage credits effectively, you can use the "Projects" feature to regenerate specific sections instead of the entire text. This can help minimize credit usage.
If you want to hear the text as a continuous audio segment, you will need to generate it first. Unfortunately, there is no way to preview it as continuous audio before generation.
is there a way to regenerate audio without playback?
Currently, there is no way to preview audio without generating it, as each generation will deduct credits. However, you can manage credit usage by utilizing the "Projects" feature, which allows you to regenerate only specific sections of your script instead of the entire text.
Additionally, you are allowed two free regenerations under specific conditions: the prompt, voice, and model must remain the same, the first generation must have been made within the last two hours, and the page must not have been refreshed since the original audio generation. These free regenerations are only available in Speech Synthesis via the website.
I have identical voiceovers in English and Spanish, and in both cases I have used deliberate and equivalent spacing using the <break time="x.xs" /> syntax. Objectively (determined using a wave editor) the same syntax results in longer pauses in Spanish than it does in English. Is this by design, or something that I need to compensate for in some way? I am using the multilingual v2 model.