Multi-voice text to speech synthesis and dialogues
16-05-2023 ,
06-01-2025
Synthesizing speech with different voices. You can do multi-voice voiceovers in different languages. It's very simple to voice with multiple voices in one track.
Here's how to do it:
- Add a new voice by clicking on this icon.
- Choose the language and voice you want.
- Highlight the sentence you want to voice with this new voice.
- Click the Wrap button .
- That's it! Now that text will be voiced with two voices. Take a listen and see how it turned out.
Example:
You can insert any number of voices and languages.
Dialogue Constructor
If you have a complex dialogue or need to generate a dialogue via API, you can create the dialogue code using a Google Document.
-
Copy the Template:
Follow the link to the provided document and save a copy to your Google Drive. Go to the "File" menu and select "Make a copy."
-
Work on the "Dialogue Constructor" Tab:
- In Column A (Voice), input the name of the voice, select the speed, and tone.
- In Column B (Your Text), enter the dialogue phrases.
- Column C (Code for Speechgen) will automatically generate the Speechgen code for each dialogue line.
-
Voice Name Requirements:
Ensure the voice name in Column A matches exactly as listed on Speechgen.
- To find the correct voice name, check the "all_voices" tab in the document, where voices are sorted by country. Copy the desired name from there.
-
Example:
- In Row 2, I entered the voice name Derek EN in Column A. Since I prefer the default speed and tone, I left those unchanged. In Column B, I added the first phrase of the dialogue.
- In Row 3, I entered the voice name Serena EN and added her corresponding dialogue line.
By repeating this process, I constructed the entire dialogue. In Column C, Speechgen automatically wrapped each phrase in the correct <dialog> tag.
-
Generate the Dialogue:
- Highlight all rows in Column C and copy them using Ctrl+C.
- Go to Speechgen and paste the content into the input field. It will look like this:
-
Final Steps:
- Click Generate speech to produce the audio.
- The entire dialogue will be generated in a single file, which you can download.
For a working example, check this link.