In short
Usually you do not need to set anything: the default is "Auto", and Speeek finds how many different people talk in the video by itself. A number from 1 to 5 is a hint for recognition: if there are more people, Speeek finds all of them.
Which voices the speakers get #
- The first speaker gets the voice from the form. If you set several speakers, you can pick a voice for each of them.
- The other speakers Speeek finds get a clone of their own voice from the video if a clone is selected in the form. If a library voice is selected, the others get different library voices, so different people sound different.
- A clone needs at least 10 seconds of clean speech. With less, the speaker gets a library voice, see Why the voice does not sound like the original.
- On Start there is one voice for the whole video: all speakers are voiced with one library voice. Different voices and clones are on Pro and Business.
Any voice can be changed in the editor. To pick voices before voicing, choose "After reviewing the text" in the form.
If speakers are mixed up #
Speaker separation sometimes makes mistakes, especially when people talk at the same time, voices are similar or the recording is noisy.
- Open the translation in the editor.
- Each phrase has a speaker dropdown. Reassign phrases to the right speaker.
- Press "Re-voice". The phrases are voiced with the new speaker's voice.
