Voice profiles having multiple samples generate audio using a random voice instead

[Flameborn]

Using VoiceBox version 0.1.11, if I create a voice profile that has multiple audio samples added to it, the generated audio uses a random voice instead. If the profile has a single sample, generation works as expected.

This is under MacOS 26.2, but it might happen under Windows as well.

[jamiepine]

Oh that explains a lot, thanks for that information. Some people were only telling me that the voice generation doesn't respect voices, but all my voices use a single sample. Looking into this now!

[FrancescoSaverioZuppichini]

@jamiepine using a single example as well

[Flameborn]

Thank you very much for the quick fix.