From b7de2988b068659d2bc608dbbdc57fcd1c4c79ae Mon Sep 17 00:00:00 2001 From: erew123 <35898566+erew123@users.noreply.github.com> Date: Wed, 13 Dec 2023 05:21:38 +0000 Subject: [PATCH] Updated note about clip length Turns out the models default config file is setting "max_ref_len": 30, allowing for up to 30 second long voice clips to be used. --- tts_server.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tts_server.py b/tts_server.py index de43cb9..709a289 100644 --- a/tts_server.py +++ b/tts_server.py @@ -750,7 +750,7 @@ simple_webpage = """

This extension will download the 2.0.2 model to /extensions/alltalk_tts/models/

This TTS engine will also download the latest available model and store it wherever it normally stores it for your OS (Windows/Linux/Mac).

How do I create a new voice sample?

-

To create a new voice sample you need to make a wav file that is 22050Hz, Mono, 16 bit and between 6 to 12 seconds long, though 7 to 9 seconds is usually good.

+

To create a new voice sample you need to make a wav file that is 22050Hz, Mono, 16 bit and between 6 to 30 seconds long, though 8 to 10 seconds is usually good enough. The model can handle up to a 30 second sample, however Ive not noticed any improvemnt in voice output from a much shorter clip.

You want to find a nice clear selection of audio, so lets say you wanted to clone your favourite celebrity. You may go looking for an interview where they are talking. Pay close attention to the audio you are listening to and trying to sample, are there noises in the backgroud, hiss on the soundtrack, a low humm, some quiet music playing or something? The better quality the audio the better the result. Dont forget, the AI that processes the sounds can hear everything, all those liitle noises, and it will use them in the voice its trying to recreate.

Try make your clip one of nice flowing speech, like the included example files. No big pauses, gaps or other sounds. Preferably one that the person you are trying to copy will show a little vocal range and emotion in their voice. Also, try to avoid a clip starting or ending with breathy sounds (breathing in/out etc).

Generating the sample!