WHEN YOU BRUSHED B, YOU SUDDENLY THOUGHT OF A QUESTION: WHY NOWAI DubbingAlways sounds mechanical? I want a more natural voice, I want to control emotions and tone, but I don't want to record it every time. Until I found IndexTTS2, developed by the B teamthisIt's a treasure tool. I'm surprised
How strong are you
In short, IndexTTS2 is oneI'M A MOVIE-CLASS I.AI DON'T KNOW. IT'S NOT AN ORDINARY TEXT-TO-SPEECH TOOL. IT'S BUILT BY THE B TEAMA groundbreaking zero sampleSpeech SynthesisModelI don't know. The point is..No limit at all!
What I tried to do was counterproductive:
- Zero samplesVoice cloning: Only a few seconds of sound will make a perfect copy of anyone's voice
- Emotional controlIf you want to be angry, if you want to be gentle, if you want to be gentle, you can even describe emotions in words
- Accurate Time Control: As long as you want, it's no longer a problem
This paper describes two sound tools that use the open source at B station: indexTTS2。
Usage:
Method 1: Cloud deployment without computer configuration
Recommended use of this method.
|
|
|
|---|---|
|
|
|
|
|
2 Cloud computers may sometimes need to be looted. Better be in the morning and at night |
Step 1: Lease of cloud computers
Open the URL:
https://www.xiangongyun.com/image/detail/7e84884a-b227-4f30-aead-c6624afd2edc
Click Deployment Now
1 Choosing the cheapest computer: $1.59 per hour。
2 Click " Confirm deployment". Deployment time, 30 seconds ~ 2 minutes。
3 CLICK "FOUNDING CLOUD OS" ON THE RIGHT WHEN RUNNING
Step 2: Run indexTTS 2
Double-click " Open application " . If not properly opened, two or three minutes will be required。
Here's the normal open page。
Close this page first
then click "open webui"
This interface shows that the software is installed successfully。
Step 3: Use
1 The sound of cloning is uploaded。
2 Enter text
3 Click Generate Voice
4 Click Download
Method 2: Local deployment
DOWNLOAD INSTALLATION PACKAGE: SIZE 27.5G。
|
|
|
|---|---|
|
|
|
|
|
|
Install package address:
Thunder grid: https://pan.xunlei.com/s/VO_QTuhJO-0l_WTQQG9k3rNA1?pwd=dbks#
quarque: https://pan.quark.cn/s/38e0cf78cd87
Step 1: Download the installation package
Unpressure after download
Step 2: Run indexTTS2
after depressurizing the package, find run.bat at the bottom, double-click it
This interface will appear。
It takes 2-5 minutes
When this address appears, the tool is installed successfully。
http://127.0.0.1:7860
Copy this address to the browser
1 Upload to clone voice
2 Enter text
3 Click Generate Voice
Downloads the results generated。
NOTE: MY COMPUTER IS 16G, FIVE SECONDS AUDIO, AND IT TOOK ABOUT FIVE MINUTES TO FINISH。
Therefore, if the computer is not properly configured, it is suggested that the first way to rent the computer should be used。