How to make deep fake lip sync | using Wav2Lip

Wav2Lip is an improved version of LipGAN. It enables to Deepfake an audio within a video .

Do you want to learn how to make deep fake lip sync using Python and Wav2lip ?
All this technique needs is a short video and an audio speech, and the speech determines what the person on the video speaks out. Moreover the lips will be moving according to you new speech audio

This tutorial can learn how to implement this Python library in a very simple process.

Make videos appear to say other things for fun creative uses. Using Wav2Lip, a pre-trained lip sync model and Python environment , you can be up an running with new lip synced creations in minutes. A very impressive Deepfake audio replacement.

~~~~~~~~~~~~~~~ recommended courses and books ~~~~~~~~~~~~~~~

I am using one of these desktop cameras in my projects :

~~~~~~~~~~~~~~~ CONNECT ~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~ SUPPORT ME 🙏~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#Wav2lip
#EranFeit
#Python
~~~~~~~~~~~~~~ Credits ~~~~~~~~~~~~~
Music by Vincent Rubinetti
Download the music on Bandcamp:
Stream the music on Spotify:

[thomasandrews]

Hey! thanks for the tutorial I got on with it well!

What is the command to run through the GPU? (I have a Nvidia Geforce GTX 1650 which works with CUDA).

Thanks in advance!

[superpanter]

nice tutorial man you helped me get the program on my PC and not in my google drive :)

[Tinny-Robot]

Must the audio and the video be of the same length. And if its not which one will be repeated?

[polonskyg]

I know this is a bit old, but were you able to do it realtime instead of creating a video and then playing it? I want to make an avatar you can interact with. Thank you!

[mohitjavali]

Hi Eran,

Can you please tell me how to use this in flask and vscode?

[dudusescariot]

Thanks for the video man,

I'm getting this error:

"PackagesNotFoundError: The following packages are not available from current channels:

- torchvision==1.1.0
- pytorch==0.3.0"

Any advice here?

[aryanmahajan]

will it take the same amount of time everytime or just for the first time

[seyed]

thanks a lot for your full explanation. ✌🏻

[amiralikhan]

Hi Eran .. Great video .. thanks a lot .. i have a quick question .. do we have to configure Google API Keys ?? And if yes then how do we do that?

[jasonwhite]

I don't understand, when i type in, cd Python-Cool-Stuff it say the system cannot find the path specified

[llewodcm]

12:10 look at their lips move... You need to be careful to target the video you're intending to change, otherwise, this will happen...

[maximgrabow]

How can I change settings to better quality and tracking

[ywueeee]

hey can you do a similar video but for wav2lipHQ?

[Damnn]

perfect but i'm getting this error:

TypeError: guvectorize() missing 1 required positional argument: 'signature'

[ywueeee]

which version do we need in order to run this on the GPU?

[karnatisaishashank]

How we do with video with without faces

[xsploit]

days 5 trying to get this to work locally. still no bueno. on google colab server yes

[lyricsworldwide]

i got this error while trying to run the last command



(python-wav2lip) C:\Users\user\Wav2Lip>python inference.py --checkpoint_path checkpoints/wav2lip_gan.pth --face media/testy.mp4 --audio media/test.mp3 --outfile results/testy.mp4
Traceback (most recent call last):
File "inference.py", line 3, in <module>
import scipy, cv2, os, sys, argparse, audio
File "C:\Users\user\Wav2Lip\audio.py", line 1, in <module>
import librosa
File "C:\Users\user\miniconda3\envs\python-wav2lip\lib\site-packages\librosa\__init__.py", line 13, in <module>
from . import core
File "C:\Users\user\miniconda3\envs\python-wav2lip\lib\site-packages\librosa\core\__init__.py", line 115, in <module>
from .audio import * # pylint: disable=wildcard-import
File "C:\Users\user\miniconda3\envs\python-wav2lip\lib\site-packages\librosa\core\audio.py", line 12, in <module>
import resampy
File "C:\Users\user\miniconda3\envs\python-wav2lip\lib\site-packages\resampy\__init__.py", line 7, in <module>
from .core import *
File "C:\Users\user\miniconda3\envs\python-wav2lip\lib\site-packages\resampy\core.py", line 11, in <module>
from .interpn import resample_f_s, resample_f_p
File "C:\Users\user\miniconda3\envs\python-wav2lip\lib\site-packages\resampy\interpn.py", line 75, in <module>
nopython=True,
TypeError: guvectorize() missing 1 required positional argument: 'signature'