Mind-Blowing Deepfake Tutorial: Turn Anyone into Your Favorite Movie Star! PC & Google Colab - roop

[FurkanGozukara]

Mind-Blowing Deepfake Tutorial: Turn Anyone into Your Favorite Movie Star! PC & Google Colab - roop

Full tutorial: https://www.youtube.com/watch?v=OI1LEN-SgLM

Newer better DeepFake tutorial : https://www.youtube.com/watch?v=RdWKOUlenaY

1 click Windows installer : https://www.patreon.com/posts/1-click-auto-88234834

1 click RunPod installer⤵️

https://www.patreon.com/posts/auto-installer-84511510

How to install C++ Compile Tools (fix C++ 14.0 or greater is required) ⤵️

https://youtu.be/-NjNy7afOQ0

Using open-source AI models and the newly released roop repository from GitHub, we can make deepfakes from a single source image. Furthermore, I have demonstrated how to use CodeFormer to significantly enhance the quality of generated video faces through the use of Automatic1111 Web UI and Davinci Resolve. I also provided a tutorial on fixing audio syncing issues. In addition, I have prepared an easy-to-use, free Google Colab that can run the roop deepfake repository and generate deepfake videos. This video is intended solely for educational and research purposes.

Source GitHub File ⤵️

https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Tutorials/1-Click-DeepFake-Tutorial.md

Our Discord server ⤵️

https://bit.ly/SECoursesDiscord

If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on 🥰 ⤵️

https://www.patreon.com/SECourses

Technology & Science: News, Tips, Tutorials, Tricks, Best Applications, Guides, Reviews ⤵️

https://www.youtube.com/playlist?list=PL_pbwdIyffsnkay6X91BWb9rrfLATUMr3

Playlist of StableDiffusion Tutorials, Automatic1111 and Google Colab Guides, DreamBooth, Textual Inversion / Embedding, LoRA, AI Upscaling, Pix2Pix, Img2Img ⤵️

https://www.youtube.com/playlist?list=PL_pbwdIyffsmclLl0O144nQRnezKlNdx3

00:00:00 Introduction to how to make amazingly easy excellent quality deepfake videos

00:01:28 Prerequisites installations

00:02:56 How to install ffmpeg and install it and set its environment path

00:03:36 How to install deepfake repo roop with a private venv

00:05:45 Start using the deepface app

00:09:10 How to use deepface on Google Colab for free

00:11:28 How to utilize ChatGPT to get Google Colab scripts / code

00:13:03 How to improve face quality with CodeFormer face restoration

00:17:12 How to reconstruct video from frames and sync the audio with Davinci Resolve free edition

🔥 Mind-Blowing Deepfake Tutorial: Turn Anyone into Your Favorite Movie Star! PC & Google Colab - roop

🎥 Learn the secrets of creating jaw-dropping deepfake videos in this step-by-step tutorial! No more relying on expensive services or online platforms. Everything you need to know is right here, on your computer, ensuring 100% privacy.

🖥️ Discover an incredibly user-friendly workflow that doesn't require a GPU or specialized training. All you need is your computer and the willingness to follow along. Plus, we've prepared a flawless Google Colab repository, allowing you to create deepfakes for FREE!

💯 Unlock the power of deepfake technology for educational and research purposes. Deepfake creation is now more accessible than ever, and this tutorial will get you prepared for the future. It's a game-changer for the video, movie, and animation industry!

📚 Access our comprehensive GitHub file, guiding you through every step of the process. Just check the video description for the link and start your deepfake journey today.

🚀 Before diving in, we'll walk you through the essential setup steps, including Python installation, git setup, Visual Studio redistributable package, and FFmpeg installation. Don't worry; we've got you covered with clear instructions and helpful links!

🔧 Once your setup is complete, we'll guide you through cloning the DeepFace repository and setting up a virtual environment. You'll learn how to install all the necessary requirements while troubleshooting any potential errors along the way.

💻 We'll demonstrate the deepfake generation process using a user-friendly interface. You'll discover how to swap faces, extract frames, and reconstruct the video. Witness the magic unfold before your eyes as we transform a regular video into a mind-blowing deepfake masterpiece!

🌟 But that's not all! We'll also show you how to enhance face quality using the cutting-edge Codeformer tool in the Automatic 1111 web UI. Plus, we'll delve into video reconstruction using the DaVinci Resolve Free Edition, taking your deepfakes to the next level.

🔍 Our step-by-step tutorial teaches you how to enhance face quality using the Codeformer of the Automatic 1111 web UI, reconstruct the video using DaVinci Resolve Free Edition, and generate deepfakes right on your PC or Google Colab. Remember, everything is set up to work on your personal computer to ensure privacy.

🎯 As we prioritize your security, we've prepared a Google Colab repository for generating deepfakes for free, all while maintaining data privacy. In this age of rapidly growing deepfake technology, it's crucial to understand the ins and outs of these technologies for research and educational purposes.

#Deepfake #DeepFace #GoogleColab #DaVinciResolve #Tutorial #Python

Video Transcription

• 00:00:00 Greetings everyone.

• 00:00:01 In this video, I am going to show you how to

• 00:00:04 generate a deepfake video with just one

• 00:00:07 click without relying on any paid or online services.

• 00:00:12 Everything demonstrated in this video works directly on your computer and remains 100%

• 00:00:17 private.

• 00:00:18 The workflow I am about to reveal is extremely user-friendly and doesn't require a GPU or

• 00:00:24 any specific training.

• 00:00:26 All you need is your computer and a willingness to follow my instructions.

• 00:00:31 Moreover, I have prepared a flawlessly working Google Colab repository and I will show you

• 00:00:37 how to make a deepfake on Google Colab for free as well.

• 00:00:42 Furthermore, I will show you how to enhance face

• 00:00:45 quality with CodeFormer of the Automatic1111

• 00:00:48 web UI and how to reconstruct video with DaVinci Resolve Free Edition.

• 00:00:53 This video is strictly intended for educational and research purposes.

• 00:00:59 Deepfake is becoming extremely easy to make, therefore everyone needs to get prepared.

• 00:01:06 Furthermore, this technology will certainly change

• 00:01:09 the video, movie and animation industry.

• 00:01:12 So I have prepared an excellent GitHub file that will guide you through for all of the

• 00:01:18 steps.

• 00:01:19 The link for this file will be in the description of the video, so make sure to open this file.

• 00:01:26 There are several things that you need to do before

• 00:01:29 start using Deepfake on your computer.

• 00:01:33 Alternatively, I also have prepared a Google Colab

• 00:01:37 and you can use this repository on this

• 00:01:40 Google Colab as well.

• 00:01:41 I will also show you how to use it in the next chapter of the video.

• 00:01:46 First we will begin with installing and using this amazing repository on our computer.

• 00:01:53 First of all, you need to install Python.

• 00:01:56 As a Python version, I am using 3.10.9.

• 00:02:00 If you don't know how to install Python and set

• 00:02:02 it as a default Python installation.

• 00:02:05 Watch

• 00:02:06 this amazing tutorial video.

• 00:02:07 So when you open a new command prompt like this

• 00:02:10 and type Python, you should get a Python

• 00:02:13 version like this.

• 00:02:16 I am using 3.10.9.

• 00:02:17 It may work with other Python versions as well, but this is working.

• 00:02:20 Then you also need to have a git.

• 00:02:23 The links are here as you are seeing.

• 00:02:25 When you open a cmd and type git, you should get git is available like this.

• 00:02:30 Also you need to install Visual Studio redistributable package.

• 00:02:33 This is the direct link and also Microsoft C++ build tools.

• 00:02:36 When you open this link, you will see download build tools here, download it and install

• 00:02:41 it.

• 00:02:42 Once you have installed all of this.

• 00:02:44 There's one more thing that you need to install, which is FFmpeg.

• 00:02:48 FFmpeg is widely used.

• 00:02:50 So once you set it up, it will help you tremendously in the future.

• 00:02:54 Other repositories that you

• 00:02:55 may use.

• 00:02:56 So go to this link, download the latest build.

• 00:03:00 For downloading

• 00:03:01 you can download this zip file, for example.

• 00:03:03 Extract it into any folder like this.

• 00:03:05 I have extracted it into this folder then, then

• 00:03:08 type environment and edit system environments

• 00:03:12 like this: open environment variables, in here

• 00:03:14 you will see path variable like this: click

• 00:03:17 edit and add that as a folder like you are seeing here.

• 00:03:22 Once you do this and click OK and OK, open a new

• 00:03:25 cmd, type FFmpeg and you should get results

• 00:03:29 like this.

• 00:03:30 This means that it is correctly installed and the path of it is set.

• 00:03:34 OK, after all of this, we will begin with cloning the DeepFace repository.

• 00:03:38 So the link is here.

• 00:03:40 Copy this command, make a new folder in any of the drive that you

• 00:03:44 want to install it, I will install it in my F drive, I will type DeepFace enter inside

• 00:03:50 that folder, you can also directly clone into the drive, open a new cmd, you see currently

• 00:03:56 this is where I am.

• 00:03:57 Right click and paste the link and it will clone into that folder.

• 00:04:01 Then

• 00:04:02 copy this command, paste it, you see the folder we are in is changed.

• 00:04:06 We will install with

• 00:04:07 making a new virtual environment so none of your other installations will be affected.

• 00:04:12 For example, Stable Diffusion, it will continue working as it is.

• 00:04:17 Execute the command, it

• 00:04:18 will make a new virtual environment with your default Python version, then we will move

• 00:04:23 into scripts folder, cd venv, cd scripts.

• 00:04:27 Type activate like this and the virtual environment

• 00:04:29 is activated.

• 00:04:30 Whatever you install, and when you are running the application, you have to have activated

• 00:04:35 this virtual environment, then type cd.. , cd.. it will go to the one of

• 00:04:41 the upper parent folder, then execute this command.

• 00:04:45 This will install all of the requirements.

• 00:04:48 However, you may get some errors like warning retrying.

• 00:04:53 These errors are caused by the connection

• 00:04:55 pypi.org.

• 00:04:57 This host is pretty unstable, therefore you may be needed to execute this command

• 00:05:05 multiple times.

• 00:05:06 Each time it will cache one of the downloaded file.

• 00:05:09 So after a while, you will be able to install all of the requirements.

• 00:05:13 You see even I had

• 00:05:14 right now.

• 00:05:15 So it re-tried to download and it was successful.

• 00:05:20 The installation has been completed.

• 00:05:21 Now one more step is left, we need to download this iiswapper onnx file.

• 00:05:29 Just click this link, it will download the file.

• 00:05:31 Once the file has been downloaded.

• 00:05:33 Right click and cut it, put it into your cloned folder

• 00:05:38 root directory like here, directly into roop directory.

• 00:05:43 And now we are ready.

• 00:05:45 There are several arguments to run this application.

• 00:05:48 This is from the official repository of this project.

• 00:05:52 So I prefer to run with this.

• 00:05:54 So what are these doing keep frames.

• 00:05:57 So when this application is running, it will extract every one of the frame of the input

• 00:06:04 video, then replace the face and then it will reconstruct the video.

• 00:06:09 Keep frames means that it will keep the generated frames.

• 00:06:12 This can be very useful because then you may apply some face restoration, such as CodeFormer

• 00:06:20 of Automatic1111.

• 00:06:21 I will also show that.

• 00:06:22 Keep FPS.

• 00:06:23 This will keep the FPS of the

• 00:06:26 video and number of cores.

• 00:06:28 Now this is important.

• 00:06:29 If you don't set this, it will use all of your

• 00:06:32 CPU cores and it will use huge amount

• 00:06:35 of RAM memory.

• 00:06:36 So I suggest you to begin with like this, turn

• 00:06:40 a small short clip into a deep fake video.

• 00:06:44 Then if it works, you can increase the number of cores.

• 00:06:48 So I will start my application with this command.

• 00:06:51 For starting the application interface

• 00:06:53 we need to first activate the virtual environment.

• 00:06:57 So enter inside the virtual environment folder, enter inside scripts, type CMD.

• 00:07:03 It will start in this directory, type activate, then cd.., cd.. and you will be

• 00:07:09 in the root directory.

• 00:07:11 Then copy this, for example, paste it and it will start the user interface for you.

• 00:07:17 You should get a message like this.

• 00:07:19 So the user interface is pretty easy to use.

• 00:07:22 I will select face.

• 00:07:23 As an example showcase, I will use face2 and select your target

• 00:07:29 target.

• 00:07:30 Means that your target video.

• 00:07:31 I will use video2.

• 00:07:33 So this is the selected face and this is the video it will be applied.

• 00:07:37 Then click start.

• 00:07:38 When you click start it will ask you which should

• 00:07:40 be the output video file name.

• 00:07:42 I will

• 00:07:43 use face_changed2.

• 00:07:44 Click Save and it will start processing.

• 00:07:48 So when it starts processing, it will generate a folder with the name of your input video

• 00:07:53 file video2 and video2 as you are seeing.

• 00:07:56 So it will first extract every one of the

• 00:07:59 frames like this.

• 00:08:01 Then it will start replacing them.

• 00:08:03 So this is the original frame and after

• 00:08:06 a while we will see it is replaced.

• 00:08:08 Patiently wait and do not touch the computer until you

• 00:08:12 see video is saved as like this.

• 00:08:15 So this was the original frame and this is the

• 00:08:18 face swapped frame.

• 00:08:20 Currently the biggest

• 00:08:21 disadvantage is the resolution of the face swapping

• 00:08:25 model is very low.

• 00:08:26 Therefore, we can

• 00:08:28 apply a face enhancer or the face has to be somewhat distant shot.

• 00:08:33 So right now we are

• 00:08:34 watching the side-by-side comparison of the original video and the converted video.

• 00:08:41 If

• 00:08:42 you encounter any problem, then you can use these video files and these images you are

• 00:08:48 seeing here as a test cases.

• 00:08:50 I have shared these video clips and the images on my Patreon.

• 00:08:55 The link is here.

• 00:08:56 If you also support me on Patreon I would appreciate that very much.

• 00:09:00 Also, please join Discord and ask me any questions that you want.

• 00:09:03 Please also don't forget to

• 00:09:04 star our repository and watch it.

• 00:09:07 Now I will show you how to use this amazing repository

• 00:09:10 on Google Colab.

• 00:09:11 So I made a Colab like this: click the link here, then click open in Colab.

• 00:09:17 Using Colab is super easy.

• 00:09:19 First, just click connect, then make sure that you have GPU.

• 00:09:24 If

• 00:09:25 you don't have GPU here, click change runtime and

• 00:09:27 select GPU here.

• 00:09:28 Click save.

• 00:09:29 It may require

• 00:09:31 you to reconnect.

• 00:09:33 Actually, yes, it has reconnected.

• 00:09:36 Once it reconnects it will delete everything

• 00:09:38 and when you click here and you see GPU, disk, and system ram, that means you are ready.

• 00:09:43 Then

• 00:09:44 click this play icon here.

• 00:09:45 Click run anyway, everything we upload here will get deleted

• 00:09:50 after we terminate our session.

• 00:09:52 So don't worry about privacy.

• 00:09:54 It will download and install

• 00:09:55 everything with just one click.

• 00:09:57 You will also see an error and warning message like this.

• 00:10:00 You can ignore them, it is just working and everything is installed and downloaded.

• 00:10:05 Now

• 00:10:06 we are ready to generate our deep fake video.

• 00:10:08 So everything will be in our runtime folder.

• 00:10:11 To open it click here as you are seeing.

• 00:10:13 We will upload our files into this roop directory.

• 00:10:18 Right click, click upload.

• 00:10:19 So for this example, I will use face1 and video1.

• 00:10:24 I will click

• 00:10:25 upload.

• 00:10:26 You see it says that ensure that your files are saved because everything will get

• 00:10:30 deleted after we terminate our session.

• 00:10:33 So meanwhile, when the file is getting uploaded,

• 00:10:36 you will see a status like here.

• 00:10:38 Okay, the upload has been completed.

• 00:10:40 Now you need to change this

• 00:10:42 input parameters.

• 00:10:44 The face that I am going to use is face1.png.

• 00:10:47 The extension is also

• 00:10:49 important.

• 00:10:50 Don't forget that.

• 00:10:51 The input video is video1.mp4 Don't forget that and the output

• 00:10:57 will be face_changed_video.

• 00:10:58 You can type any name you wan.

• 00:11:01 I will keep frames, keep fps and

• 00:11:04 I will use gpu.

• 00:11:05 Just click play icon and now we will see more system ram and gpu ram usage.

• 00:11:10 First

• 00:11:11 it will download the necessary model files.

• 00:11:14 This is only one time.

• 00:11:15 This will also happen on your

• 00:11:16 computer as well, but these are all one-time downloads.

• 00:11:20 Then it will extract all of the frames.

• 00:11:23 So it is starting to extract frames.

• 00:11:25 You see the frames are being extracted into video1 folder.

• 00:11:29 You can also download this folder.

• 00:11:31 For downloading you need a special code.

• 00:11:33 Actually I will get that

• 00:11:34 code with ChatGPT.

• 00:11:36 Give me a Google Colab code to download a directory.

• 00:11:42 You see I am showing you

• 00:11:43 how you can also use this.

• 00:11:44 Just copy code, add a code, paste it here and let's give its path.

• 00:11:49 Copy

• 00:11:50 path.

• 00:11:51 So this is the path and it will allow us to download

• 00:11:54 all of the frames like this.

• 00:11:56 I will show

• 00:11:57 that once this is done.

• 00:11:58 So when you see swapping in progress like this, it means that it started

• 00:12:03 swapping the faces so all frames were extracted.

• 00:12:06 It is only using three gigabyte gpu.

• 00:12:09 Once the

• 00:12:10 author figures out how to use it on our computer.

• 00:12:13 Even with four gigabyte vram having gpu, you can

• 00:12:16 use this on your computer.

• 00:12:17 Okay, the file is generated.

• 00:12:20 It is in the roop folder.

• 00:12:22 Face changed

• 00:12:23 video, right click and download.

• 00:12:24 And let's also download the repository of our frames.

• 00:12:28 Here.

• 00:12:29 Yes.

• 00:12:30 Click play icon.

• 00:12:32 It will generate a video one zip file.

• 00:12:35 The zip file is generated but it failed to

• 00:12:37 automatically download.

• 00:12:38 Just right click and download and inside there we will see all of the

• 00:12:43 frames.

• 00:12:44 Of course this file is bigger.

• 00:12:45 Now I will show you the face swapped video.

• 00:12:51 Okay, currently we are seeing the face swapped video.

• 00:12:54 The audio is also there as you are seeing.

• 00:12:57 It's pretty decent quality when we consider we

• 00:13:00 only use single image without any training.

• 00:13:03 We can even improve the quality.

• 00:13:05 How?

• 00:13:06 Now I will show you that with face restoration of automatic

• 00:13:10 1111 web ui.

• 00:13:11 So I will start my automatic 11 web ui from here.

• 00:13:15 If you don't know how to use

• 00:13:16 Automatic1111 web UI.

• 00:13:18 I have amazing tutorials for this on my Stable Diffusion repository.

• 00:13:24 You

• 00:13:25 see I start from beginning.

• 00:13:26 I show pretty much everything that you need with Automatic1111 web

• 00:13:30 UI.

• 00:13:31 I will also put the link of this page in the linked file so you can watch all of these

• 00:13:38 videos

• 00:13:39 and learn many more things about Automatic1111 web UI, including training, face training,

• 00:13:45 some custom scripts like this, and many more things.

• 00:13:49 Also, I did put two video links in this

• 00:13:52 file which will show you how to install and use Automatic1111 web UI.

• 00:13:56 It's in the bottom of the

• 00:13:58 this page.

• 00:13:59 So my Automatic1111 web UI is getting started.

• 00:14:02 Okay, the frame zip file also downloaded.

• 00:14:05 Let's move it into our directory.

• 00:14:07 Let's extract it video1.zip file.

• 00:14:10 So all of the frames

• 00:14:12 are here.

• 00:14:13 You see the face is changed.

• 00:14:15 However, we can improve the face quality.

• 00:14:17 So web UI started.

• 00:14:19 Let's open it.

• 00:14:20 How we are going to improve the face quality: go to the extras tab in here

• 00:14:25 for

• 00:14:26 example let's drag and drop this first frame into here and you may upscale it if you want.

• 00:14:32 However, I will just show you without upscaling.

• 00:14:35 This is the original resolution.

• 00:14:38 So what we are

• 00:14:39 going to do is select your upscaler from here.

• 00:14:41 You don't need to select anything if you wish.

• 00:14:43 I am

• 00:14:44 going to use 4x ultra sharp and CodeFormer visibility.

• 00:14:47 I will make this full like this and

• 00:14:51 I will generate.

• 00:14:52 So without upscaling it will only fix the face and let's compare it.

• 00:14:57 This is working

• 00:14:58 especially good if your face is close distance.

• 00:15:01 So let's see.

• 00:15:02 So this is the original frame and this

• 00:15:05 is the corrected frame.

• 00:15:07 You see there is a significant difference in terms of quality.

• 00:15:10 Okay, let me show you as a side-by-side comparison.

• 00:15:13 So here code former effect you are seeing right

• 00:15:16 now.

• 00:15:17 It is looking pretty amazingly better.

• 00:15:21 So now I will batch process them.

• 00:15:24 For batch processing

• 00:15:25 go to the batch from directory, give your zip

• 00:15:29 directory like this, then the output will be

• 00:15:32 let's say enhanced.

• 00:15:34 Okay and you see the same settings are applying.

• 00:15:38 Just hit generate.

• 00:15:39 It will

• 00:15:40 process every one of the frames into this folder

• 00:15:44 and then we will use these frames to regenerate,

• 00:15:48 reconstruct our video.

• 00:15:49 This may take a while depending on your computer hardware.

• 00:15:53 Moreover,

• 00:15:54 if you feel that the face is too much changed, you

• 00:15:56 can reduce the code former visibility, and

• 00:15:59 code former weights from here like 50 percent.

• 00:16:02 In both cases, you can play with these values

• 00:16:05 and decide how much effect should CodeFormer make

• 00:16:10 on your fixed faces.

• 00:16:12 Moreover, once you are

• 00:16:13 done with Google Colab, click here, disconnect and

• 00:16:16 delete runtime and it will delete everything

• 00:16:20 and you won't have any privacy issues.

• 00:16:22 The face restoration is completed.

• 00:16:25 All of the frames are

• 00:16:26 here.

• 00:16:27 318.

• 00:16:28 However, I noticed something else.

• 00:16:32 Currently when I open the output video file

• 00:16:36 of the repository when I click properties I see that

• 00:16:40 the the video fps is 23.

• 00:16:43 However, when I open

• 00:16:45 the original video, the video fps is 23.976 therefore,

• 00:16:54 there is a significant issue here

• 00:16:56 which may cause audio desynchronization, so reconstructing

• 00:17:03 video from the extracted frames

• 00:17:05 looks like better way currently instead of depending the output video of the repository.

• 00:17:11 So for reconstruction of the video, I will use

• 00:17:13 Davinci Resolve free edition and I will rename

• 00:17:16 all of the frames.

• 00:17:18 To be sure just copy paste and make a backup of them, then select all of

• 00:17:22 them.

• 00:17:23 Right click rename a1.

• 00:17:25 So all frames are renamed like this.

• 00:17:27 Start a new project.

• 00:17:29 Go to the first

• 00:17:30 tab here in the three dots here.

• 00:17:33 Make the frame display mode sequence, then in the right bottom

• 00:17:37 click project settings, set the settings as you

• 00:17:40 wish.

• 00:17:41 I will set the frame rate like this.

• 00:17:43 Save

• 00:17:44 and also let's set up the our resolution like this

• 00:17:48 one.

• 00:17:49 You can set any resolution, then drag and drop

• 00:17:51 all of the frames into here.

• 00:17:54 You see now it is imported as a sequence and then drag and drop

• 00:17:58 it here and now we can see the generated video with correct fps value.

• 00:18:04 When I right click and

• 00:18:05 see the clip attributes I will see the video frame

• 00:18:08 rate which is the original one and now this

• 00:18:11 is our face restored video.

• 00:18:14 Now I will compare it with the output of the repository.

• 00:18:18 So let's also

• 00:18:20 add it into our Davinci Resolve which is here.

• 00:18:24 Drag and drop here.

• 00:18:26 Let's also append it.

• 00:18:28 You see it has

• 00:18:29 a longer duration than the reconstructed video because

• 00:18:34 of the incorrect fps value.

• 00:18:36 So let me set

• 00:18:37 its position like this so we can watch side by side.

• 00:18:41 So the left side is the face enhanced and

• 00:18:44 the right side is the original output of the repository,

• 00:18:50 deepfake repository.

• 00:18:51 Now let's also

• 00:18:52 compare it with the original footage which is here.

• 00:18:55 I will drag and drop it in here so we can compare

• 00:18:58 better.

• 00:18:59 By the way, with this way, you can also apply the correct sound so I will just unlink them.

• 00:19:06 Delete this sound and now it has the original sound.

• 00:19:10 So let's also disable this with d key on

• 00:19:14 your keyboard.

• 00:19:15 Let's put the original video to left and let's put our deep face to the right which

• 00:19:21 is

• 00:19:22 here.

• 00:19:23 Let me see.

• 00:19:24 Okay, in the left now we see our deep face and in the right we see the original

• 00:19:28 footage.

• 00:19:29 Let's watch it.

• 00:19:30 Okay, right now we are seeing on the left, our deep fake and on the

• 00:19:35 right we are seeing the original footage.

• 00:19:38 You see it is pretty amazing.

• 00:19:40 However, since we have applied

• 00:19:42 face restoration, there are some minor issues that can be visible on the high resolution,

• 00:19:49 but on the low resolution it is pretty accurate and pretty consistent and pretty high quality.

• 00:19:55 Okay, let's watch it one more time.

• 00:19:57 On the left side we are seeing the deep fake.

• 00:19:59 On the right

• 00:20:00 side we are seeing the original footage.

• 00:20:03 You can see the lips synching, the emotions and everything

• 00:20:07 looking pretty good and accurate.

• 00:20:09 And then you can export it.

• 00:20:11 I already have an amazing tutorial for

• 00:20:13 Davinci Resolve free edition as well.

• 00:20:15 You can use any software you want.

• 00:20:17 So both Automatic1111

• 00:20:18 web UI tutorials and Davinci Result tutorial are also

• 00:20:22 added to the file that you will find the link in

• 00:20:24 the description.

• 00:20:25 Then all you need to do is just export the video as you wish after you set it up.

• 00:20:30 It will also have synchronized audio of the original

• 00:20:34 video as well.

• 00:20:35 This is all for today.

• 00:20:36 I hope you have enjoyed.

• 00:20:37 Please subscribe, join, support me on Patreon.

• 00:20:40 This is really, really

• 00:20:41 important if you support me on Patreon, it helps me

• 00:20:44 tremendously.

• 00:20:45 I'm also sharing Patreon exclusive

• 00:20:48 content as well.

• 00:20:49 You can see them here.

• 00:20:50 So you won't regret by supporting me on Patreon.

• 00:20:54 Also, the link

• 00:20:55 for this file will be in the description of the video.

• 00:20:58 Let me show you as an example.

• 00:21:01 You will see

• 00:21:02 in here in this video and also you will see it in

• 00:21:04 the comment section as well.

• 00:21:06 I am 100% open to

• 00:21:09 consultation, helping you privately.

• 00:21:11 One by one.

• 00:21:12 Just become our Patreon supporter.

• 00:21:15 Also, please

• 00:21:16 join our Discord server.

• 00:21:18 You can join our Discord server from clicking this link here it will

• 00:21:22 open

• 00:21:23 our Discord server like this as you are seeing right now.

• 00:21:25 This is the best way to contact me.

• 00:21:27 You can also click the Discord server in the description

• 00:21:30 of the video and also in the pinned

• 00:21:32 comment of the video.

• 00:21:34 If you have any questions, just leave a comment.

• 00:21:36 Ask me.

• 00:21:37 You can also join

• 00:21:38 our Discord server as I said.

• 00:21:40 Hopefully see you in another amazing tutorial later.