Now I have a working version to show.
Currently it outputs a video with an orvelay, showing where the mask was created, there is a image sequence output in PNG with a mask, and there is also a possibility of creating inpaint.
All done throught blender.
The video was edited to be faster.
There are some things to tweak and more testing to do, but probably a WIP version is coming out this week for the Patrons and subscribers of 2023-06-13 23:21:49 +0000 UTC View Post
Here is a preview of something I'm working on, with this solution you will be able to interactively select where do you want a mask to be created and export it to be used for compositing or something else.
As soon as I finish the part of the process to create a maks throughout the video I'll release a WIP version
This addon lets you create 4 maps from a single ou multiple images, that you can use to texture your assets
It was made based on the code at https://github.com/valentin-deschaintre/multi-image-deepNet-SVBRDF-acquisition
To install the addon you need the zip file attached and the model, that you can download from the CEB Multi Image SVBRDF Chackp...
2023-06-03 19:33:19 +0000 UTC View PostThis is a test for a solution to make easy replacement of character. This was a suggestion from a discord user.
The idea is to make a fully working solution to use inside blender, and to make it to be possible to do in simple steps.
Here is a simple test of something I'm working on.
On this solution you will be able to create a PBR material (diffuse, normal, roughness and specular maps) from a single or multiple images
2023-05-27 21:05:59 +0000 UTC View Post
Here is a test done with RVC to clone a voice in Portuguese BR.
One of the best and easiet free voice clone solutions I've tested.
I made this test folowing this tutorial from NerdyRodent
2023-05-25 09:55:02 +0000 UTC View Post
Another great code to create voice from text is BARK, that you can see the original code at: https://github.com/suno-ai/bark
Using this addon you will be able to generate voice from several languages, unlike Tortoise TTS.
Some information from the github page that you can use on the text prompts:
Head to https://carlosedubarreto.gumroad.com/l/sinmdm to get your copy.
This version is the 0.02, on gumroad also has 11 models that are ready to use.
Here is the tutorial on how to install
And the tutorial on how to create a custom mod...
2023-05-13 09:55:36 +0000 UTC View PostThis is another test done using 2 cellphone cameras.
I feel that having this kind of information on the 3d space can be very helpful to learn the real movements, it seems to be more interesting because even with 2 cameras you can have some nice results
I plane to make some tests using this method to help on animation creation
Attached you have the source files for this 4d scan
This scan was done using 2023-05-10 08:18:22 +0000 UTC View Post
Today I recorded the process to install the vid2avatar code.
This repository is very complex to make it work on a windows machine. and as my pc already have a lot of complicated tool installed, I went for a AWS EC2 G5 instance to record what would be a tutorial
But... I had lots of problems and couldnt make it work, and also got confused several times because, it really is a complicated task.
In the end the video I recorded dont show the vid2avatar work...
2023-05-02 11:00:18 +0000 UTC View PostThis was the first test running vid2avatar and having the result in blender.
I still have to try to create using my own data. But I think the worst part is done.
If you want to try the code, its available at:
https://github.com/MoyGcc/vid2avatar
2023-04-30 22:39:04 +0000 UTC View PostATTENTION: you can find the videos I used on the attachment of this post
This video shows a test done using the 4D Scan of Agisoft Metashape Pro (this option is available on the Professional version)
https://www.agisoft.com/features/professional-edition/
You can have a trial of 30 days
To do it was very simple. I recorded a video u...
2023-04-25 16:08:00 +0000 UTC View PostTesting some ways to make a self scan.
The process involves:
How to do it:
Setup your smartphone in a place that you can be at your own chair and can record your face.
while on the chair test rotating yourself on the chair and look to see if you are going to be inside the frame of the recording.
press record on your smartphone and...
2023-04-23 08:46:29 +0000 UTC View Post
Install version 0.02 to follow this tutorial
Get version 0.02: https://www.patreon.com/posts/81862227
Steps done in the video:
- get an animation
- retarget it to a mixamo armature
!!IMPORTANT STEP
ROTATE X -90
APPLY TRASNFO... 2023-04-21 21:50:06 +0000 UTC View Post
This version has the Advanced button to make it easier to follow the "creating new model" tutorial
Creating new model tutorial: https://www.patreon.com/posts/81861586
2023-04-21 21:49:14 +0000 UTC View PostModels from the original repository but configured to work on CEB SinMDM and Model created on the tutorial showing how to create your own model
Get the Addon (wip version for Patrons)
https://www.patreon.com/posts/81862227
2023-04-21 19:50:07 +0000 UTC View Post
This addon enables you to create multiple animations using a trained model as input.
you can create huge amount of new variations and even bigger animations then the original one.
This addon was made using the repository https://github.com/SinMDM/SinMDM
This is a work in progress and its already working, I'm just taking the time to proper organize it and make it in a more pro...
2023-04-20 22:18:12 +0000 UTC View PostA very interesting repository showed up, SinMDM. With it we can generate random animations based on a single animation.
We can extend it, create variations and much more.
The video is a simple example using one of the models that were ready to use.
The armature on the front is the original animation, the 5 animations behind, were randomnly generated based on the front.
For more information
This update solves a problem in the python dependencies on the installation
download all the parts and unzip it, you will have the zip file to install in blender
2023-04-11 15:46:56 +0000 UTC View Post
To get the version used on this tutorial got to this page:
https://www.patreon.com/posts/ceb-easymocap-1-80851440
For the latest free version, go to:
https://carlosedubarreto.gumroad.com/l/ceb_easymocap_multicam/ckwed2w
2023-04-02 18:41:13 +0000 UTC View Post
For Non_Commercial Use
Here is a video showing how ot use.
I plan to make a full tutorial tomorrow, but the way the addon was created should not be complicated to use.
Importante information, you must have OpenPose in your pc to make it work, and you can get it here https://github.com/CMU-Perceptual-Computing-Lab/openpose/releases/tag/v1.7.0 2023-03-31 22:48:59 +0000 UTC View Post
On this tutorial I explain how to use CEB 3D Face DFNRMVS to create high quality meshed using a couple of images
Important Info: The input images needs to be 240x240 pixels
You can download the addon attached on this post, or get it from the gumroad for free
https://carlosedubarreto.gumroad.com/l/dfnrmvs
On the gumraod you have also more information on how to install, mainly the matl...
2023-03-20 10:29:37 +0000 UTC View PostHere is a detailed tutorial on how to make the code DFNRMVS to work on your pc.
on the link you have attached a txt file with this explanation, might be better to follow.
Install miniconda, get it from:
https://repo.anaconda.com/miniconda/Miniconda3-latest-Windows-x86_64.exe
Download the ZIP from the sites bellow
This code aims to make high density (better definition) of human face. Its advantaga is that you can more pictures of different angles that will make the final result even better.
To use this code, you can get it at:
https://github.com/zqbai-jeremy/DFNRMVS
One important information. The installation is not simple and needs Matlab to generate one of the needed files. 2023-03-18 11:38:11 +0000 UTC View Post
Here is a test done using ZeoDepth (https://github.com/isl-org/ZoeDepth)
This tool helps you to create depth maps using monocular data.
I found that the results are very detailed.
You can have the blend file with the test I've done attached bellow (I had to remove 2 images so it can fit the 200mb limit from Patreon.
Do you think its a valid blender addon to be done?
2023-03-15 07:46:31 +0000 UTC View Post
Just found about some interesting links to download objects packs.
You have free option and paid option.
If you wnt to try the free ones, you can get packs with 10 files at: https://cubebrush.co/dadadesign
If you liked it and want to get the full 600+ files, you can get it at:
a few days ago was released Next3d, this code make possible that you create random 3d faces and also reenact some other video.
this example will show only the random 3d face creation.
If you would like to use it you have to go to this repo
https://github.com/MrTornado24/Next3D
for windows user, the worst part will probably be the installation of pytorch3d.
There is an explanation on how...
2023-03-12 05:55:02 +0000 UTC View Post
Testing another solution to create faces from photo, this one has better texture quality.
This was done using the script
https://github.com/cleardusk/3DDFA_V2
2023-02-27 14:59:32 +0000 UTC View Post
This addon lets you integrate Gamesynth with blender.
In case you dont know gamesynth, its a software that enables you to create all sorts of sound effects.
With this addon, now you can create the sound files without leaving blender.
If you are interested in gamesynth, you can buy a copy at:
http://tsugi-studio.com/web/en/products-gamesynth.html
This addon lets you make the color of one image to match another image color
You need to have a source image and a reference image to make it run.
Get the addon attached on this post
2023-02-24 23:34:00 +0000 UTC View Post