BSLIVE Using DeepBump and Midas AI to Generate Normal and Depth

Introduction

In this episode, we will explore several methods to generate normal and depth maps using AI, enhancing your 3D modeling workflow in Blender. We've previously looked at the Material Map Generator for creating PBR maps such as metalness and roughness. In this article, we'll focus on two powerful tools: DeepBump and Midas, which allow for the generation of depth maps based on images.

DeepBump for Blender

We'll start with the DeepBump add-on, compatible with both Mac and Windows. This tool can create normal maps, height maps, and curvature maps from a single image. To set it up, download the zip file and follow the installation process through Blender's preferences. After installation, it’s crucial to click the button to install the necessary dependencies.

Once you have DeepBump installed, you can begin by creating a plane in Blender and assigning a material. Here’s how to do it:

1. Open Shader Editor: Create a new image and select your texture (you can use AI-generated textures).
2. Connect the Texture: Link the texture to the base color of the material.
3. Generate Normal Map: Select the options to generate the normal map, which takes just a few seconds.
4. Adjust Settings: Tweak the strength of the normal map and switch to Material Preview shading. Add light sources to see the effects.
5. Create Height and Curvature Maps: After generating the normal map, you can easily create height and curvature maps by selecting the appropriate options.

Displacement in Cycles

You can use the height map for displacement purposes. To do this:

• Switch to the Cycles rendering engine.
• Connect the height map to the height displacement input and adjust the scale.
• Use a Subdivision Surface modifier for better detail and select the appropriate settings under material options.

Midas AI for Depth Generation

The next method we’ll discuss is the Midas algorithm, suitable for processing photographs and videos. Midas generates a depth map based on its understanding of the 3D structure in the images.

1. Image Preparation: An image must be in JPEG format. Convert your PNG images if needed.
2. Generate Depth Map: Upload your JPEG image to Midas, wait for a few seconds, and download the output.
3. Import into Blender: With the depth map ready, import it into Blender for further use.
4. Texture Displacement: Utilize the depth map from Midas for displacement on your 3D meshes.

Combining Outputs for Enhanced Results

By using both DeepBump and Midas, you can achieve more detailed and nuanced 3D results. You can combine normal and height maps for better surface detail, especially for complex models. Utilize techniques like mixing colors in Blender to merge the outputs from both methods effectively.

Conclusion

With the combination of AI tools like DeepBump and Midas, creating visually striking 3D models becomes more efficient and accessible. Both tools provide valuable resources for enhancing depth and normal mapping, crucial for high-quality texture work in Blender.

Keywords

DeepBump, Midas, normal maps, depth maps, height maps, curvature maps, Blender, PBR, texture generation, AI.

FAQ

Q1: What is DeepBump?
A1: DeepBump is a Blender add-on that generates normal maps, height maps, and curvature maps from a single image, optimizing the texture workflow in 3D modeling.

Q2: How can I use Midas?
A2: Midas uses an AI algorithm to generate depth maps from photographs. Simply upload an image in JPEG format, wait for the processing, and download the depth map for use in Blender.

Q3: Can I combine outputs from DeepBump and Midas?
A3: Yes, you can combine normal and height maps generated from both tools for enhanced surface detail and effective displacement in your 3D models.

Q4: Is DeepBump compatible with different operating systems?
A4: Yes, DeepBump is compatible with both Mac and Windows systems.

Q5: How do I adjust the displacement settings in Blender?
A5: In the Cycles rendering engine, connect your height map to the height displacement input, and you can adjust the scale for the desired effect.