User Guide Cancel

Generate workspace | Substance 3D Viewer

The Generate workspace is where you can generate beautiful imagery by using your 3D scenes as input to Adobe Firefly. It's a great way to quickly create 2D visuals without relying on complex and time-consuming 3D rendering techniques.

To start generating content:

  1. Open the Generate workspace.
  2. Select either Text to 3D, or 3D model to image for your preferred generation workflow.
  3. Enter a prompt and adjust available settings.
  4. Click Generate and wait a few seconds for your content to be generated.

More information is available below on the different workflows and settings available in the Generate workspace.

Text to 3D

With Text to 3D, you can generate a brand-new 3D model using only a text prompt as input. These generated models can be very useful when combined with the “3D model to Image” feature.

Here are a few prompts you can try to get you started:

A white rabbit with long earsDescription automatically generated
Plastic white matte bunny with a mustache toy figurine

A brown boot with lacesDescription automatically generated
A brown boot that features a thick sturdy design with a high top and a chunky sole

A group of objects on a white backgroundDescription automatically generated

What 3D format does Text to 3D use?

3D models generated by Text to 3D use a specific representation called “Gaussian splatting”. These models have some advantages but also some limitations.

Advantages

  • It’s really quick to generate!
  • It’s lightweight and fast to render
  • It is good at representing things typically hard to represent in 3D, like fuzzy or translucent surfaces.

Limitations

  • Materials and appearance cannot be edited
  • Lighting changes have no effect
  • It cannot be exported to most traditional 3D file formats.

Some of these limitations may be lifted in the future as this technology evolves and matures.

3D scene to image

With 3D scene to image, you can generate stunning visuals based on your own 3D models and a text prompt in just a few seconds.

There are many options available in Substance 3D Viewer (beta) to control and fine-tune the generated image, but the most important are the Composition modes:

Generate a composite scene

This mode will keep your 3D model intact and fill the frame around it with a context of your choosing. This is useful when you are working with a high-quality 3D asset of an existing product and you want to make sure to keep the silhouette, look and colors of your model untouched.

Decreasing the Match Color value underneath the Composition dropdown allows you to experiment with different looks for your 3D object while maintaining its silhouette and overall shape.

Compositing a 3D model of a pair of headphones into a Firefly generated context: Prompt: “Headphones in a library”
Compositing a 3D model of a pair of headphones into a Firefly generated context: Prompt: “Headphones in a library”

Generate a new image

This mode will generate an image over the entire frame, using your 3D model only as a reference to drive the generation. This is a great way to create brand new illustrations while keeping creative agency over the image composition.

Decreasing the Match 3D shapes value underneath the Composition dropdown allows Firefly more freedom when interpreting the 3D scene. The lower the value, the more Firefly will interpret your 3D scene as a suggestion rather than a strict direction.

Generating a brand new illustration using the 3D model as a reference.
Generating a brand new illustration using the 3D model as a reference.

Generate a new look for my model

This mode will only replace your 3D model with a generated image and keep the rest of the frame transparent for easy insertion and compositing in any document or project. It’s useful to quickly create a variation of a model, ideate on new looks, or generate a high-quality image from an imperfect or even blank 3D model.

Increasing the Match color value underneath the Composition dropdown allows Firefly to consider the actual colors of the 3D model. A higher value will for example allow you to create a variation in a different style but still using the same overall colors as the original.

Generating an image of a new look for the headphones: Prompt: “Headphones with geometric graphic prints and black fabric earpads”
Generating an image of a new look for the headphones: Prompt: “Headphones with geometric graphic prints and black fabric earpads”

Settings and Effects

In addition to the 3 generative modes, Substance 3D Viewer gives you access to many effects and settings allowing you to achieve different looks and styles without overly complex text prompts.

Content Type

You can choose the generation to target either a more artistic or photorealistic look with this setting. Enabling the Auto switch will let the app choose the right type based on your subject and prompt.

Styles

There are 2 settings in the Styles section.

Visual Intensity: A lower value will make the result of the generation more subdued and neutral while a high value will push the generation towards more extravagant and extreme results.

Strength: This setting allows to finetune how much the selected Effects will affect the result.

Effects

Firefly features over 100 effects to choose from and combine to generate interesting new styles and looks for your images. You can also select various presets for Color and tone, Lighting and Composition to get as much control as possible on the generation output without editing your prompt or engaging in “prompt engineering”.

Previewing and exporting generated images

When an image has been generated, you can switch between 4 variations by selecting the thumbnails under the Variation section.

To save generated images:

  • Hover over the image in the viewport and use the Download button that appears.
  • Hover over a thumbnail in the Variation section, and use the options > download to save the variation.
Note:

The Snapshot and Export an image buttons do not include generated content. Additionally, generated images are always at a 1:1 ratio, and do not take canvas size into account.

 Adobe

Get help faster and easier

New user?