What is Stable Diffusion?

Introduction to AI Image Generation

Brief Model Overview

Stable Diffusion is an AI model for image generation developed by Stability AI. It belongs to the family of diffusion models, which are based on a step-by-step removal of noise.

The core functions of Stable Diffusion are:

  • Text-to-Image Generation
  • Image-to-Image Transformation
  • Inpainting and Outpainting (Completing existing images)
  • Controlled Image Generation (ControlNet)
Stable Diffusion Process

The Diffusion Process: From Noise to Image

Online versus Local Operation

Online Services

  • ✓ No installation required
  • ✓ Immediately usable
  • ✓ No powerful hardware needed
  • × Usually paid or limited
  • × Limited customization options
  • × Potentially less privacy

Examples: Midjourney, DALL-E, Leonardo.AI

Local Installation

  • ✓ Complete control
  • ✓ No running costs
  • ✓ Custom models & extensions
  • × Requires powerful GPU
  • × Technical know-how required
  • × More complex setup

Interfaces: A1111 Webui, ComfyUI, InvokeAI

Did you know?

Stablehub uses Stable Diffusion in the backend, offering you the benefits of local operation without the installation effort!

Positive ↔ Negative Prompt & Parameters

Positive Prompt

The positive prompt describes what should appear in the image. The more detailed and precise your wording, the more targeted the model can generate.

A majestic castle on a mountain, sunset, dramatic lighting, detailed architecture, 8k, cinematic, fog, mystical atmosphere

Negative Prompt

The negative prompt specifies what should be avoided in the image. This helps reduce unwanted elements or quality issues.

blurry, bad anatomy, distorted, deformed, disfigured, poor quality, low resolution, text, watermark

Parameter Control

CFG Scale 7.5

Controls how strongly the prompt is followed. Higher values = more prompt adherence.

Steps 30

Number of diffusion steps. More steps = more details, but longer generation time.

Seed 1234567890

Starting point for the random generator. Same seed = reproducible images.

Learning Objective Check

You should now understand:

  • What Stable Diffusion is and how it fundamentally works
  • Differences between online services and local installation
  • The role of positive and negative prompts
  • Why parameters are important and how they influence the result