santa hat
deerdeer nosedeer glow
Sign In

Samplers, Schedulers, and Sigmas, Oh My! Cutting Through Denoise

Samplers, Schedulers, and Sigmas, Oh My! Cutting Through Denoise

Samplers, Schedulers, and Sigmas, Oh My!

NEW! Companion workflow with some advanced techniques and more implementation examples: https://civitai.com/models/987817

Samplers and schedulers are an intrinsic part of AI image generation, but confusing vocabulary and lack of information can shroud these concepts in mystery. The reality is actually quite boring - they dictate the formula used for the 'noise' added to and taken from images during training and image creation, and they specify and sometimes guide the strength of image creation in each step, typically by dividing the steps evenly along a xy curve (see above graphic depicting the curves of some popular schedulers).

Sigmas: The Step Weights

Sigmas are simply a list of numbers that control how strong each denoising step is:

  • Example: [17, 15.2, 12.0, 10.3, 7.3, 3.0, 2.1, 3.2, 0, 0]

  • Some schedulers add redundant zeros at the end, others don't

  • Higher numbers = stronger denoising (image generation) steps

Schedulers: The Step Managers

Schedulers are rule sets that:

  • Generate sigma weights for your requested steps

  • Only need to know the model's proper range (starting number)

  • Create steps using specific curve formulas

  • In ComfyUI, this is all that "samplers" actually do

  • More advanced schedulers can:

    • Evaluate predicted noise shifts

    • Adjust weights based on actual vs. predicted noise shifts

    • Perform other optimizations

Samplers: The Noise Controllers

These define the mathematical patterns for:

  • Adding noise during training

  • Driving the AI's noise prediction algorithms during image creation

The Denoising Parameter: Just the final % of steps

  • 1.0 for txt2img

  • 0.01-0.8 for img2img

  • Specifies where in the sigma schedule to start/stop

  • Example: 0.2 denoise = using only the final 20% of sigma steps

How Steps and Denoise Interact in Most ComfyUI Nodes

When you set:

  • 0.2 denoise at 10 steps

    • System calculates 10/0.2 = 50 total steps

    • Gets all 50 steps from scheduler

    • Uses only the final 10 steps (last 20%)

Alternative approach:

  • Some nodes allow you to specify total steps and start/end points directly

    • 50 total steps, starting at step 40 = identical to 0.2 denoise/10 steps

    • More transparent for what's actually happening

    • Easier to automate

  • ComfyUI's 'split sigmas' nodes can divide steps between different operations

    • (see example images below)

Key Takeaways

Understanding that denoise is simply the percentage of steps where you start during img2img, and that sigmas are just a list of step strengths (high to low), you can more confidently use different interfaces and also set up more complex workflows. This knowledge enables you to 'split sigmas' between different schedulers, changing settings, prompt or lora weights, upscaling, or more.

A final note to leave with - not all platforms implement schedulers the same way, and you can get more out of them if you use software that supports a more advanced scheduler. The diffusers back end implements flowmatch schedulers that are capable of adjusting sigma weights during generation for better results (and in fewer steps). That being said, you can still get similar quality with simpler schedulers, it will usually just take more steps and more tweaking.

Here's a link to some good information on schedulers: https://huggingface.co/docs/diffusers/api/schedulers/overview


EXAMPLE WORKFLOW

companion workflow pictured below - download it and try out some of the different techniques!

https://civitai.com/models/987817

and here are some basic configurations for reference (these do not contain specific setting suggestions, but simply illustrate how to wire the split sigmas node)



12

Comments