High-Frequency Detail Sampling based on Res Sampling

This is a ComfyUI custom node that enhances fine detail preservation in diffusion model outputs using spectral high-frequency emphasis (HFE).

Installation

Clone or copy this folder into your ComfyUI custom_nodes directory:

ComfyUI/
  custom_nodes/
    RES4SHO/
      __init__.py
      sampling.py

Restart ComfyUI. The new samplers and schedulers will appear in the dropdown menus of any KSampler node.

Samplers

All samplers are exponential integrators with phi-function coefficients. The HFE enhancement extracts high-frequency detail from inter-stage correction deltas via a 3x3 spatial high-pass filter and re-injects it with configurable strength.

Fixed-Strength Presets

Each stage count offers 8 strength levels (s1 = no emphasis, s8 = maximum potential sharpness):

Sampler	Stages	Model Evals/Step
hfe_s1 .. hfe_s8	2	2
hfe3_s1 .. hfe3_s8	3	3
hfe4_s1 .. hfe4_s8	4	4
hfe5_s1 .. hfe5_s8	5	5

Higher stage counts provide better ODE integration accuracy at the cost of more model evaluations per step.

Adaptive (Auto) Samplers

Per-step adaptive eta based on sigma envelope and content gating:

Sampler	Stages	Description
hfe_auto	2	Variable c2, eta, and kernel per step
hfe3_auto	3	Per-step eta with 3-stage integrator
hfe4_auto	4	Per-step eta with 4-stage integrator
hfe5_auto	5	Per-step eta with 5-stage integrator

How auto adapts:

• Sigma envelope (smoothstep): suppresses emphasis at high noise (early steps), full strength in the detail-forming range
• Content gate: reduces emphasis when the model correction is already HF-rich; increases it when the correction is smooth and needs boosting

Experimental Samplers (hfx_*)

10 fundamentally different enhancement modes, each operating in a distinct mathematical domain. All use a 2-stage exponential integrator base.

Spatial Domain

Sampler	Method	Description
hfx_sharp	Spatial high-pass	Unsharp mask on eps_2 via 3x3 box blur residual
hfx_detail	Post-step injection	Extracts HF from denoised_2, injects after the integrator step

Value Domain

Sampler	Method	Description
hfx_boost	Uniform scalar	Amplifies eps_2 magnitude uniformly (effective "lying sigma")
hfx_focus	Power-law contrast	Nonlinear gamma curve on eps_2 element magnitudes -- large corrections amplified, small corrections unchanged

Frequency Domain

Sampler	Method	Description
hfx_spectral	FFT power-law	Reshapes eps_2 frequency spectrum with distance-based power-law boost
hfx_coherence	FFT phase gating	Amplifies frequency bins where eps_1 and eps_2 agree in phase; suppresses where they disagree

Temporal Domain

Sampler	Method	Description
hfx_momentum	EMA across steps	Accumulates denoised differences across steps via exponential moving average
hfx_stochastic	SDE noise injection	Adds scaled noise proportional to local HF content -- non-deterministic

Inter-Stage / Geometric Domain

Sampler	Method	Description
hfx_orthogonal	Gram-Schmidt projection	Extracts the component of eps_2 orthogonal to eps_1 (novel information only)
hfx_refine	ODE curvature map	Uses

Each mode has 4 graduated strength presets (_s1 .. _s4), e.g. hfx_sharp_s1, hfx_spectral_s3, hfx_refine_s4, etc. A per-step safety cap prevents compounding artifacts at higher strengths.

Schedulers

Arctangent S-curve schedulers that concentrate step density in the detail-forming sigma range:

Scheduler	Description
atan_gentle	Mild mid-sigma concentration
atan_focused	Moderate detail-range concentration
atan_steep	Aggressive detail-range concentration
karras_tan	Karras-Tangent hybrid (experimental)
logistic	Logistic sigmoid S-curve (experimental)

An ASCII sigma chart is printed to the console when a scheduler is used.

Recommended Combinations

Getting Started

Goal	Sampler	Scheduler	Notes
General use	hfe_auto	atan_focused	Best all-rounder -- adaptive emphasis handles most content
Subtle enhancement	hfe_s3	atan_gentle	Light touch, minimal risk of artifacts
Strong detail	hfe_s6	atan_steep	Noticeably sharper textures and edges
Maximum sharpness	hfe_s7 or hfe_s8	atan_steep	Aggressive -- inspect for over-sharpening

By Content Type

Content	Sampler	Scheduler	Why
Portraits / faces	hfe_auto	atan_focused	Auto gate protects smooth skin while sharpening eyes, hair, pores
Landscapes / nature	hfe_s5	atan_gentle	Fixed mid-strength avoids over-enhancing skies and gradients
Architecture / hard surfaces	hfe_s7	atan_steep	Strong emphasis on edges and geometric detail
Text / UI renders	hfx_sharp	atan_steep	Spatial high-pass targets glyph edges specifically
Fabric / organic texture	hfx_spectral	atan_focused	Frequency-domain emphasis across texture scales
Illustrations / anime	hfe_s4	atan_gentle	Light emphasis preserves flat shading without adding unwanted texture

High-Accuracy Integrators

More model evaluations per step for better ODE integration -- useful at low step counts or with difficult models:

Sampler	Scheduler	Use Case
hfe3_auto	atan_focused	Good balance of accuracy and speed (3 evals/step)
hfe4_auto	atan_focused	High accuracy for complex prompts (4 evals/step)
hfe5_auto	atan_gentle	Maximum integration accuracy (5 evals/step)
hfe4_s5	atan_steep	Fixed-strength detail + 4-stage accuracy
hfe5_s6	karras_tan	High emphasis + high accuracy + Karras hybrid spacing

Experimental Combinations

Sampler	Scheduler	Character
hfx_sharp	atan_focused	Spatial high-pass -- good default experimental choice
hfx_spectral	atan_steep	Frequency-domain power-law sharpening
hfx_refine	atan_focused	ODE curvature-adaptive -- sharpens where the model is least certain
hfx_coherence	atan_focused	Phase-coherence gating -- amplifies structurally confident frequencies
hfx_orthogonal	atan_focused	Novel-information extraction via Gram-Schmidt
hfx_momentum	atan_gentle	Temporal accumulation -- builds detail across steps
hfx_focus	atan_focused	Value-domain contrast -- amplifies dominant correction directions
hfx_stochastic	atan_gentle	Stochastic texture injection -- adds micro-variation
hfx_boost	atan_gentle	Uniform eps amplification -- simple signal boost
hfx_detail	atan_focused	Post-step HF injection from denoised output

Scheduler Pairings

Scheduler	Best With	Character
atan_gentle	Low-strength samplers (s1-s4), stochastic modes	Mild concentration, safe for all content
atan_focused	Auto samplers, mid-strength presets (s4-s6)	Balanced step density in detail range
atan_steep	High-strength samplers (s6-s8), architectural content	Aggressive detail-range concentration
karras_tan	High-stage integrators (hfe4_*, hfe5_*)	Karras optimal spacing + tangent warp
logistic	Any -- alternative S-curve shape	Sharper transition through detail range, flatter extremes

How It Works

Base integrator: Multi-stage singlestep exponential integrator (res_Ns) with phi-function coefficients, giving exact treatment of exponential decay and higher-order corrections from intermediate evaluations.

HFE enhancement (hfe_* samplers): The inter-stage correction delta captures what the model reveals at lower noise -- texture, edges, micro-structure. A spatial high-pass (residual after box blur in latent space) extracts the fine detail component, which is re-injected with extra weight eta. This compounds across every step.

HFX modes (hfx_* samplers): Each mode modifies the second-stage prediction (eps_2) using a different mathematical operation before the integrator update step. The 10 modes span 5 domains:

• Spatial: high-pass filtering (sharp), post-step HF injection (detail)
• Value: uniform scaling (boost), nonlinear power-law contrast (focus)
• Frequency: FFT power-law reshaping (spectral), inter-stage phase coherence gating (coherence)
• Temporal: EMA across steps (momentum), stochastic noise injection (stochastic)
• Inter-stage: Gram-Schmidt novel-component extraction (orthogonal), ODE curvature-adaptive gain (refine)

Safety: A per-step cap limits eps_2 modifications to 10% of the original RMS, preventing compounding artifacts across steps. A sigma warmup gate suppresses all enhancement at high noise levels (early steps). An img2img denoise gate scales down enhancement for partial-denoise schedules.

Cost: One 3x3 avg_pool per step for spatial variants; one FFT pair for spectral/coherence modes. All negligible vs. model evaluation. Auto samplers add a few scalar ops on top.

License

MIT