2.4 KiB
2.4 KiB
Meld Auto Exposure (MeldAutoExposure)
A node that analyzes image brightness (luminance) and automatically applies gamma correction to approach a target brightness level. It is useful for brightening generated images that are too dark, or conversely, toning down images that are too bright.
Connection Image:
KSampler -> VAE Decode -> [Meld Auto Exposure] -> Save Image
Input
Required
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE |
- | The input image to correct. Supports batch input, where each image is analyzed and corrected individually. |
| correction_strength | FLOAT |
0.5 |
Correction Strength (0.0 - 5.0). Higher values apply stronger force to reach the target brightness. At 0.0, no correction is applied (gamma 1.0). Start around 0.5 and increase if the effect is too weak. |
| target_brightness | FLOAT |
0.5 |
Target Brightness (0.1 - 0.9).0.5 is medium brightness.Increase (e.g., 0.7) for brighter results, or decrease (e.g., 0.3) for darker results. |
Optional
| Name | Type | Default | Description |
|---|---|---|---|
| gamma_limit_min | FLOAT |
0.2 |
Minimum gamma limit. Prevents the image from being corrected to be extremely dark. |
| gamma_limit_max | FLOAT |
2.5 |
Maximum gamma limit. Prevents the image from being corrected to be extremely washed out (white). |
Output
| Name | Type | Description |
|---|---|---|
| image | IMAGE |
The brightness-corrected image. |
| calculated_gamma | FLOAT |
The actual applied gamma value (calculated). Returns the average for batch inputs. |
| current_brightness | FLOAT |
The average luminance of the original image. Returns the average for batch inputs. |
Tips & Mechanism
-
Auto-Adjustment Logic: Calculates the average luminance (Rec.601) of the image and derives a gamma value based on the difference from
target_brightness.- If the original is darker than target → Corrects to be brighter.
- If the original is brighter than target → Corrects to be darker.
-
Tuning Tips:
- First, set
target_brightnessto your preferred brightness level. - If the correction causes quality degradation (e.g., contrast is too harsh), lower the
correction_strength. - Conversely, if the correction is insufficient, increase
correction_strengthor widen the limiter range (gamma_limit_min/max).
- First, set