ComfyUI Node
⚗ Alchemica Image Mix (blended start latent)
Blends up to 4 photos (with weights) into one start latent for ⚗ AlchemicaKrea Pro img2img. Use denoise 0.65-0.80 so the sampler resolves the mix into ONE coherent person (lower = keeps more of the blend).
⚗ Alchemica Image Mix (blended start latent)
- vae
- image_1
- image_2
- image_3
- image_4
- latent
- preview
- width
- height
- report
◄weight_11.00►
◄weight_21.00►
◄weight_31.00►
◄weight_41.00►
◄aspect2:3►
◄megapixels2.00►
◄fitcenter crop►
◄modelatent►
◄contrast1.00►
Category⚗ AlchemicaMente/latent
Inputs (14)
| Name | Type | Default | Description |
|---|---|---|---|
| vae | VAE | — | |
| image_1 | IMAGE | — | |
| weight_1 | FLOAT | 1.000–4 | Weight of photo 1. Weights are normalised (2 and 1 = 67% / 33%). 0 = ignore this photo |
| weight_2 | FLOAT | 1.000–4 | Weight of photo 2. Weights are normalised (2 and 1 = 67% / 33%). 0 = ignore this photo |
| weight_3 | FLOAT | 1.000–4 | Weight of photo 3. Weights are normalised (2 and 1 = 67% / 33%). 0 = ignore this photo |
| weight_4 | FLOAT | 1.000–4 | Weight of photo 4. Weights are normalised (2 and 1 = 67% / 33%). 0 = ignore this photo |
| aspect | COMBO | 2:3 | Output shape. 'dal primo' = same aspect ratio as photo 1 |
| megapixels | FLOAT | 2.000.25–8 | FINAL resolution (the sampler works at this size) |
| fit | COMBO | center crop | How photos with a different shape are fitted. Cropping to the centre is best for portraits |
| mode | COMBO | latent | latente = blend the encoded photos; pixel = blend the pixels, then encode (very similar, sometimes cleaner) |
| contrast | FLOAT | 1.000.5–1.5 | 1.0 = plain mean. Averaging flattens fine contrast; >1 gives it back (try 1.1-1.2), <1 blurs the blend more |
| image_2opt | IMAGE | — | |
| image_3opt | IMAGE | — | |
| image_4opt | IMAGE | — |
Outputs (5)
| Name | Type | Description |
|---|---|---|
| latent | LATENT | — |
| preview | IMAGE | — |
| width | INT | — |
| height | INT | — |
| report | STRING | — |