← All models
Convert standard videos to HDR by recovering lost detail in bright highlights and dark shadows—smooth, stable results with room for re-exposure, guided by text or reference images.
images
Details
- Accepts
- text
Pricing
- Price
- 1.25 cr / megapixel
Prices in credits (1 credit = $0.01).
Data schema
Input
| Field | Type | Description |
|---|---|---|
| seed | — | Random seed for reproducible results. |
| grid_r | integer | Grid overlap stride for sliding window inference. Lower values produce smoother results but are slower. |
| long_exposure | integer | Exposure value (EV) for the long exposure image. Common values: 1, 2, 3. |
| short_exposure | integer | Exposure value (EV) for the short exposure image. Common values: -3, -2, -1. |
| medium_exposure | integer | Exposure value (EV) for the medium exposure image. Typically 0. |
| num_inference_steps | integer | Number of DDIM sampling timesteps. Lower is faster, higher may produce better quality. |
| long_exposure_image_urlrequired | string | URL of the long exposure (overexposed) LDR image. |
| short_exposure_image_urlrequired | string | URL of the short exposure (underexposed) LDR image. |
| medium_exposure_image_urlrequired | string | URL of the medium exposure LDR image. |
Output
| Field | Type | Description |
|---|---|---|
| seed | integer | The seed used for generation. |
| image | — | The reconstructed HDR image (tonemapped to LDR for display). |