ImageSDRToHDR - ComfyUI Built-in Node Documentation

The Convert SDR to HDR node expands standard dynamic range sRGB images into Rec.2020 HLG HDR.

The Convert SDR to HDR node expands standard dynamic range sRGB images into Rec.2020 HLG HDR. It keeps the existing look of the image and maps its brightness into a wider range, so the result must be saved with an HDR color space selected. Alpha is preserved unchanged.

Inputs

Parameter	Description	Data Type	Required	Range
image	The SDR image(s) in sRGB to expand into HDR.	IMAGE	Yes	-
algorithm	HDR expansion algorithm. Selecting an algorithm reveals its parameters (target_luminance).	DYNAMIC_COMBO	Yes	"lottes"

lottes Inputs

These parameters appear when algorithm is set to "lottes".

Parameter	Description	Data Type	Required	Range
target_luminance	Peak-white target in nits. 0 keeps the calibrated rendering. Dark images are not normalized to this peak. (default: 650.0)	FLOAT	Yes	0.0 to 10000.0 (step 1.0)

Each image of the batch is converted separately: the sRGB encoding is inverted to linear light, the brightness is expanded with a Lottes-style rational curve blended with a calibrated monotonic log-gain curve, and the result is re-encoded as Rec.2020 HLG. Because the output is HDR, the save node must use the HDR color space; saving it as sRGB renders the brightness incorrectly.

HLG assumes a 1000-nit reference display, so target_luminance values below 1000 keep that reference and a value of 0 leaves the calibrated rendering untouched. Single-channel images are expanded to three channels before conversion, and the alpha channel is copied through unchanged.

Outputs

Output Name	Description	Data Type
IMAGE	The Rec.2020 HLG HDR image(s).	IMAGE

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

---

Source fingerprint (SHA-256): 51eb0beac3d48345a30ba66d80d95873384cf5155fc06c33353b6d1dc029658a