Rendering HDR Sequences

This tutorial shows how to render high-dynamic-range sequences suitable for downstream sensor emulation. HDR images store linear intensity values, unlike display-referred formats like PNG/JPEG.

HDR File Formats

HDR outputs can be saved as either .exr (OpenEXR) or .hdr (Radiance HDR):

Format	Extension	Color Depth	Color Mode	Compression
OpenEXR	.exr	16 or 32 bit float	RGB or RGBA	NONE, PXR24, ZIP, PIZ, RLE, ZIPS (lossless); DWAA, DWAB (lossy)
Radiance HDR	.hdr	32 bit float (always)	RGB only	lossy RLE

The default DWAA codec for OpenEXR provides a good balance of compression and quality. For lossless archival use ZIP or PIZ.

Using the Python API

The include_frames method defaults to PNG with 8-bit color depth. Configure it for HDR output as follows:

# OpenEXR with lossy default codec (DWAA) at 32-bit float
client.include_frames(file_format="OPEN_EXR", color_mode="RGB", bit_depth=32)

# OpenEXR with lossless compression
client.include_frames(file_format="OPEN_EXR", color_mode="RGB", bit_depth=32, exr_codec="ZIP")

# Radiance HDR (always 32-bit float, RGB only)
client.include_frames(file_format="HDR", color_mode="RGB")

Composite outputs can also be configured for HDR:

client.include_composites(file_format="OPEN_EXR", color_mode="RGB", bit_depth=16)
client.include_composites(file_format="HDR", color_mode="RGB")

Using the CLI

The same configuration is available through the render-animation command:

visionsim render-animation scene.blend ./output \\
    --frames.file-format OPEN_EXR \\
    --frames.bit-depth 32

visionsim render-animation scene.blend ./output \\
    --frames.file-format HDR \\
    --frames.color-mode RGB

Tip

Some sensor emulation might require HDR inputs. When running the downstream sensor emulation pipeline, ensure the rendered frames are in .hdr or .exr format.