Load EXR Sequence
Scans every frame of an EXR sequence and outputs an ordered list of part descriptors, without reading pixel data.
Category: EXR
TL;DR
- Takes a
Sequence (from Scan Sequence or Scan Split Sequence) and outputs one EXRPartArtifact per frame, ordered by frame number.
- Connects directly into a ForEach Start loop for per-frame work such as Display EXR Part.
- Header-only: always reads just the frame headers, whatever the
openexr.header_only setting says. Downstream nodes read pixels themselves.
- Reads the same part (
part_index) from every frame.
Typical workflow position
Scan Sequence → [Load EXR Sequence] → ForEach Start → Display EXR Part → ForEach End
Node preview
| Name |
Type |
Required |
Notes |
sequence |
Sequence |
Yes |
EXR frame sequence from Scan Sequence. |
Parameters
| Name |
Type |
Default |
Notes |
part_index |
int |
0 |
Zero-based EXR part index to scan in every frame. |
on_mismatch |
skip \| fail |
skip |
What to do when a frame fails to scan or doesn't match the sequence's resolution and channel layout. skip drops that frame; fail fails the whole run. |
Advanced (collapsed by default)
| Name |
Type |
Default |
Notes |
threads |
int |
4 |
Number of concurrent header-scan workers (1–16). |
Outputs
| Name |
Type |
Notes |
parts |
list[EXRPartArtifact] |
One descriptor per scanned frame, ordered by frame number. May be shorter than frame_count with skip. |
frame_count |
int |
Number of frames in the sequence. |
Tips & pitfalls
- List position isn't the frame number. Each artifact's
frame_number field carries the real sequence frame number; with skip, dropped frames leave no gap in the list.
- The most common layout wins. Frames are checked against the resolution and channel layout shared by most of the successfully scanned frames, so an outlier such as a slate frame is the one dropped (or, with
fail, the one reported). If no layout covers more than half the scanned frames, the node fails. Frames that fail to scan don't count toward that majority.
- Single part per run. For multi-part frames, add one node per
part_index you need.
See also