Exporting images and movies¶
Fused OME-BigTIFF¶
This reads level zero of <stem>_fused.ome.zarr and
<stem>_max_z_fused.ome.zarr, writing corresponding .ome.tif files beside
them. All timepoints, channels, pixel values, and the source dtype are retained.
Both exports use BigTIFF, 512 x 512 tiles, zlib compression level 8, and
predictor=True. Tiles are streamed without loading the full volume.
Progress reports timepoints, channels, and Z planes per channel separately.
OME-XML records dimensions, voxel spacing, registered positions, and available channel names, colors, wavelengths, exposures, timing, acquisition date, and instrument metadata. Acquisition settings, source XML, NGFF metadata, and the processing journal are embedded as annotations when available. Keep the source acquisition and journal accessible to retain their metadata.
--metadata-only repairs the XML encoding of existing exports without rewriting
or recompressing pixels. See export-ome-tiff.
Annotated projection TIFFs¶
Supply the directory containing one full deconvolved, deskewed store. Each timepoint, channel, and position produces a full-resolution 8-bit display TIFF: XY at upper left, XZ below, and YZ to the right. Panels share a physical display scale and include labels, a scale bar, and an acquisition timestamp.
Contrast remains fixed to timepoint 0 for each position/channel. Its 0.001st and
99.999th intensity percentiles define the limits, with a min/max fallback when
they coincide. Timestamps use acquired scan-plane count times summed channel
exposures, excluding moves and other acquisition overhead. The raw acquisition
is located through the journal or supplied with --acquisition.
Depth coloring uses the brightest voxel's depth: Z for XY, Y for XZ, and X for
YZ. Exact ties select the first voxel. The turbo color map is the default;
--depth-colormap changes it. Depth legends stay fixed across time.
TIFFs are losslessly compressed display renderings, including their annotations.
Use the OME-Zarr or fused OME-BigTIFF for numerical image analysis. Output is
under projection_frames/<dataset>/p000/c000/, with a depth_color subdirectory
for colored images. Re-exporting replaces matching TIFFs.
Movies¶
The zero-based stop is exclusive. This exports and encodes timepoints 100 through 199 using their original acquisition timestamps. Movie encoding follows TIFF export; existing TIFFs outside the selected range remain on disk.
Encode an existing dataset's TIFF sequences instead:
Movies use NVIDIA NVENC, defaulting to H.264 and 8 Mbps constant bitrate.
Playback defaults to the recorded acquired volume rate, including fractional
rates. --fps overrides playback without changing burned-in timestamps.
Each TIFF supplies one frame in timepoint order. Existing-frame encoding rejects
missing or duplicate indices and skips legacy downsample_* directories.
Hardware encoding requires an NVENC-capable NVIDIA GPU and compatible driver. MP4s are lossy presentation copies. See export-projections and encode-projections for options.
Raw timelapses¶
convert-timelapse exports the time series at selected raw positions and scan planes. Raw timelapse TIFFs retain camera counts; optional time-mean projections apply camera calibration. This command does not export a registered fused volume.