Skip to content

ChangeLog.md

Latest commit

 

History

History
95 lines (82 loc) · 5.37 KB

ChangeLog.md

File metadata and controls

95 lines (82 loc) · 5.37 KB

v0.5.6 ?

  • Bug fix: incorrect assert in dm_preview_metric.py
  • Fix: disabled some unnecessary warning messages when dm-preview was selected
  • Fix: YCbCr to RGB transform changed from BT.601 to BT.709 (issue #36)
  • Bug fix: --dump-channels works again (was broken after adding the batch mode)
  • Fix: --device=mps works again (on Apple Silicon)
  • Added: Progress bar
  • Added: --debug will display the full stack trace when error is encountered. Otherwise, the stack trace should be hidden.
  • Added: --count-frames will use slow but accurate method to ensure that we know the exact number of frames in the stream.
  • Added: --temp-padding=symmetric is now implemented.
  • Fixed: reduced frame delay with --temp-resample
  • Fixed: support for reading Yuv files
  • Added: 422 chroma subsampling is now natively supported
  • Added: new display models in the display_models.json
  • Fixed: pu_psnr_y was compared linear instead of PU21 values (thanks to jgerster0)

v0.5.5 (31/10/2025)

  • Bug fix: Distogram export works again (it was broken after introducing the batch mode)
  • Fixed: A more meaningful message when a metric does not generate a histogram or heatmap but it is asked to generate such
  • Fixed: --temp-resample accepts an optional parameter - the maximum frame rate to use for temporal resampling

v0.5.4 (24/09/2025)

  • Bug fix: Caching was causing a huge mem-leak

v0.5.3 (22/09/2025)

  • Added: Batch mode for faster processing
  • Bug fix: small discrepancy in the predictions depending on how many frames are processed at once (gpu memory) (thanks to Yancheng)
  • Removed dependecy on the depreciated pynvml package
  • Minor changes to avoid warnings in Python 3.13

v0.5.1 (20/07/2025)

  • Fixed missing files in the PIP release

v0.5.0 (20/07/2025)

  • Added two new metrics: cvvdp-ml-saliency and cvvdp-ml-transformer. See metrics.md for more info.
  • Fixed: Heat map visualization when the test image has very small dynamic range (thanks to Yuta).
  • Fixed: More accurate color transform is used when dumping channels (thanks to Simon)
  • Added documemntation for --dump-channels.

v0.4.3 (30/05/2025)

  • Interpolation of the CSF is a bit faster now (thanks to Dongyeon)
  • Fixed: Updated broken calibration/*.py scripts. Now they use --config-paths instead of discontinued --config-dir.
  • Fixed: Base bands were incorrectly weighted when generating heat maps.
  • Fixed: Improved Matlab interface - handles negative JODs and video heatmaps (thanks to CaptainS5)
  • Now can handle videos that do not contain the number of frames in the meta-data
  • Added --temp-resample option that allows to compare videos of different frame rates
  • Fixed: The default device is now 'cuda' instead of 'cuda:0' (thanks to soCzech)

v0.4.2 (29/09/2024)

  • Files are now sorted after the wildcard expansion
  • Updated PU21 encoding parameters so that they are in sync with those in https://github.com/gfxdisp/pu21/
  • Added: --dump-channels for generating videos with intermediate processing stages (debugging and visualization)
  • Added: Support for HLG EOTF (e.g. iPhone HDR video) - thanks to Cosmin Stejerean
  • Added: Processing of videos stored as image frames, described using the C-notation frame_%04d.png. New arguments: '--fps' and '--frames'
  • Fixed: A better memory model for estimating how many frames can be processed at once on a GPU. Added '--gpu-mem' argument.
  • Added: 'exposure' field in a display model JSON file.
  • Added: --result argument to store results in a CSV file.
  • Fixed: Added examples to README.md and improved documentation.
  • Added: ColorVideoVDP logo.

v0.4.1 (27/04/2024)

  • Added --metric dm-preview for debugging of the display model
  • Added a loss method to cvvdp and examples/ex_adaptive_chroma_subsampling.py
  • Added "pixels_per_degree" as an optional field in display_models.json
  • Platform is printed when running with --verbose (for reporting bugs/issues)

v0.4 (19/01/2024)

  • A new calibration with a small improvement in performance
  • Improved predictions for supra-threshold contrast across color directions
  • Improved masking model
  • Added Matlab wrapper
  • CLI now has --interactive mode to process multiple images/video without restarting PyTorch
  • Fixed distogram generation for images
  • "--quiet" flag now ensures that no warning messages end up in the stdout
  • CSF updated to the latest castleCSF fit
  • added 'luminance' color space to handle luminance-only data

v0.3 (05/09/2023)

  • Added value range checks when the metric is running on HDR data (to avoid passing relative values)
  • Added SSIM as an alternative metric
  • Better handling of paths to configuration files

v0.2

  • Updated ColorVideoVDP model with cross-channel masking and more advanced pooling, different calibration and better prediction accuracy.
  • Added distograms
  • Changed handling of paths to configuration files

v0.1.1

  • Heatmaps are now saved using ffmpeg's mpeg4 codec for better comparibility across platforms
  • Fixed freezing on Windows when reading long videos (due to bug in python's /dev/null implementation)
  • Added plain psnr_rgb metric, which operates on display-encoded values (or PU21-encoded if needed)
  • Updated PU-PSNR-* to use 100 nit as the peak + to computed PSNR for all pixels in the video (rather than mean PSNR over all frames).
  • Reorganized display_models.json - now color space is a part of the display model json spec, EOTF is a part of color space json spec.

v0.1.0 Initial beta release (27 March 2023)