Spatio-Temporal Noise Sequences: Multipurposed Pseudo-Random Visual Test Signals and Programmatic Content Examples

Author: Florian Friedrich | FF.de

General Description

This repository hosts a range of spatio-temporal noise sequences designed for use in various technology evaluations in the video field, including display metrology, workflow testing, encoding stress tests, display warm-up, and algorithm design. In August 2024, this repository has also been extended to serve as a resource for content described in the paper "Defining and Characterizing Programmatic Image Sequences for Multi-Disciplinary Applications", intended to be used for multiple purposes such as implementation into HDR video generators, HDR DUT Stabilisation, Power Consumption and more (see paper, referenced below).

Features (Spatio-Temporal Noise sequences)

• Format: TIFF (16Bit half float, LZW compressed), MOV (H.265 lossless 12Bit)
• Resolution: 3840x3840 pixels, lower resolutions by utilizing cropping, higher resolutions by tiling)
• Aspect Ratio: Agnostic
• Framerate: Agnostic
• Colorspace: Agnostic
• File Size: Up to 50GB per sequence

Data Download

The full dataset is too large to be hosted directly on GitHub. You can download the sequences from Zenodo as described below.

Directory Structure

This repository is organized as follows:

• Example_Implementations/: Folder planned for containing sample code and use-case scenarios (currently empty; see Planned_Implementations.md).
• LUTs_and_Transformations/: Folder planned for containing Lookup Tables and transformation files (currently empty; see Planned_LUTs_and_Transformations.md).

Media Directories (Hosted on Zenodo)

The following directories contain a ZENODO_LINK.md file with the DOI link for downloading the actual media sequences from Zenodo:

• STNOISE_alpha1_MOV_sequence/: MOV file for alpha value 1.
• STNOISE_alpha1_TIFF_sequence/: TIFF files for alpha value 1.
• STNOISE_alpha2_MOV_sequence/: MOV file for alpha value 2.
• STNOISE_alpha2_TIFF_sequence/: TIFF files for alpha value 2.
• STNOISE_alpha3_MOV_sequence/: MOV file for alpha value 3.
• STNOISE_alpha3_TIFF_sequence/: TIFF files for alpha value 3.
• STNOISE_alpha4_MOV_sequence/: MOV file for alpha value 4.
• STNOISE_alpha4_TIFF_sequence/: TIFF files for alpha value 4.
• STNOISE_alpha5_MOV_sequence/: MOV file for alpha value 5.
• STNOISE_alpha6_TIFF_sequence/: TIFF files for alpha value 6.
• STNOISE_alpha7_MOV_sequence/: MOV file for alpha value 7.
• STNOISE_alpha7_TIFF_sequence/: TIFF files for alpha value 7.
• Example_compressed_IDMS_HDR_Test_Signal: Compressed examples (.mp4) of programmatic HDR content, including Python script to combine it into longer sequences.
• Programmatic_content_stats_examples: Examples of image statistics for the sequences provided as SVGs.
• Programmatic_content_HEVC_12Bit_lossless: Programmatic HDR content examples, as .mov files with 4:4:4 12 Bit lossless HEVC encoding.
• Programmatic_content_TIFF_sequences: Programmatic HDR content examples, as .tiff sequences (16 Bit half-float, full range).

See ZENODO_LINK.md in each media directory for the corresponding download link.

Usage Variabilities (Spatio-Temporal Noise sequences)

You can modify these sequences in numerous ways, including:

• Mapping, swapping, or flipping the R, G, B channels.
• Changing the seed order or applying different blending methods.
• Cropping, tiling, or scaling the images.
• Converting to different color gamuts through computational transformations or LUTs.

Analysis and Fine-Tuning (Spatio-Temporal Noise sequences)

For specific project goals, seeds and sequences can be selected, excluded, or combined for desired spatio-temporal distribution. Filters in the frequency domain can be used for additional customization.

References

1. Kunkel, T., & Friedrich, F. (2022). Utilizing advanced spatio‐temporal backgrounds with dynamic test signals for high dynamic range display metrology. Special Section Paper. DOI: 10.1002/sdtp.15469 (Distinguished Paper), DOI: 10.1002/jsid.1125 (Full Paper).

2. Kunkel, T., & Daly, S. (2020). 57-1: Spatiotemporal noise targets inspired by natural imagery statistics. SID Symposium Digest of Technical Papers. DOI: 10.1002/sdtp.14001.

3. Field, D. J. (1987). Relations between the statistics of natural images and the response properties of cortical cells. Journal of the Optical Society of America A. DOI: 10.1364/JOSAA.4.002379.

4. Webster, M., & Mollon, J. (1997). Adaptation and the color statistics of natural images. Vision Research. DOI: 10.1016/S0042-6989(97)00125-9.

5. Lennon, J. J. (2000). Red-shifts and red herrings in geographical ecology. Ecography. DOI: 10.1111/j.1600-0587.2000.tb00265.x.

6. Friedrich, F., & Kunkel, T. (2024). Defining and Characterizing Programmatic Image Sequences for Multi-Disciplinary Applications SID, Color and HDR Metrology. DOI: 10.1002/sdtp.17727 (Invited Paper)

License

This project is released under the Creative Commons Attribution 4.0 International License. For attribution, mention "Spatio-Temporal Noise Sequences: Multipurposed Pseudo-Random Visual Test Signals and Programmatic Content Examples by Florian Friedrich | FF.de".

Author Information

• ORCID: 0009-0007-2868-8432