JLab: Cinematographic Data Science

Overview

This repo intends to make getting started with cinematographic data science (CDS) easy, by maintaining a working toolchain of useful programs.

CDS is a catch-all term for analyzing, generating, and transforming cinematic data. This includes but is not limited to:

1. color science
2. image formats, compression, and encoding
3. high dynamic range

Architecture

We maintain a Docker image with a large toolset of useful programs, including:

1. OpenColorIO
2. OpenImageIO
3. OpenEXR
4. ffmpeg
5. OpenCV
6. Python:
   1. numpy
   2. scipy
   3. colour-science
   4. openexr

Goals

We want the following to be easy for an end-user to use:

Image Analysis

• vectorscope, histogram, parade
• false color
• color checker detection and registration

Tmage Generation

• test patterns such as BT.2111
• color checkers
• save images with embedded ICC profiles

Image Transforms

Precise control over:

• color primaries, white-point, transform function
• bit-depth
• color Channels e.g. (RGB, YCbCr)
• chroma Sub-Sampling (e.g. 4:4:4 vs. 4:2:0)

Local Install

These directions are tailored towards a general audience.

Pre-Requisites

Install the following applications onto your computer.

1. Docker
2. Visual Studio Code
3. GitHub Desktop or another Git Client.

JLab

1. Use GitHub Desktop to download JakeGWater/JLab.
2. Open the project using Visual Studio Code.
3. Create a new terminal, and type ./scripts/start to start the server.
4. Navigate to http://localhost:8888

Development

1. To build a local Docker image:

   1. Download ~500mb of OCIO data: type make download-ocio

   2. Type make build

2. To delete all temporary data, including the docker image: type make clean

The start script will use the jlab docker image by default.