Working with ARRIRAW
ARRIRAW Processing Pipeline
ARRI built long-term relationships with postproduction equipment manufacturers through the ARRI Partner Program. These relationships have facilitated all of the leading compositing and color correction tools to be enabled with direct processing of the ARRIRAW file format. ARRI offers a software development kit (SDK) for ARRIRAW processing, containing documentation of the ARRIRAW processing pipeline, and a library – that is, executable code – that the vendor can incorporate into their application. ARRI also supports vendors who wish to implement the ARRIRAW processing procedure on their own, through comprehensive documentation of a three-phase color processing pipeline and continuous direct support.
The first phase of ARRIRAW processing is the most compute-intensive. ARRIRAW images (like all 'camera raw' images) have only one channel. A color reconstruction algorithm calculates the missing components of each pixel based on the type and position of the array of colored filters on the camera sensor. The ALEXA uses the Bayer pattern color filter array. The term 'color reconstruction' therefore is also known as 'debayering'. The Bayer pattern filters the light hitting the sensor so that 50% of the sensors photosites are used to represent green, 25% of the photosites represent red and the remaining 25% represent blue.
The image above shows a single-channel capture from the sensor on the left. At center is the color filter array pattern used by the ALEXA. At right is the reconstructed image. Half of the reconstructed image's green values are interpolated from the surrounding photosites rather than captured, as are three-quarters of the red and three-quarters of the blue.
The output quality of the image depends on the debayering algorithm. Generally speaking, a simpler algorithm will process faster, but also involve a higher probability for color errors. Color errors are also more likely as the output resolution gets closer to the resolution of the ARRIRAW image. The ALEXA sensor provides a horizontal resolution of about 3K pixels, from which downsampling routines produce optimized HD and 2K resolution images. For VFX, however, the images often are processed using the native sensor pixel count and then downscaled to 2K, for example, at a later stage. Using this approach takes advantage of the luminance resolution, which correlates to the sensor pixel count.
ARRIRAW SDK and 3rd Party Implementations
In many cases, ARRI's SDK is fast enough and adaptable enough to satisfy the processing needs of an application. But in some cases, especially when the product uses custom hardware (or standard hardware in a nonstandard way), partner program members may want to implement the ARRIRAW processing pipeline themselves. In this scenario, the SDK serves as a reference against which the partner program member's developers can test their results before submitting evaluation imagery to ARRI's workflow group. In a few applications, the vendor's product offers both processing solutions, giving the user the option of maximum throughput with a very good match to the SDK, or bit-for-bit matching of other products that use the reference SDK implementation.
Variances and Tolerances
When the ARRIRAW Processing pipeline has been implemented into a product, a company can submit the image processing results to ARRI. Our engineers then evaluate the images based on a standardized set of tests, which must produce completely satisfactory results, so we will issue the "Certified for ARRIRAW Processing" label for this product.
ARRI allows partner program members some variation in the implementation of the first (debayering) phase of the processing chain. There are no compromises allowed, however, during the second and thrid phases of the processing chain where LogC or video encoding is completed. The reason for the flexibility in debayering is that some platform vendors traditionally produce a 'softer' image as a result of their debayering process. These platforms might choose to preform a sharpening operation after the conversion from ARRIRAW to the desired working format. ARRI will certify a product’s default settings, even when producing debayered images that are less sharp than the SDK reference, under one condition: the product must provide settings that render the converted image to match the SDK reference, free from artifacts and matching in geometry, tonal balance, color rendition, bit depth and metadata.
The ARRIRAW Converter application provides a graphical user interface for ARRI’s reference SDK. In addition to the GUI ARRIRAW converter, a command-line tool for processing images or image sequences is available for Mac OS X, Windows, and the RedHat, SUSE and Centos distributions of Linux.
You can download your copy of ARC in the ARRIRAW workflow downloads section.
Testing Your Workflow
ARRI recommends running tests to define:
- How the intended look can be achieved using the chosen toolset(s).
- The different tools that need to be set up to meet the production's expectations of consistent results.
- The settings required for a chosen tool to most closely match the SDK reference results, if no other reference set of results are available.
You can use an ARRIRAW frame grabbed with an ALEXA as test material. ARRI also provides ARRIRAW test footage containing useful test imagery. Download instructions can be found in the general downloads section.