With iOS 27, Apple introduces a new version of the system-level RAW image processing engine. It uses machine learning to significantly improve detail and reduce noise when reprocessing old RAW photos. Here are the details.
iOS 27 includes RAW 9
If you’re not familiar with RAW, it’s basically an image format that preserves data captured directly by the camera’s sensor, giving photographers more flexibility when editing elements like exposure, color, and white balance.
Apple has its own system-level pipeline for processing RAW files from third-party cameras that are exposed to apps via Core Image. It now includes support for nearly 800 camera models and camera-specific calibration, with a complete and regularly updated compatibility list. here.
Over the years, Apple has updated the RAW processing algorithm eight times, improving the way it handles sensor data and adjustments such as demosaicing, denoise and white balance, exposure, color and tone.
With iOS 27, Apple is introducing RAW 9, which the company says is its “biggest update yet.” Here’s David Hayward, Apple’s principal imaging engineer WWDC26 session Improve RAW image processing with Core Image:
(RAW 9) Dramatically improves the rendering of RAW files. It is built on top of a tiled CoreML model that combines demosaicing with denoise for the best quality. The model is run on the device using Apple Neural Engine cores for optimal performance.
In the session, Hayward shows several examples of RAW 9, comparing the results to RAW 8 and sometimes the original, unprocessed sensor data:
This is an enlarged crop of a low noise image using RAW 8. The rendering of this Sony Alpha 7 II vintage dial actually looks pretty good. However, when you examine the same image under RAW 9, the image is sharper, clearer, and fine text is easier to read.
The differences are even more dramatic when you look at high-noise images. First, observe the actual RAW data contained in this very noisy ISO 51,200 image. In this example from the Canon 5D Mark III, the image is a 10x crop of a box of pencils. There is so much luma and chroma noise in the RAW data that it is impossible to distinguish the unique color of each pencil. Here is the result using our previous algorithms! RAW 8 did a reasonable job of reproducing the true colors in the scene. But if you examine the results under RAW 9, the result is significantly better. Colors are accurate and well defined. This is what the bright stamped highlights on the pencils look like.
This last example is a crop of an embroidery thread photo taken with the Fujifilm X-T5 at ISO 12,800. This camera has an unconventional sensor design, which makes demos a challenge. RAW 8 results have some color artifacts and loss of detail in thread. However, if you observe the same image under RAW 9, the results are noticeably better. Smaller text is more legible and the texture on the thread is clearer.
The developer session details how to enable RAW 9, optimize performance for editing and exporting, and more.
To learn more about RAW 9 and other Core Photo improvements coming with iOS 27, watch the full session here.
It’s worth checking out on Amazon
FTC: We use automatic affiliate links that generate income. More.
