Oppo announced the Oppo MariSilicon X in its Inno Day event in China a few days ago. The MariSilicon X comes with Oppo’s custom NPU, ISP, and Memory Subsystem. It is said to improve the photos and videos quality as well as give the phone better performance during AI-related tasks. The first phones to feature the Oppo MariSilicon X will be the next-generation Find X series promising “the most advanced and powerful imaging experience ever”. It is due to be announced in Q1 2022.
Continue reading to find out everything you need to know about the Oppo MariSilicon X. Read about the Oppo Find N, Oppo’s first-ever foldable.
Oppo MariSilicon X Features
The Oppo MariSilicon X is built with three major parts; the Image Signal Processor, the Neural Processing Unit, and the Memory Subsystem. All of these parts are said to help improve the overall performance of phones that have it, in the photo, video, and AI tasks.
The Oppo MariSilicon X is a chipset on its own and it is built on a 6nm fabrication process. This is kind of non-energy efficient as the Qualcomm Snapdragon 8 Gen 1 is built on a 4nm fabrication process and comes with all the parts embedded.
First, let’s talk about the ISP (Image Signal Processor) that is embedded in the MariSilicon X as Oppo won’t be relying on the ISP in the Snapdragon 8 Gen 1. The ISP has support for 20-bit which means it’ll be able to get more data out of the image sensor. The ISP also supports Ultra HDR in videos, 120db dynamic range in images, and 4K night videos. It gives you support for 20-bit RAW images and RGBW PRO so you can do some nice post-processing to the photos you take. The IPS works together with the NPU so the AI can apply different algorithms for noise reduction, better color reproduction, HDR, improved detail, dynamic range, etc.
The ISP comes with RGBW Pro mode which allows you to take the full potential of Oppo’s RGBW sensor as it supports the separation of the white channel from the RGB.
Now, let’s talk about the NPU (Neural Processing Unit) which is can perform up to 18 TOPS. This makes it great for crunching through AI-related tasks. Whilst being powerful, the NPU is also power efficient.
The Memory Subsystem comes into play now as it is paired with the NPU which means the NPU would no longer rely on the system memory. This helps to reduce processing time and also increase power efficiency. This Memory Subsystem comes with dedicated memory capable of 8.5 GB/s data transfers.
