Chip technology company Arm announced a set of multimedia processors intended to deliver “compelling visual experiences to the mainstream”, while also detailing why its chosen path to artificial intelligence (AI) is different to some rivals.

The company unveiled its Mali Multimedia Suite of Video, Display and Graphics technology, which includes a video processor enabling 4K content at 60fps or 120fps; a display processor delivering “more complex visual experiences”; and graphics processors which delivers “premium experiences” to digital TV, mainstream and entry-level mobile.

“Whether they realise it or not, consumers are demanding more from today’s mainstream and low-cost devices and support for a rich, multi-layered user interface and a broad range of the latest applications and technologies is an absolute must,” Ian Smythe, senior director in the client line of business at Arm, wrote.

In a statement, Arm also said machine learning is no longer solely the domain of premium smartphones: its Mali-G52 graphics processing unit (GPU) offers 3.6-times the machine learning performance of its previous-generation product, ensuring next-generation use cases are supported across all classes of device.

The company said it is “not always practical to dedicate silicon and engineering effort” to machine learning in mainstream devices, noting “it requires every component of the system-on-chip to be used to its best advantage on a workload-by-workload basis”, across the CPU and GPU.

This argument had been used before by Arm-partner Qualcomm, which said “dedicated hardware requires a very deliberate choice and if you don’t make that choice correctly, you’ve wasted silicon.”

In contrast, Huawei launched its Kirin 970 processor with much fanfare in 2017, which features a dedicated “neural processing unit”. The chip is already being used in smartphones from the company.

Apple’s A11 Bionic processor, as used in iPhone 8 and iPhone X, includes a dual-core “neural engine” designed for specific machine-learning algorithms.