Dimensity 1000 Mediatek unlocks AV1 decoding for YouTube

Google

Mediatek unlocks AV1 decoding for Youtube

Smartphones with Mediateks Dimensity 1000 require less data and energy thanks to AV1.

Soc developer Mediatek has announced that AV1 decoding will be used by their own Dimensity 1000. The chip is the first to date for smartphones, which can accelerate the open video codec in hardware. Mediatek releases the AV1 decoding of the Dimensity 1000 by firmware update for Youtube videos, partners like Oppo at the Reno 3 5G must update their devices accordingly.


Google has been using the AV1 codec in a beta test for Youtube since autumn 2018, but still only in SD quality. Mediatek himself generally speaks of an AV1 acceleration up to 4K resolution with 60 images per second. The aim of the AV1 H.264/H.265 or VP9 is to reduce the required data volume with the same image quality, since the compression is more efficient. This results in less load on the LTE/5G modem and the Wi-Fi block, which is why MediaTek uses AV1 decoding to increase battery life.

In addition to Youtube, Google supports the AV1 codec for the video chat application Duo, and Netflix uses the AV1 codec in its own Android app. The US company uses the free decoder Dav1d, which was developed by the Videolan team. The Aomedia Consortium (Alliance for Open Media) is developing the Videocodec AV1 itself and is expected to be significantly better compressed than in the past and widely distributed thanks to its free patent licence.

Mediatek’s latest premium-class chip for Android smartphones. 

The Dimensity 1000(L), in turn, is Mediatek’s latest premium-class chip for Android smartphones. The Soc integrates four Cortex A77 cores with up to 2.6 Ghz and four Cortex A55 cores with up to 2 Ghz as CPU cores, plus Mali G77-MP9 as graphics unit and a 64-bit wide interface for LPDR4X memory. The Dimensity 1000 also features an NPU (Neural Processing Unit) for artificial intelligence, that is to say inferencing on the device, and a 5G modem. The chip is produced by TSMC in a 7 nm process.