Qualcomm presents the Snapdragon Sound Elite Gen 2: 30% smaller, 40% less consumption and double the AI performance
Qualcomm has announced the Snapdragon Sound Elite Gen 2 SoC, a chip intended for Wireless headphones and speakers with built-in AI acceleration. It replaces the previous Snapdragon Sound S7, a chip that it improves with a smaller size, less consumption, improvements in connectivity and the integration of an NPU for AI.
Now, the Snapdragon Sound Elite Gen 2 is integrated into a chip that occupies only an area of 31.18 mm², which represents a 30% reduction compared to 44.22 mm² of the S7. This allows its integration into smaller headphones, or those of equal size, but leaving room for other elements such as larger capacity batteries.
Besides, Its consumption has also been reduced, dropping by 40% compared to his predecessor. According to Qualcomm, the new chip has a consumption of 0.6 mA at rest with an active WiFi connection, while the S7 consumed 1 mA.
Double the AI performance of the last generation
In an era where it seems that everything must involve AI, Qualcomm has followed the trend by integrating an eNPU, an integrated NPU to locally process certain modest AI tasks. This module has a speed of 250 MHz with which it achieves 128 GOPs (giga operations per second) and 256 MACs (multiplication-accumulation) for each clock cycle. This is double the performance of the previous model at the same frequency.
The new Snapdragon Sound Elite Gen 2 now supports “Micro-Power WiFi 6E”a technology with which you can connect directly to the internet to play music or use AI systems in the cloud, for example. However, it continues to maintain support for Bluetooth 7.0since it is the connectivity with which most headphones and speakers connect to the mobile phone and other devices.
XPAN (Expanded Personal Area Network) technology allows you to increase the bandwidth and signal range.
For the audio processing system itself, Qualcomm’s Kymera architecture is used in three DSPs, plus an additional one dedicated to ambient sound amplification, operating with “ultra-low” latency and also to noise cancellation (5th generation ANC).
It supports lossless audio decoding using aptX Lossless up to 24-bit and 96 kHz, and can handle multiple microphone input via cVc algorithms and Audio Sense hardware.
The Snapdragon Sound Elite Gen 2 platform has already been introduced to device manufacturers, and brands like Cleer already have concept prototypes. However, there are still no estimated dates for its arrival on the market.
