
Fully integrated into DSP Concepts’ Audio Weaver, GPU Audio’s engine runs Dirac’s Unison convolution on the GPU integrated into Pioneer’s automotive audio amplifier platform — a drop-in module inside the Audio Weaver graph that unlocks substantial additional audio processing at no extra hardware cost. The solution will be demonstrated at the 2026 AES Automotive Audio International Conference in Detroit.
Detroit, July 27, 2026 - GPU Audio, Pioneer and Dirac today announced a joint demonstration showing premium in-car sound experiences running in real-time on the GPU (Graphics Processing Unit) already built into Pioneer’s automotive hardware. The work is fully integrated into DSP Concepts’ Audio Weaver, the audio-development environment that automakers and Tier 1 suppliers already build on, and the breakthrough will be shown at the AES Automotive Audio convention in Detroit at the end of the month.
The demonstration targets the most computationally demanding stage of Dirac’s Unison tuning - the FIR filtering that corrects the cabin’s acoustics and shapes the sound for every loudspeaker and seat in the car. Running on Pioneer’s production-ready amplifier platform powered by Texas Instruments’ AM62 SoC, GPU Audio’s processing engine performs this filtering on the chip’s onboard GPU in real-time and at full 32-bit floating-point precision, with headroom to spare and output that matches Dirac’s own reference within the limits of floating-point computational noise.
The significance is simple. In today’s car amplifiers the GPU on these chips sits almost completely idle. By moving audio processing onto it, GPU Audio turns unused silicon into vast parallel audio compute, unlocking substantial additional processing power at no extra hardware cost (estimated to be around 35% extra compute on standard DSP chips) and the journey has just begun.
“The GPUs inside automotive chips have been sitting idle for years. GPU Audio harnesses the immense power and delivers massive audio compute - with the same premium tuning, running in real-time, at no additional hardware cost.”
- Alexander Talashov, CEO, GPU Audio
"Every car already has more audio compute sitting idle than most people realize. Unlocking that 'free' compute lets us actively treat the cabin as a listening space—running Dirac Unison to deliver a premium in-car sound experience, treating the acoustic space with the same care we give source and system, for every seat, at no extra hardware cost."
- Anders Storm, CEO, Dirac
“Using the GPU enables us to run much more complex audio algorithms that demand 50 GFLOPS of processing power and more, enabling new listening experiences and applications such as individual sound zones, real-time head-tracking or venue emulation.”
- Sebastian Wendt, Chief Global System Architect, Pioneer
Crucially, this is not a research prototype. GPU Audio’s engine is fully integrated into DSP Concepts’ Audio Weaver: inside the Audio Weaver graph, the GPU-accelerated block is a drop-in replacement for the standard convolution block. The same FIR filters run on the in-car GPU instead of using dedicated DSP - there’s no change to the rest of the signal chain, other than substantially more compute power.
For automakers and Tier 1 suppliers who already use Audio Weaver, that makes adoption a graph-level change rather than a re-engineering effort - a ready-to-integrate solution to take the compute-intensive work handled today by dedicated DSPs and move it onto silicon that would otherwise be unutilized.
The AES Automotive Audio demonstration is a first step, ready to demo in-person at the show. Building on GPU-accelerated FIR filtering, the partners are mapping a path to move progressively more of Dirac’s sound-optimization pipeline onto the same idle GPU. Because the GPU offers a large pool of untapped parallel compute, this opens the door to richer immersive audio, more simultaneous channels and private listening zones - with ultra-fine tuning and lower latency - without additional hardware.
The same “offload” approach lets automakers scale one premium audio platform across vehicle tiers: delivering flagship sound on existing systems, freeing up DSP resources, and keeping hardware costs down while raising performance. Longer term, the partners see the GPU becoming a standard audio accelerator in automotive amplifiers, with the potential to host complete tuning pipelines that today’s DSP-only platforms cannot fit.
GPU Audio is the world’s first company to unlock your GPU to power premium immersive audio in a range of industries
Pioneer is a global technology company in automotive electronics, with decades of expertise in automotive audio systems, loudspeakers, infotainment platforms, and human-machine interface technologies.
Dirac technology improves in-car sound through advanced signal processing, precise speaker optimization and accurate cabin correction.
DSP Concepts is the creator of Audio Weaver, the real-time audio infrastructure for Intelligent Systems.
GPU Audio - [Sergey Korobov, sergey@gpu.audio]
Pioneer, Corporate Communications Department - [pioneer_prd@post.pioneer.co.jp]
Dirac - [Nareh Rostamian, nareh.rostamian@dirac.com]