Blog: The Challenges of Engineering Audio for Self-Driving Cars

Most of the recent talk in the automotive industry – and in the tech world, urban planning and the energy industry, too – has been about the impact that future self-driving cars will have on the way we live and get around town. But while there’s been plenty of debate about how safe these cars will be, and how soon they’ll come to market, there’s been little discussion of what these cars will be like on the inside – or how their entertainment and telecommunications systems may differ compared with the systems in today’s cars.

Having participated in the automotive infotainment arena for 30 years, most recently as Sr. Manager of Wireless Engineering for Tesla, DSP Concepts COO John Whitecar has a great deal of insight into what challenges and opportunities the self-driving car presents for automotive audio engineering.

“The concept we have now of everyone facing forward may change,” he says. “With no one driving the car, it won’t be necessary for anyone to face forward, or even sit toward the front. If you have one or two people in the car, they might sit in what we now think of as the back of the car. You might have club-style seating, where people are facing each other. This is very different from today, where you know there’s always going to be someone in the driver’s seat facing forward, and if they have one passenger, that passenger will probably be in the right-side front seat.

“In autonomous vehicles, because there’s no driver, everyone could watch video,” he continues. “But the screens could be anywhere in the car, even where the windshield would normally be.”

Knowing where a car’s occupants sit makes the task of designing today’s automotive audio systems much easier. But with autonomous vehicles, the sound system will have to adapt to the passengers’ habits, which will require new audio processing tools.

“You’re going to need to be able to pan sound to any location in the vehicle,” Whitecar says. “Of course, some of the passengers may want to access different audio programming – someone might be listening to the radio or music from their smartphone, while someone else is on a phone call. You can easily imagine that every passenger will want to maintain their own connection with the outside world, especially in for-hire vehicles that might be picking up multiple passengers and servicing multiple destinations.”

According to Whitecar, the solution is, “You want to create environments where these conversations and entertainment options don’t interfere with each other.” Achieving this goal will require flexible advanced audio processing that can handle dozens of channels of sound, routing the desired audio to each passenger. It will also require advanced techniques to manipulate the amplitude and phase of multiple speaker drivers to create localized sound fields enabling one passenger to hear clearly while the other passengers barely notice – and also to isolate that user’s voice for voice commands, even in a noisy environment.

DSP Concepts’ Audio Weaver Automotive algorithms greatly simplify these tasks for automotive audio engineers, because Audio Weaver’s modular, graphical audio processing environment makes it easy to assemble and experiment with complex processing chains in a matter of minutes, all with no coding necessary. Audio Weaver now includes more than 400 processing modules, to meet practically any need in automotive audio and telecommunications, and custom modules can also be created. The algorithms are already optimized for the DSP or SoC they’re running on, so there’s no need for debugging or laborious optimization of code after the programming is done. 

Audio Weaver Automotive can handle as many audio channels as the DSP or SoC’s processing power will allow, and it also incorporates voice input algorithms, which DSP Concepts has refined through its extensive work on microphone array and smartspeaker platforms and reference designs.

“With Audio Weaver, you have total freedom to design whatever soundstage you want and reconfigure it as the vehicle environment changes,” Whitecar says. “This is exactly what automotive engineers need to meet the challenges of autonomous vehicles, and it’s already available now.”