Google explains Pixel sound design, like using gen AI for Pixel 8 ‘Gems’ collection

The latest episode of the Made by Google Podcast (S5 E2) goes into the sound (and haptics) design on Pixel devices. 

At a high-level, sounds are meant to reflect the Google Pixel design principles of “simple,” “human,” and “playful.” The goal is to “offset the burden of information that we’re putting on the visual domain onto other senses,” like sound and haptics.

The most recent example of their work is the “Gems” collection of ringtones, notification sounds, and alarm sounds. Introduced with the Pixel 8 series and made available to all devices, generative AI was used as a “collaborative partner.”

“All of the sounds in this collection were produced using Google MusicFX, a generative AI model that converts text input into music.”

Originally MusicLM, MusicFX is available for you to experiment with in AI Test Kitchen.

It started with a prompt. In the case of “Amber Beats,” this was: “Behind the groovy bass line, there’s an atmospheric pad synth, which adds depth to the track. Bouncy drums are featured with a clear glass hit predominant sound. There are layers of synth sounds.”

The audio designer was particularly inspired by a two-second snippet from the 30-second clip that was generated: “I took that audio clip and began to rhythmically and melodically alter it, building an original idea that felt authentic to me.”

“Our goal with the Gems sound collection was to explore how Google’s in-house tools can unlock creative inspiration and expression and provide an optimistic glimpse into that future.” 

This joins other methods for creating sounds. For example, the Pixel Camera shutter sound was originally inspired by scissors opening and closing, which was “used as the basis for manipulating the final sound that you hear today.” Another consideration in this case is making sure the “shutter motion that you see on the screen fits really well with the sound that you’re hearing.”

Another factor is making sure sounds can be heard in all the noisy environments we use our devices, as well as the constraints of the actual hardware. Optimizing for form factors is another consideration. 

“If it’s a ringtone, I want to make sure that that’s going to sound really good in my home, but it’s also going to sound really good on the street. And then we optimize the sound for the hardware. So we make any orchestration or EQ or whatever it is, any types of changes that need to be made to get the sound sounding as good as possible on the actual device.”

Google also touched upon the process for updating the default “Big Adventure” ringtone to “Your New Adventure” in 2021:

So I would say with the Your New Adventure sound, that was one where we found it challenging to get the transition that we have from this like slow, gentle state to a more faster upbeat, lively tempo and sound. We worked on that a lot, and that was one where there was a long iterative process of trying to get that one right.

Another modernization was the “Popcorn” default notification sound getting modernized into “Eureka.” 

In terms of naming, the sound designers come up with a “musical or aesthetic description” of each clip and then a different team (UX content strategy) names it based on that.