In theatre there's sometimes a moment most people don't consciously register: a sound before the character speaks. A breath, a room tone, a minimal shift in atmosphere. It determines how the line lands. Before a single word.
That's exactly my job as a sound designer.
Now think of your AI assistant. It receives a query. It processes. What does the user experience? A spinner. A pulsing dot. Acoustically, often nothing. Neutral at best.
Even the simplest level would be valuable: acoustic feedback. Listening. Processing. Done. Technically simple. And already this can be more than status: character, warmth. Is my assistant just a neutral tool, or could it be a friend?
In natural conversation, turn-taking happens in about 200ms. With AI, we're at several seconds. Most products treat that gap as dead time.
Designing this moment is structural work.
#AIDesign #VoiceAI #SoundDesign #ConversationalAI
Proud and happy: I brought back Taryn Simon's sound installation "Start Again The Lament" in the cloister and the cells of a former monastery in Coimbra, together with my close friend and lighting designer Urs Schönebaum. I did the sound composition and design for this unique piece.
After the premiere at Cisternerne Copenhagen in 2024, @anozerocoimbra Bienal de Coimbra is the first re-staging. It's very different in Coimbra than in the cisterns. Possibly even more amazing.
"You need the nerves of a ghost-hunter to walk through the pitch-black ground-floor corridor of the dormitory wing, lit only by a neon strip at either end, where tortured wails ambush you from the monkish cells. Sung in Albanian, Chinese, Kurdish, Kyrgyz and Turkish, these laments are part of an installation by US artist Taryn Simon, but they feel like spectral reminders of the nuns who lived in these quarters for two centuries."
Philip Oltermann writes about the festival in The @guardian : "create art which can only happen here and nowhere else." Exactly that. These voices, this place, this acoustic. Impossible elsewhere.
Anozero—Bienal de Coimbra runs through 5 July 2026. Anyone here who has heard it, in Copenhagen or Coimbra?
Sound on.
Start-up sound. Shut-down sound. And everything in between.
For @harman_intl AutoOne, we started with the brand. What does @harmankardon actually stand for, sonically? Once we had answers, @joerg.huettner and @mathisnitschke shaped start-up and shut-down sounds in Dolby Atmos from there.
What started as a sound design brief became an impulse for redefining how Harman Kardon sounds inside a vehicle. CES 2025.
Thank you Sabrina Ceylan, Tanaponn Kulsedzeranee and Philipp Siebourg @siebourg and team for the super fun and creative collaboration and trust!
A soundwalk that works anywhere in the world. Instead of pre-composed content for a fixed route, City Songs classifies your real-world context (weather, sun position, temperature, environment) and generates a site-specific sonic experience in real time. Custom OpenStreetMap parsers. GPT-4 for location-specific poetry. The context classification pipeline behind this project is the direct precursor to our adaptive automotive sound systems.
An experimental project funded by @fffbayern
#CitySongs #SoundArt #GenerativeAudio #Soundwalk #AdaptiveSound
Today: Beyond Silicon in Nuremberg. Tomorrow: Financing 4.0 in Hamburg.
These two events are connected. The Hamburg panel at the @gces.hamburg is held in cooperation with Bayern Innovativ, the same institution running Beyond Silicon in Nuremberg today: Quantum, Neuromorphic, Photonic and Biological Computing.
The hyperscaler criticism of the past months is justified. At CORPUS we are working on a different geometry: models that live inside the product itself. Low latency, context-aware, no data stream flowing back to a rack.
What is on stage in Nuremberg today decides what can be said in Hamburg tomorrow about financing in the creative economy.
The fun starts when we leave the silicon age behind.
#CORPUS #BeyondSilicon #OnDeviceAI #CreativeEconomy
Products make sounds. Some on purpose, most not.
Our coffee machine squeaks while heating up because it needs descaling. Nobody designed that squeak, and yet it's information.
For maybe 90% of operating life, a product isn't signaling anything explicit. But it's rarely truly silent. It hums, clicks, vibrates softly. What nobody meant is still being heard.
What would change if a product team took the 90% as seriously as the 10%?
#SoundDesign #ProductDesign #UXDesign #AutonomousProducts
The public AI music debate has narrowed to a single question: who gets paid when a machine generates a song.
The question is legitimate. It operates inside a frame that is already being outgrown.
In cars, hospitals, games, rehabilitation, the finished song is not the product. What is needed is sound that reads a situation and keeps responding to it, in real time, on device. These markets are regulated, context-specific, machine-to-machine. Anything built around the prompt-in, song-out flow is architecturally incompatible with them.
We wrote a new Journal article about this, because the infrastructure these adjacent markets need (a semantic layer, real-time generation on device, and an auditable training corpus) is what we are building at CORPUS.
What is arriving is something older: sound that responds to the situation it is in.
Go to journal.corpus.music to read it.
#AIMusic #GenerativeAudio #MusicTech #CORPUS
Engineering builds the product. Later, someone realizes it needs audio feedback. Firmware picks a frequency. UI says sound isn't her department. PM says "later." Later arrives without budget. The product ships with a placeholder beep that was never meant to be anything.
Every one of these was a design decision. Just made by accident.
The fix is to recognize sonic communication as a design discipline that belongs in the product architecture from the start.
#ProductDesign #SoundDesign #UXDesign #DesignLeadership
Your robot has a friendly face, rounded edges, smooth movements. And still it feels off when it comes toward you.
It's the sound. Motors loud, servos whining. Generic beeps that say "I'm doing something." The eye sees a sleek machine; the ear hears a device pulling itself together.
Speech would be the obvious answer. But talking robots create conversational expectations no one wants to fulfill.
Research on pedestrian-robot interaction shows: multimodal signals (sound and light combined) are the most effective way to communicate intent without permanent noise.
I come from music theatre. The core problem is the same. How do you make presence and intent audible without being intrusive?
#Robotics #HumanRobotInteraction #SoundDesign #TrustInTechnology
The CORPUS Journal has been relaunched.
New architecture, new website, new publishing pipeline. The articles we wrote ourselves have been cleaned up sentence by sentence: tighter focus, sharper language, fewer rhetorical reflexes. Dare I say: less sloppy?
If something on the Journal doesn't meet the bar we set for the project itself, tell us.
A particular thank you to our guest contributors Kasia Glowicka, Lars Ullrich, and Steffen Holly. Their pieces shaped the Journal beyond what we could write ourselves.
Link in bio.
Vergehen (2016). An opera you walk yourself through, along Munich's Isar river. GPS, speed, direction mapped to music. @sarah_aristidou 's voice, @cellanja Anja Lechner's cello, analog electronics, digital sound art. All layered onto the river and whatever the river happened to be doing that day.
The piece is about memory. What if technology could record experiences so perfectly that the recording becomes indistinguishable from the original? Would we want that?
It was the starting point for my work with new technologies. I deliberately opened the door with it, out of the concert hall and out of the theater. How much does the system determine, and how much does it leave to context?
Still available on Android. Head over to
#SoundWalk #InteractiveAudio #SoundArt #Munich #AdaptiveSound
Most sound design stops at the active moments. Startup chime. Notification. Error tone. Each gets weeks of refinement. Everything in between gets treated as "no sound needed."
But that in-between is the default state. A car at highway speed. A robot on standby. A voice assistant that isn't speaking. These states aren't silent. They have an acoustic texture: ventilation, mechanics, the absence of something that was just there. Nobody designs them.
The active moments might account for ten percent of usage time. The other ninety percent is this in-between. If you don't design that, you leave the acoustic character of your product to chance.
Designing idle states doesn't mean filling them with sound. It means shaping the transition. The difference between a room that's quiet and a room that's empty.
#SoundDesign #ProductDesign #HMI #AutomotiveUX #AmbientSound
