Coming soon
A long form video and audio explainer dropping soon. The full project, the multi quarter roadmap, why it's open source, and how riders, tuners, and engineers fit in. Get on the list.
Open source. Physics. Built in the open.
An open-source engine and exhaust simulation framework. From acoustic visualisation today, through audio synthesis and inverse exhaust design, to engineering grade simulation and consulting by late 2028.
A demo of two bikes lives in the Acoustic Lab today. Custom signatures available via the merch page.
What lives here
A working demo of two bikes. Drag the RPM slider; watch the ECG, DNA, and Genome shift. To see your own bike, head to the merch page.
Open the LabHear what your machine sounds like, generated from your engine spec and full exhaust geometry. No recording needed. A neural residual model fills the gap between physics and reality. Targeting late 2027.
Describe a target sound. Get the exhaust geometry that produces it. A reinforcement learning agent searches the topology space; Bayesian optimisation tunes the parameters. Targeting mid 2028.
The engineering
Three methods, one framework. Each layer adds a capability the previous one cannot deliver alone.
Transfer matrix method
The physics backbone. Munjal's plane-wave acoustic theory cascaded across every element of the exhaust: pipes, expansion chambers, perforations, junctions, side branches.
Currently in progress. Multi-element TMM with mean flow and branching topology.
Neural residual synthesis
Physics predicts the spectrum. A neural model learns the gap between prediction and recorded reality. The output is audible exhaust sound, generated entirely from engine and pipe geometry.
Targeting Q4 2027. Trained on a dataset of ~30 motorcycle recordings.
RL inverse design
A reinforcement learning agent learns which exhaust topologies match which acoustic targets. A second stage optimises the continuous dimensions. The result: from a target sound to a buildable pipe geometry.
Targeting mid 2028. Publication-worthy methodology.
The fidelity climb
A multi-stage build, in public. The framework matures from acoustic visualisation today, through audio synthesis and inverse exhaust design, to engineering grade simulation by Q4 2029.
Open source commitment
The framework is being built privately while the core products mature. Code, methodology, calibration data, and papers will be shared openly when ready. MIT licensed throughout. The release happens when the work is solid, regardless of timeline.
Engineer, researcher, or hobbyist who wants early access? Email me.
Engineering consultancy
Acoustic engineering consultancy for motorcycle and motorsport exhaust design. Pre-fabrication design from a target sound. Validation and optimisation of existing systems. Ongoing R&D partnership.
The open-source simulation framework being built here is the engineering foundation for this work. If you design or fabricate exhaust systems and have a problem worth solving, the line below reaches me directly.
Get in touchWho is building this
I am a Chartered Professional Engineer (CPEng) registered in Australia with an aerospace background. I ride a CFMoto SR-S 450. I built this because I wanted to understand what my exhaust note actually is. As physics, as data, as something I could look at.
This is not a corporate product. It is a workshop. I am building and publishing the engineering as I go. I want to collaborate with riders, builders, and engineers who think the same way.
Read the Field NotesDirect line
aboude@exhaustnote.engineer
Have a machine. A question.
A build. An idea. Write to me.