Results
Demo: Beethoven Piano Sonata No.12, Op.26 Excerpt
This is a working demo of Mattheson, a research platform I am building for studying how the expressive choices a performer makes in the moment actually land with an audience. This is not how a critic describes them afterward, but what listeners feel, tap by tap, as the performance unfolds. Here is an excerpt from Beethoven's Piano Sonata No. 12 in A♭ major, Op. 26. I deliberately chose the vocabulary of the passions from historical writing on musical affect (in this case, The Emotions of the Human Mind by Thomas Brown in 1820), the same kind of language historical theorists used to describe how music was supposed to move a listener.
Audience responsei
Each dot is one audience response: bigger = felt more intense, and overlapping dots darken where several people responded around the same moment. Shaded bands are the intended passions marked for this passage — deeper color means it was meant more strongly. Hover a dot or band for the exact detail, or hover empty space on a row to see how many nearby taps there are. Scroll or drag to pan the timeline, Ctrl/⌘+scroll (or the +/− buttons in the corner) to zoom into a stretch of it, and the ⤓ icon saves the graph as an image. Drag the scrub bar under the chart, or the video's own controls, to seek the recording. Whatever's currently in view here also sets the time region the sections below summarize.
Currently showing 0:00–2:47 — full timeline.
Responses by passion
Data from the full timeline (0:00–2:47) — zoom the chart above to narrow this to one passage.
How often each passion was tapped, and how intense those taps ran on average (0–1, the same scale as the ladder’s weights).
Intent vs. response
Data from the full timeline (0:00–2:47) — zoom the chart above to narrow this to one passage.
Of every response, the share that landed inside a window you marked as intending that same passion — counting a match against any passion intended at that instant, not just one. A rough descriptive signal, not a statistical test.
Match rate over time
Same data as above, broken into short stretches — where the audience tracked intent well vs. where they didn’t, instead of averaging the two into one number (scale on the left). The dashed line along the bottom marks stretches with no responses at all, left as a gap rather than drawn as a false zero. The vertical dashed marker tracks the video's current position — drag the strip below to seek it.