Build up to six strings on a sound box. Tune your instrument and compose your own musical signal.
Same note, different sound?
Your next step and Nallo tips are in the workbench.
Select “Test sound” or play a string.
First select “Test sound” or pluck a string. Check the volume here and on your device, and whether the browser tab is muted. Some devices require you to tap before the first sound can play. A string without a sound box is deliberately quieter.
Your build is ready. What do you predict?
Choose a mission above or build freely. Test a build, save A and change only one thing if possible.
Build up to six strings on a sound box. Tune your instrument and compose your own musical signal.
Pluck strings in the picture. Add note blocks, then tap notes into the eight steps.
Build or change one thing, discuss your prediction, then start the trial. “Undo” restores recent build changes and restarts the trial.
Play a string, change its length or tension, and listen again. Build three notes and compose a signal with rests. Save A, change only the sound box, and replay the same sequence as B. Note labels follow your actual tuning. You may also duplicate an existing note. The yellow band shows your goal, and Nallo gives the next step in the workbench. Each mission has three instructions under “How the experiment works”.
With other conditions equal, a shorter string vibrates faster. Higher tension also raises the fundamental frequency. Greater mass per unit length lowers it. Stronger plucking primarily increases amplitude in this linear model. Visible motion is slowed by a factor of 400; audio uses the calculated real frequency. Trial duration follows the calculated decay, including delayed notes in a sequence, so it ends after playback has faded.
A bare string radiates little sound. Through its support it transfers movement to the top plate and air in the sound box. These systems have resonances of their own. Volume, opening, stiffness and damping affect the frequencies radiated strongly and their decay. A larger box does not simply make every string note louder or lower.
Top-plate thickness changes its modelled natural frequencies. More damping broadens resonances and shortens decay. Without a sound box, body settings have no effect and are disabled. Duplicate strings, tune them independently and play the whole instrument. The 220 Hz reference tone is separate. Missions assess model frequencies and controlled changes, not musical beauty.
Note blocks fit steel strings with 60 N tension and 0.6 mm diameter. Their calculated length is close to the labelled note; no recording is used. All strings remain adjustable. Choose a note with the play keys and tap it into eight steps. Tapping the same note again makes a rest. The sequence plays your strings at your chosen tempo. Change one string or the sound box and play it again. Plans preserve up to six strings and eight steps.
Save a note with the same selected string. Change only between no sound box, a box and guitar form. Replay A and B. Did pitch change or mainly timbre? Then change only the plucking point. Some partials weaken because they are barely excited at that point.
An adult can stretch an elastic band around a stable empty box. Pluck gently and compare with a band held without the box. Change only vibrating length for one comparison. Keep it away from faces; use no glass or sharp edges.
The fundamental follows the ideal string model. Fourteen damped partials come from plucking position and strength. The body combines a Helmholtz resonance and a few idealised plate modes. Shape and material affect model parameters; this is not a full simulation of a real guitar. Speakers, headphones and volume settings affect playback. Sound starts only after a deliberate play action and output is limited. A real flexible body has more complex coupled behaviour. Overtone bars are relative to the strongest partial; the separate sound-strength value is a relative model amplitude.
Predict before starting, save a completed trial as A, and change only one thing. The result preserves the prediction and lists changed parts. Feedback explains when a comparison has changed too many conditions. Free building remains available.
A learning simulation with simplified, transparent models. Save your build as a plan. A/B comparisons last until the page reloads. Your plan, prediction and results are not sent to a server.