The short answer
Talk into a cardboard tube and your voice bounces back and forth between the two ends. Some notes fit the length of the tube exactly, so each bounce arrives in step with the one before it and the two add up. Those notes get louder and keep ringing after you stop. The rest arrive out of step, partly cancel, and fade.
That is resonance: a space handing some frequencies back to you louder than the rest. The pitch of the ring is set by the size of the space, because a longer tube fits longer waves and longer waves are lower notes. A paper towel tube rings around 600 Hz. A long pipe rings low. A thimble rings high enough to barely register.
Which notes a tube lets ring
A wave also fits a tube when it goes round twice, or three times, or four, so a tube rings at its own note and at the multiples of it. The pattern of which multiples survive is as audible as the note itself, and it is what separates a cardboard tube from a tin can.
Why a closed end sounds hollow
When sound reaches the open end of a tube it bounces back flipped, upside down compared with how it arrived. A tube with two open ends flips it twice, so the wave comes back the right way round and lines up after one round trip.
A tin can has one closed end, and a closed end does not flip it. The wave comes back inverted after one trip, which cancels rather than adds, so it needs a second trip before it lines up. Two trips is the same as a tube twice as long, which is why a can rings an octave below an open tube of the same length.
The flip takes something else with it. Every second multiple above that low note arrives inverted as well and cancels, so the can rings only on the odd ones. A missing every-other-peak is what your ear reports as hollow or boxy, and it is the difference you hear between shouting into a tube and shouting into a can.
Resonance on a filter, and why Q is the same idea
The same word shows up on tools that have no tube in them. A filter's resonance control decides how much the filter feeds its own output back to its input right at the cutoff point, and feeding a frequency back to itself is exactly what the bounces in a tube do. Turn it up and the filter stops being a slope and grows a peak that rings.
On an EQ band the same quantity is called Q, and it sets how narrow the band is. A narrow band is a sharp ring and a wide one is a gentle lift, which is why a high-Q boost sounds like a note and a low-Q boost sounds like tone. On a high-pass or low-pass filter the same control puts a bump right where the filter turns over.
Your own voice runs on this. Your throat and mouth are a tube, and the notes that fit it are your formants: change the shape and you change which notes ring, which is how you say one vowel instead of another. Helium moves all of them at once by changing how fast sound crosses that tube.
Putting a voice inside something
In Patchd the node for this is the Comb, one of the Studio voice-fx nodes. It plays your voice back into itself a little late, over and over, which is what the bounces in a tube do without needing a tube.
- Frequency (50 to 2000 Hz, default 440). How late each copy arrives, stated as the note that fits. Turn it up for a short tube and a high ring, down for a long pipe.
- Resonance (-90 to +90 percent, default 60). How much of your voice comes back on each pass. More means the ring lasts longer. Turn it negative and each copy comes back flipped, which is the tin can: an octave down, every second peak canceled, hollow.
- Damping (0 to 100, default 40). How much of the top end is lost on each pass, the way a real tube absorbs the high notes faster than the low ones.
One of the Persona voices is built from exactly this. Bullhorn's Horn control is the Comb's frequency, running 800 to 2000 Hz and sitting at 1200 by default, because the flare of a bullhorn is a tube like any other.
The takeaway
Resonance is a space giving you back the notes that fit it. The size sets the note, a longer space rings lower, and a tube rings at its note plus the multiples of it. Closing one end flips every returning wave, which drops the ring an octave and cancels every second peak, and that gap is the hollow sound of a can. On a filter the same idea wears the name resonance or Q, and in your own throat it is what makes a vowel.
Patchd is a Windows audio mixer with a full effects rack on every channel, and the Comb node is in the Studio voice-fx rack. It is in open beta now, and if you clear the terms your invitation arrives straight away. The Studio tier is $5.99 a month, $39.99 a year, or $119.99 once for a lifetime license, and the real-time engine is identical on both tiers.