Experimental conditions across the five experiments (rows).

Hardness discrimination with softened fingertips.

Participants wore an inflated bubble to reduce fingertip stiffness and tapped two identical rigid blocks; one tap was augmented with a brief vibration. Confusion values denote the proportion that the row stimulus was judged harder than the column stimulus. Judgments increase monotonically with the delivered total spectral energy, largely independent of waveform family.

Material identification with a soft fingertip.

A foam tap augmented by synthetic vibration was matched to one of five real materials. Response distributions shift systematically from foam → metal as total spectral energy increases, indicating that energy conveys material identity even when fingertip mechanics are softened.

Hardness discrimination on a compliant surface.

With bare fingers, participants tapped identical foam blocks; one tap included a vibration burst. Confusion matrices show that perceived hardness follows the signal’s total spectral energy across waveform types, with small residual confusions only at the highest two levels (wood-metal).

Material identification on a soft substrate.

Foam taps augmented by synthetic vibrations were matched to real materials. Despite reduced precision under dual compliance (soft surface; bare finger), perceived category still scales with total spectral energy, shifting from foam/silicone at low levels to wood/metal at high levels.

Cue-conflict classification.

While tapping foam, participants received brief vibration transients in which the dominant frequency class and the RMS (total spectral energy) were crossed. Choices tracked the rendered spectral energy (RMS target) rather than dominant frequency or base material, demonstrating that total spectral energy dominates perceived material identity under conflict.

Spectral analysis of vibration recordings from 20 everyday materials.

Harder materials clustered together by broadband spectral energy, whereas softer materials formed distinct groups. Five representative materials (foam, silicone, rubber, wood, metal) were chosen for the main experiments.

Examples of processed compensation signals generated via frequency-domain subtraction.

Behavioral paradigms and tasks.

A Fingertip-discrimination setup: bubble-softened index fingers tap two rigid blocks; one tap includes vibration (2AFC: harder). B Fingertip-matching setup: bubble-softened finger taps foam with vibration and is matched to a real material (5-AFC). C Surface-discrimination setup: bare finger taps two foam blocks; one tap includes vibration (2AFC). D Surface-matching setup: bare finger taps foam with vibration and is matched to a real material (5-AFC). Triggers align transients to contact with sub-millisecond latency.

Apparatus and instrumentation.

A Experimental bench with photoelectric contact sensors and Bela real-time playback. B Inflatable fingertip bubble (5 psi) used to reduce fingertip stiffness while leaving the nail accessible for actuation.