Sound Fundamentals · Waveforms & Playback Hardware

Waveforms and playback: what changes between headphones and speakers

How a sound is generated in a circuit or file is only half of the acoustic story. Whether you listen through stereo headphones, room loudspeakers, or a mono speaker determines whether left and right channels stay isolated, mix in the air, or sum electrically.

How the sound works How it is perceived Page updated: 11 November 2026 · Literature search cutoff: 10 November 2026

1. How to read time-domain and spectral plots

The figures in this guide are analytical or illustrative models computed from mathematical equations, not measurements of a particular playback system. Reading them accurately requires checking four structural elements on each figure:

  • Horizontal axis and units: Time t in milliseconds (ms) or seconds (s) for waveforms; frequency f in hertz (Hz, often on a logarithmic scale) for spectral plots.
  • Vertical axis and normalization: Normalized instantaneous amplitude (-1.0 to +1.0), channel gain (0.0 to 1.0), or relative power spectral density in decibels (dB) referenced to 1 kHz.
  • Physical waveform vs. envelope overlay: Rapid oscillations show the audible carrier wave, while dashed or solid outer curves trace the slower amplitude envelope E(t).
  • Left vs. right stereo channels: Separate traces show whether modulation already exists inside a single channel or only appears when two channels are compared or mixed.

2. What changes between stereo headphones, speakers, and mono playback

The acoustic signal that reaches the eardrums depends on both the generated audio channels and the playback path. Stereo headphones isolate the left and right ears, whereas open-air loudspeakers radiate into a shared room where sound from both speakers reaches both ears (acoustic crossfeed) and interacts with room reflections (Hartmann, 1997; Moore, 2012).

Sound family Stereo headphones Stereo loudspeakers in a room Mono downmix or single speaker
Binaural beats
f1 Left, f2 Right
Preserves separate channels: Each ear receives an unmodulated sine wave; the beat percept arises from central binaural interaction. Mixes channels in the room: Acoustic crossfeed between speakers partially sums f1 and f2 in the air, converting part of the stimulus into a physical monaural beat alongside room reflections. Converts to a physical monaural beat: An averaging mono downmix M(t) = 0.5·(L + R) produces 0.5·[cos(2π·f1·t) + cos(2π·f2·t)] (or an unattenuated sum L + R doubles the peak excursion), creating a physical amplitude-modulated wave.
Monaural beats
f1 + f2 in same channel
Delivers the modulated signal to the ear: Both tones are summed in the same channel before the transducer (subject to headphone fit and response). Radiates the modulated signal into the room: The source signal carries the |f2 - f1| modulation, though room standing waves, reflections, and speaker frequency response can alter the relative balance of the two tones and the depth of the envelope at the listener's ears. Preserves the source modulation: For identical left and right channels (L = R), an averaging mono downmix 0.5·(L + R) preserves the original signal level (0 dB change), whereas an unattenuated sum L + R doubles amplitude (+6.02 dB) without altering the beat rate |f2 - f1|.
Isochronic tones
Periodic amplitude pulses
Delivers the pulse envelope to the ear: Onset/offset ramps and quiet gaps are reproduced directly at the ear canal without room reverberation. Radiates the pulse train into the room: The source signal retains the periodic pulses, while room reflections and reverberation can partially fill in the quiet intervals between rapid pulses at the listening position. Preserves the source pulse train: Single-channel or identical-channel pulses retain their repetition rate and duty cycle in the electrical signal.
Alternating & bilateral tones
Left ↔ Right panning or pulses
Preserves full left–right separation: Events alternate cleanly between the left and right ears. Narrows spatial separation: Both ears hear both speakers with interaural level and time cues, creating a narrower left–right stage in front of the listener. Removes the left–right movement: Summing L and R collapses spatial alternation into a centered stream; whether the mono level stays flat or dips/peaks at center depends on the panning law (linear vs. constant-power) and whether the two channels carry phase-aligned carriers.
Colored noise & soundscapes Maintains stereo separation without room reflections: Delivers independent left–right spatial textures directly to each ear, while perceived tonal balance depends on the headphone transducer, acoustic seal, and any device signal processing. Shapes spectrum via speaker & room response: Perceived low- and high-frequency balance depends on the loudspeaker's frequency extension, placement, and room acoustics, while room reflections blend stereo soundscapes. Collapses stereo width: For an unattenuated sum L + R of two equal-level channels, independent (uncorrelated) noise sources add in power (+3.01 dB RMS), whereas identical (correlated) signals add in amplitude (+6.02 dB). Under an averaging downmix 0.5·(L + R), identical channels stay at 0 dB while uncorrelated noise drops by -3.01 dB RMS.

3. Analytical waveform and spectral diagrams across the six families

Each diagram below illustrates the mathematical structure of one sound family using explicit time or frequency axes and normalized scales.

A. Dichotic binaural presentation vs. physical linear sum (0 to 100 ms)

Three-row analytical waveform plot from 0 to 100 ms showing an unmodulated 211 Hz sine wave in the left channel, an unmodulated 221 Hz sine wave in the right channel, and their physical linear sum with a destructive node at 50 ms.
Analytical model: xL(t) = cos(2π·211t) and xR(t) = cos(2π·221t) over 100 ms. In dichotic headphone listening (rows 1 and 2), neither ear receives a physically modulated envelope; row 3 shows the physical interference pattern that arises only if the two signals are mixed in air or in a mono circuit.

B. Monaural two-tone superposition and carrier zoom (0 to 200 ms & 0 to 50 ms)

Two-panel analytical waveform plot of equal-amplitude two-tone superposition at 211 Hz and 221 Hz showing two 100 ms beat lobes over 200 ms and a 0 to 50 ms zoom into the product waveform where the 5 Hz cosine factor scales the 216 Hz rapid cosine factor to zero at 50 ms.
Analytical model: x(t) = 0.5·[cos(2π·211t) + cos(2π·221t)] = cos(2π·216t)·cos(2π·5t), where 216 Hz is the rapid cosine factor, 5 Hz is the signed slow cosine factor, and 10 Hz is the repetition rate of the unsigned envelope |cos(2π·5t)|. Panel A shows two 100 ms physical envelope lobes within one 200 ms period of the signed 5 Hz factor; Panel B zooms into 0 to 50 ms as the signed 5 Hz factor scales the product waveform to zero at t = 50 ms.

C. Isochronic pulse shaping: Raised-cosine ramps vs. rectangular gate (0 to 200 ms)

Two-panel plot over 200 ms comparing a 10 Hz raised-cosine pulse envelope (duty cycle 0.65, ramp fraction 0.14) against an unsmoothed rectangular on/off gate on a 200 Hz carrier.
Illustrative model: A 200 Hz sine carrier modulated at 10 pulses/s (100 ms period, duty cycle d = 0.65). Gradual raised-cosine onset and offset ramps (e = 0.14) avoid the rectangular gate's step discontinuity in the envelope function Grect(p) and the resulting slope discontinuity (or waveform step jump when gated at a nonzero instantaneous carrier amplitude) in the modulated signal.

D. Continuous stereo panning vs. alternating left–right pulses (0 to 1000 ms)

Two-panel plot over 1000 ms comparing continuous 2 Hz sinusoidal stereo panning against discrete alternating left and right pulses at 4 total pulses per second.
Analytical model: Panel A shows continuous sine/cosine panning (gL2 + gR2 = 1) completing two full left–right cycles in 1000 ms (2 Hz cycle rate). Panel B shows discrete alternating pulses at 4 total pulses/s (2 pulses/s in the left ear and 2 pulses/s in the right ear).

E. Idealized power spectral density (PSD) slopes of noise colors (20 Hz to 20 kHz)

Log-frequency plot from 20 Hz to 20 kHz showing idealized power spectral density slopes per 1 Hz bandwidth normalized to 0 dB at 1 kHz for white (0 dB/oct), pink (-3.01 dB/oct), and brown (-6.02 dB/oct) noise.
Idealized spectral model: Relative power spectral density per 1 Hz bandwidth on a logarithmic frequency axis (20 Hz to 20 kHz), normalized to 0 dB at 1 kHz. Because octave bands double in width with each octave, power per octave band has a slope +3.01 dB/octave higher than PSD per Hz.

F. Procedural soundscape macro-envelopes and filter modulation (0 to 25 s)

Three-row schematic plot over 25 seconds showing a 10-second wave-like surge envelope, slow multi-band texture drift with short droplet impulses, and a slow resonant filter center-frequency trajectory from 400 to 900 Hz.
Illustrative model: Procedural soundscapes combine slow periodic or quasi-periodic macro-envelopes and filter sweeps with continuous stochastic noise and transient impulses, avoiding fixed audio-file splice seams.