Speaker Test

After reconnecting speakers or changing headphones, sound may come from the wrong side, seem uneven, or lose bass. Short listening comparisons make these symptoms easier to describe than a busy music recording. A repeatable observation about where and how the sound changes can guide the next connection, placement, or playback-setting check.

Stereo audio carries separate left and right channels. Left means the listener's left when facing the speakers, or the ear marked L on headphones. Channel routing determines where each signal goes through the selected output, connections, and playback equipment. A mono setting combines the channels, so a signal intended for one side may reach both. Sound being audible therefore does not establish that stereo separation is intact.

Relative polarity describes how the two waveforms line up in sign. Inverting one channel turns its positive values negative and its negative values positive without exchanging the channels or adding a delay. When the channels interact, the sound may become fuller, thinner, or differently located. A change identifies a difference between the two listening conditions. It does not, by itself, identify reversed wiring or establish the absolute polarity of either speaker.

Loudspeakers in a room

Both speakers reach both ears, and reflected sound joins the direct sound. Speaker spacing, distance from walls, and listening position can strengthen some frequencies and weaken others. Moving your seat can change a polarity or bass comparison even when every connection stays the same.

Headphones

The channels largely reach separate ears, so the sound does not combine in the room in the same way. A polarity change may alter the listening impression without the same acoustic cancellation that can occur between room speakers. Headphone fit and processing that mixes some of each channel into the other also affect the comparison.

Frequency, measured in hertz (Hz), is the number of waveform cycles per second. Separate low-frequency tones make it possible to compare a few bass pitches. A sweep changes pitch continuously and can draw attention to a rattle, dropout, or change in audibility across a wider range. A faint bass tone may reflect limited low-frequency reproduction or a cancellation at the listening position. Equal digital amplitudes do not guarantee equally loud tones, and a fading sweep cannot distinguish equipment response from listening conditions or hearing.

Digital level is another separate quantity. Decibels relative to full scale, written dBFS, describe a signal's amplitude relative to a digital reference. Sound pressure level (SPL) describes physical sound pressure, while loudness is the listener's perception. System volume, amplification, and the speakers or headphones determine how a digital signal becomes sound. Begin with low system volume, keep comparisons brief, and stop if uncomfortable. Raising the volume to chase a faint tone can make the next tone unexpectedly loud. Listening comparisons provide neither a hearing assessment nor a calibrated response measurement.

How to Use This Tool:

Use the output you want to investigate and keep your seating position or headphone fit consistent between attempts.

1. Choose the intended output in your system's sound settings and turn its volume down. The tool uses that output and does not change system volume or select a device for you.
2. Select a Check. Use Left / right identification for channel routing, Stereo polarity comparison for the effect of reversing one channel, Bass steps for separate low tones, or Frequency sweep for a rising tone.
3. Set Digital peak between −48 and −18 dBFS, inclusive. The starting value is −30 dBFS. This setting controls the generated signal amplitude and cannot establish a safe acoustic volume.
4. Adjust the optional sweep fields in Advanced only if you want to change the sweep. Sweep start accepts 20 to 15,999 Hz, Sweep end accepts 21 to 16,000 Hz, and Sweep duration accepts 2 to 12 seconds. The end must be higher than the start. Blank fields use the defaults shown below; other checks ignore these refinements.
   If the summary says Check settings, correct the named field before playing. Digital peak cannot be blank. For a reversed or equal sweep range, lower the start or raise the end within the permitted bounds.
5. Press Play test and follow the channel and tone cues. Every check ends automatically. Changing settings during playback affects the next attempt. Use Stop or Escape to end early; hiding the tab, leaving the page, or an audio interruption also ends playback. Returning to the page does not restart sound.
   If Audio unavailable appears, follow the accompanying message. Audio may require browser sound permission, a supported browser, or an output that exposes stereo channels. For a frequency-range error during a sweep, lower Sweep end. If a fixed-tone check cannot run, try a supported browser or output. Retry with Play test after correcting the cause.
6. Select What you heard after playback ends, open Advanced to add optional Listener notes, and press Save observation. The entry belongs to the latest attempt, including an interrupted attempt. Read its Suggested next check in the Listening report, then change one relevant setup condition before repeating.

Interpreting Results:

The L and R cues show which digital channels are being generated. Confirm the physical sound yourself. These checks use sine tones, with a 0.5-second quiet gap between separate tones. The frequencies, durations, and digital level limits are declared presets, not an acoustic testing standard.

Complete means the planned sequence ended. It is not a hardware pass. Stopped early, Tab hidden, and Audio interrupted identify incomplete attempts that should be repeated before drawing a comparison. The duration in a report's signal description is the planned total, not a measurement of how long sound was heard.

Your saved selection supplies the observation and the suggested follow-up. An As expected entry is your judgment; the observation count is not an automatic test score.

• No sound heard calls for checking tab mute, system output, connections, and a familiar audio source.
• Wrong side / both sides calls for checking L/R labels, stereo connections, balance, and mono-audio settings.
• One side quieter calls for comparing seating or headphone fit, balance, and connections before repeating with another source.
• Buzz / rattle / dropout calls for a quiet comparison with familiar audio and a check of nearby objects and connections.
• For Difference heard, compare again from the same position. For No clear difference, inspect mono mixing, placement, and processing before repeating.

Signal plan and Frequency path describe the currently selected settings, which can change while an earlier selection is playing. The Listening report retains each attempt's original signal and digital peak. The frequency chart plots generated frequency against time, with a logarithmic frequency scale and five plotted points for a sweep. Its line shows no measured speaker response, loudness, or microphone evidence.

Technical Details:

A sine tone varies smoothly around zero. Its frequency determines the rate of oscillation, while its amplitude determines the size of that variation. Routing, polarity, and frequency can therefore be compared separately while the nominal digital peak remains fixed.

Formula Core: digital peak and amplitude

A digital peak expressed in dBFS converts to a linear amplitude ratio relative to full scale.

D is Digital peak in dBFS, and A is the positive amplitude ratio used to form the channel gains. At −30 dBFS, the ratio is approximately 0.0316228, or 3.16% of full-scale peak amplitude. At −20 dBFS it is 0.1; at −40 dBFS it is 0.01. A 20 dB reduction divides amplitude by ten without changing frequency or duration. These ratios are neither percentages of perceived loudness nor system-volume settings.

A negative dBFS value means the peak is below the full-scale reference. A negative channel gain means the waveform is inverted. Those minus signs have different meanings. The peak setting also does not describe an average signal level or provide a sound-pressure reading.

Channel routing and relative polarity

Each channel receives a scaled copy of the same sine waveform. A zero gain leaves that generated channel silent, while changing the sign reverses the waveform without changing its magnitude.

For an ideal equal-weight mono mix, adding identical opposite-polarity signals cancels them. A quiet or missing inverted segment can therefore result from mono mixing before sound reaches the speakers. With separate loudspeakers, unequal distances, reflections, and processing change how the signals combine at the ears. A stronger same-polarity segment or a stronger inverted segment cannot identify which physical connection, if any, is reversed.

For a single steady sine tone, inversion is equivalent to a 180-degree phase difference. It does not add a time delay. A real delay changes phase by different amounts at different frequencies, which is why a polarity switch cannot generally correct the effects of unequal acoustic paths across a whole sweep.

Formula Core: exponential sweep

An exponential sweep covers equal frequency ratios in equal time intervals. An octave is a doubling of frequency, so equal octaves take equal amounts of time.

f₀ is Sweep start, f₁ is Sweep end, and the resulting frequency is in Hz. T is Sweep duration in seconds, and t is elapsed time from the beginning of the sweep. The end frequency must exceed the positive start frequency.

A calculated example using 100 Hz to 1,600 Hz over 4 seconds doubles the frequency every second. These are generated frequencies, not observations from a physical speaker.

The midpoint frequency is the geometric mean of the endpoints, giving 400 Hz in this example. The generated sweep moves continuously between the tabulated values.

Timing, precision, and playback limits

Each tone's amplitude rises from zero over 40 milliseconds and falls back to zero over its final 40 milliseconds. These fades are included in the stated tone duration. A normal stop fades the overall signal over approximately 30 milliseconds; a hidden-tab or audio interruption releases playback immediately.

Displayed signal-plan values use at most four decimal places. Playback calculations retain the underlying values. For example, a displayed gain of 0.0316 corresponds to the calculated ratio of approximately 0.0316228.

The highest generated frequency must be strictly below half the browser's audio sample rate, the Nyquist frequency. Playback is blocked at or above that boundary, even if the requested sweep end is within the form's allowed range. The sample rate and stereo capability exposed by a browser do not establish a speaker's acoustic bandwidth or confirm the physical left/right connections.

Privacy Notes:

Sound generation and the listening report run in your browser. The tool does not request microphone access, record audio, or upload the listening report. Observations are the choices you save after listening, and optional notes describe your own setup.

The report remains in the current tab until it is cleared or the page is reloaded. It holds up to 30 attempts, with up to 500 characters of notes for each attempt. Download a report you want to keep before reloading or using Clear report. At 30 attempts, clear the report before starting another check.

The page address can retain the selected check, digital peak, and sweep settings. It does not include observations or listener notes. Opening that address restores the setup with audio off and without the previous report.

References:

• Web Audio API, World Wide Web Consortium, 17 June 2021.
• Deafness and hearing loss: Safe listening, World Health Organization, 6 March 2026.
• Quantization and decibels, Carnegie Mellon University.
• Polarity, phase and delay, DPA Microphones.
• How to Position Studio Monitors and Find the Best Listening Position, Genelec.
• HSP (Headphone Spatial Processing), Meridian Audio.