Hear speech through the lens of someone with hearing loss.
Explore how different hearing thresholds can shape speech. Adjust a profile and compare the original with a simple frequency-based simulation.
Shape a hearing profile
Enter thresholds in dB HL. Higher numbers mean softer sounds may be harder to hear.
This simulator uses the classic five-band scheme applied to the pure-tone average (PTA4) of each ear: the mean of thresholds at 500, 1000, 2000 and 4000 Hz. 250 and 8000 Hz are not part of that average.
How the degree is calculated
- Normal: 25 dB HL or less
- Mild: above 25 to 40 dB HL
- Moderate: above 40 to 60 dB HL
- Severe: above 60 to 80 dB HL
- Profound: above 80 dB HL
These classic WHO five-band ranges are adapted here for each ear separately; newer standards use different categories. An average can hide hearing loss at individual frequencies. This is an educational estimate, not a diagnosis. View the source criteria.
Left output Right output · Press play to see audio levels
Estimated output level (dB HL), assuming 65 dB SPL speech. Not calibrated.
Solid: at/above the applied threshold. Dashed: below. Louder levels plot lower. Includes noise; excludes the listening-volume setting. Levels outside 0–120 dB HL are clipped at the chart edges. Preparing the recording reference…
How the sound is processed
Thresholds interpolate linearly in dB on a logarithmic frequency axis between 125, 250, 500, 1,000, 2,000, 4,000, and 8,000 Hz. A Web Audio AudioWorklet measures changing energy in fine frequency sections and applies a smooth audibility gate based on the estimated level relative to the audiogram.
The nearest measured threshold is used outside 125–8,000 Hz. At or below the estimated threshold, each section targets 80 dB of suppression. It opens smoothly over the next 6 dB, retaining louder sections without extra fixed loss. Overlapping windows smooth transitions, so brief crossings can contain transition energy.
Profile changes update processing during playback. There is no extra overall attenuation or volume boost. Reconstruction introduces approximately 171–186 ms of delay at typical 48/44.1 kHz sample rates.
This is not a calibrated mapping to headphone sound pressure. Softer parts may become inaudible depending on the recording, listening volume, headphones, and listener. Processing uses an assumed presentation level and an approximate free-field reference, not individual headphone calibration. It does not identify phonemes or simulate loudness recruitment.
Original uses an independent unfiltered path. The volume slider changes listening loudness, not the audiogram. This demonstration cannot predict speech recognition or reproduce every effect of hearing loss.
Listen & compare
Listen to Anita Slusser’s recording and compare Original with your simulated hearing profile.
Simulation approach
Level-dependent audibility
Assumed, not measured. This sets the estimated dB HL level on the chart, not the listening-volume setting. Lower assumed levels suppress more quiet cues; higher levels let more through.
Analyzing the recording’s digital level… Original remains available.
How the estimate works
Level-dependent processing uses 2,048-sample overlapping windows, updated every 256 samples (about 43 ms and 5 ms respectively at 48 kHz). Window-normalized levels are compared with the recording reference, assumed speech SPL and frequency-specific reference threshold to estimate dB HL. The live curves summarize wider display regions; these are not the individual processing sections. In Level-dependent mode, output levels are measured from the actual reconstructed signal before listening volume.
Fine frequency sections are compared with the interpolated audiogram. At or below the estimated threshold, each section targets 80 dB of suppression. It opens smoothly between the threshold and 6 dB above it, then is retained without extra fixed loss. Short, overlapping analysis windows smooth transitions; brief crossings can still contain transition energy. Zero-threshold sections remain unchanged.
The recording reference averages 20 ms frames, excluding frames more than 30 dB below the loudest frame so long pauses do not inflate the estimated level of speech cues. This is an amplitude-based estimate, not a standardized active-speech measurement. The frequency reference uses historical free-field pure-tone threshold values from ISO 226:2003, Table 1. These are not headphone calibration values; applying them to speech bands is an approximation. References outside the measured range are held at the nearest endpoint.
Zero-threshold bands remain unchanged. Output volume and headphone settings do not calibrate this model. It cannot determine actual audibility, delete named phonemes, simulate masking, or reproduce an individual’s hearing. No extra fixed attenuation is added.
Both ears applies the left profile to the left channel and the right profile to the right channel. Use stereo headphones to hear each side separately; mono recordings are duplicated before the profiles are applied.
Check headphone channels
OptionalPlay each brief tone and confirm it comes from the named side. Pause speech playback first. This check does not change your hearing profile.
Choose a side to test. The tone is brief and low level.
Preparing the recording level reference. Original remains available.
Background noise
A soft, generated ambient hiss, only while speech plays. Added to both Original and Simulated comparisons; the selected ear profile filters speech and noise together. This level is illustrative, not a calibrated signal-to-noise ratio.
Start at a comfortable, low volume. This tool does not estimate safe listening levels.
A human recording of Emily Dickinson’s poem, read by Anita Slusser for LibriVox. LibriVox recordings are public domain in the U.S.; check local rights elsewhere. Source and rights
Read transcript for Anita Slusser’s recording
Because I Could Not Stop for Death by Emily Dickinson, read by Anita Slusser.
Because I could not stop for Death – He kindly stopped for me – The Carriage held but just Ourselves – And Immortality. We slowly drove – He knew no haste And I had put away My labor and my leisure too, For His Civility – We passed the School, where Children strove At Recess – in the Ring – We passed the Fields of Gazing Grain – We passed the Setting Sun – Or rather – He passed Us – The Dews drew quivering and chill – For only Gossamer, my Gown – My Tippet – only Tulle – We paused before a House that seemed A Swelling of the Ground – The Roof was scarcely visible – The Cornice – in the Ground – Since then – ’tis Centuries – and yet Feels shorter than the Day I first surmised the Horses’ Heads Were toward Eternity –
Read by Anita Slusser. Recording source and rights.
