Volume Level Discrimination Test

Every round of the Volume Level Discrimination Test starts with an obvious volume gap and gets harder from there — click Start Test, then pick First or Second for whichever sounds louder as the gap steps down toward fractions of a decibel. There's nothing to configure: the test finds the smallest level change you can reliably detect just from your answers. Try the free free vocal range test for a quick, no-install way to check this yourself.

If you've ever wondered why a hearing clinic asks you to repeat words instead of just beeping tones in your ear, you've already brushed up against the idea behind a volume level discrimination test. This is the family of exams — clinically known as speech audiometry — that measures not just whether you can detect that a sound exists, but whether you can actually understand it once it's loud enough to hear.

Your results tell an audiologist far more about real-world listening than a tone-only exam ever could, because the ability to discriminate speech at a given volume is the single best predictor of how you'll function in noisy, everyday conversations. Along the way you'll see how the physics of sound and word recognition come together to produce the audiogram your clinician reviews with you, and why two people who "hear" the same tones can walk out of the same appointment with very different real-world outcomes.

What Is a Volume Level Discrimination Test?

A volume level discrimination test asks a simple question with a surprisingly layered answer: at what intensity, and with what accuracy, can you understand spoken words rather than merely notice that a sound occurred? Audiologists distinguish this from basic tone testing because two people with nearly identical pure tone audiometry results can have very different real-world comprehension. Use the speaker distance and delay calculator online any time you need a fast, accurate figure without doing it by hand.

One may follow a conversation across a noisy restaurant with ease; the other may struggle even in a quiet office. Speech audiometry exists to capture that gap. It is a core test in a clinician's test battery precisely because pure tone thresholds are a limited predictor of a person's ability to recognize speech, and improving someone's access to speech sounds is usually the entire point of fitting them with amplification in the first place.

Consider two patients walking into the same clinic with the same tone-test result: both can just barely detect a 40 dB tone across the board. On paper their hearing loss looks identical.

Yet when each is handed a list of words to repeat back at a comfortable volume, one scores 92% correct while the other scores 58%. That 34-point gap is invisible to a pure-tone test — it only shows up once you measure discrimination, not just detection — and it's the entire reason speech audiometry exists as a separate discipline rather than being folded into a standard tone test.

How Loudness Is Measured: Decibels and Frequency

Every discrimination test like this is built on two physical dimensions of sound: how loud it is and how high or low it is. Loudness is expressed in decibels (dB), and the scale spans a wide range of everyday experience:

• A whisper sits around 20 dB
• Ordinary conversation is roughly 60 dB
• City traffic or a vacuum cleaner runs 80–90 dB
• Loud music or a concert can reach 100–120 dB
• A jet engine at close range approaches 180 dB

Pitch is expressed as frequency, measured in Hertz (Hz); low bass tones fall in the 50–60 Hz range while high-pitched tones climb past 10,000 Hz. Normal hearing typically covers 250–8,000 Hz at 25 dB or lower.

Because the decibel scale is logarithmic rather than linear, small numeric jumps represent large jumps in actual sound energy — the reference sound intensity used to anchor the scale is defined at roughly 10^-12 W/m², so a rise from a quiet room to a loud room is a far bigger physical change than the numbers alone suggest. A jump from 40 dB to 70 dB isn't "a little louder," it's closer to a thousand-fold increase in acoustic energy. Every subsequent test in this article — from a basic audiometry test to full speech-in-noise testing — is really just varying these two dimensions and recording how well you cope as they change.

Speech Discrimination vs. Speech Detection Testing

It helps to separate two terms that sound alike but measure different things. Speech detection testing only asks whether you noticed that speech was present — the patient's job is to signal "I heard something," not to identify what was said. It's the most basic form of speech testing and correlates well with the pure-tone average, which is why its main clinical use is confirming an audiogram rather than diagnosing a comprehension problem. Speech discrimination, by contrast, requires you to correctly recognize the word itself.

This is the more clinically useful measurement, because a person can detect that sound is present while still misunderstanding a large share of the words spoken to them — the exact pattern that sends people back to the clinic confused about why "I can hear people talking, I just can't understand them." A patient reporting that exact complaint is one of the clearest clinical signals that a full speech discrimination test, not just a repeat tone test, is what's actually needed next.

How Audiologists Perform Speech Audiometry

Because languages and clinical conventions differ across countries, the exact mechanics of speech audiometry vary by clinic, but the underlying logic is consistent everywhere. Before any speech testing begins, the audiologist typically performs otoscopy and pure tone audiometry to rule out obvious contraindications — wax, debris, an outer-ear infection such as otitis externa, or a perforated eardrum can all distort speech-testing results if left unaddressed, and testing over an occluded canal risks measuring the blockage rather than the patient's actual hearing. The check whether your audio latency is working properly runs entirely in your browser via the Web Audio API — nothing is uploaded anywhere.

Words are then delivered through headphones, inserts, or a sound field loudspeaker setup, and the patient repeats each one back so the tester can score it correct or incorrect. The whole sequence is designed to move from the simplest, most objective measurement (can you detect a tone at all) toward the most functionally meaningful one (can you understand a word once it's audible).

Pure Tone Audiometry as the Diagnostic Starting Point

This baseline tone test measures the softest sound you can hear at each pitch and plots the result as an audiogram, with red circles marking the right ear and blue X's marking the left. It's the reference point every subsequent speech test is compared against.

If a patient's WRS turns out to be far worse than their pure-tone average would predict, that mismatch is itself diagnostic information — it can point toward a deeper sensory or neural problem rather than a simple loss of sensitivity. Audiometers capable of running both pure tone and speech tests are common in clinical audiology precisely so the two measurements can be taken back to back and compared directly, without switching machines or recalibrating between stages.

Masking the Non-Test Ear

Just as in tone audiometry, cross-hearing can occur in speech testing — a loud enough signal presented to one ear can be picked up faintly by the other, contaminating the result. Masking solves this by playing noise into the non-test ear so it can't "cheat" on behalf of the ear being measured. Because speech spans a wide range of frequencies, narrowband noise isn't an appropriate masker; audiologists use speech-shaped noise instead.

One widely cited formula for the masking level needed is:

$$\text{Masking Level} = D_S + \max(ABG_{NT}) - 40 + E_M$$

Here \(D_S\) is the dial setting used to present speech to the test ear, \(ABG_{NT}\) is the largest air-bone gap in the non-test ear between 250 and 4,000 Hz, the constant 40 reflects the minimum interaural attenuation for headphones (55 dB for insert earphones), and \(E_M\) is effective masking, which modern calibrated audiometers already account for automatically. As a worked example: if speech is presented to the test ear at a dial setting of 60 dB, the non-test ear's maximum air-bone gap is 15 dB, and effective masking is already built into the audiometer's calibration, the masking level works out to roughly \(60 + 15 - 40 = 35\) dB of noise fed into the untested ear — enough to keep it from contributing to the score without being so loud it bleeds back into the ear actually being measured.

Standard headphone or insert delivery isn't always possible. When the canal is blocked by wax, swelling, or a foreign object, presenting speech through air conduction risks measuring the blockage instead of the patient's true hearing, so the tester switches strategy rather than pushing ahead with a flawed setup.

One option is bone conduction delivery, where a small vibrating oscillator is placed against the skull behind the ear and speech is presented directly through bone rather than air, bypassing the blocked canal entirely. Another is a loudspeaker-based setup in a treated room, useful for patients who can't tolerate headphones — young children especially — or for aided testing where the point is to measure performance with a hearing device physically in place rather than removed for the exam.

Each delivery method has its own calibration requirements and its own limitations, and part of a tester's job is recognizing early which one actually fits the patient in front of them rather than defaulting to whichever setup is fastest to run. None of these adjustments change what's being measured conceptually — they simply make sure the measurement itself is clean before any interpretation of the result begins, which matters just as much for a first-time patient as it does for someone being retested years later for comparison.

Understanding Your Speech Discrimination Test Results: SRT, WRS, and the Audiogram

Once testing is complete, your results distill into a small handful of numbers that carry a lot of clinical weight. Two calculations dominate: the threshold at which you begin to recognize speech, and the percentage of words you get right once speech is comfortably audible.

Speech Recognition Threshold (SRT) Explained

The speech recognition threshold — sometimes called the speech reception threshold in older literature — is the intensity, in dB HL, at which you correctly identify 50% of the test material — typically two-syllable spondee words like "baseball" or "hotdog," chosen because their stress pattern makes them easy to recognize once audible. The SRT is usually measured first and then used to set the presentation level for the next stage of testing, commonly 20 to 40 dB louder than the SRT itself. A well-run test should find close agreement between your SRT and your earlier tone-test results; when the two diverge by more than about 12 dB, it's a signal that something beyond a simple sensitivity loss may be going on, and the clinician will typically flag the discrepancy for further investigation rather than accepting it at face value.

A Worked Example: Reading a Word Recognition Score

The word recognition score itself is calculated with a simple percentage formula:

$$\text{WRS} = \frac{\text{Words Correctly Identified}}{\text{Total Words Presented}} \times 100\%$$

Say your tester presents a standard 25-word list at a comfortable, elevated volume and you correctly repeat 22 of them. Your WRS is \(22 / 25 \times 100\% = 88\%\), which clinicians generally describe as excellent discrimination ability.

A score in the 70–80% range is typically labeled fair to good, and scores below 50% suggest that even amplified speech may remain difficult to follow without additional strategies such as lip-reading cues or assistive listening technology. Because the number of words in a list directly affects how precise the score can be — too few words and a single mistake swings the percentage dramatically, too many and the test drags on and patient fatigue starts to erode accuracy — most clinical word lists settle on somewhere between 10 and 25 items as a practical middle ground. Clinics don't build these lists from arbitrary vocabulary either — standardized lists such as CID-W22 use phonetically balanced words, meaning the mix of speech sounds across the list mirrors how often those sounds occur in everyday spoken language, so no single list is accidentally easier or harder than another simply because it leans on unusual consonant or vowel combinations.

Roll-Over and Signs of Retro-Cochlear Hearing Loss

In most people, WRS improves as volume increases and then plateaus. In some patients it actually gets worse at higher volumes, a pattern called roll-over, calculated as:

$$\text{Performance Index} = \frac{\text{Maximum Score} - \text{Minimum Score}}{\text{Maximum Score}}$$

Suppose a patient scores 88% at a moderate volume but only 60% once the volume is raised further. The resulting index is \((88 - 60) / 88 \approx 0.32\).

A performance index above roughly 0.4 is generally treated as a possible sign of retro-cochlear pathology — a problem further along the auditory pathway, beyond the sensory organ itself — and typically prompts additional testing such as reflex-based measurement or auditory brainstem response before a diagnosis is finalized. A milder result like the 0.32 example above wouldn't necessarily trigger alarm on its own, but a clinician would likely note it and watch for whether it recurs on repeat testing.

Speech-in-Noise Testing: Word Recognition Under Real-World Conditions

A patient can score beautifully on a quiet-room word recognition test and still struggle badly at a dinner party. That gap is why speech-in-noise testing exists as its own discipline within audiometry, dating back to some of the earliest speech-perception research in the 1940s. Difficulty understanding speech against background noise is consistently reported as the single most common complaint among people being treated for hearing loss, and it remains true even though regular speech-in-noise testing is, somewhat surprisingly, still performed in fewer than half of routine clinical evaluations.

Why Quiet-Room Scores Don't Tell the Whole Story

Two patients with nearly identical audiograms and nearly identical quiet-room word recognition scores can post wildly different results once background noise enters the picture — some tolerate a signal-to-noise ratio (SNR) of just a few dB while others need the target speech to be far louder than the noise around it before they can follow along. Researchers describe this using the signal-to-noise ratio at which a listener achieves 50% accuracy, often written SNR₅₀.

Because everyday environments — classrooms, restaurants, open offices — typically present SNRs between 0 and 15 dB, a patient with a poor SNR₅₀ can be functionally impaired in daily life even with an unremarkable quiet-room audiogram. Some researchers describe speech-in-noise testing as putting real pressure on the auditory system, closer to a stress test of hearing than a resting-state measurement, which is exactly why it surfaces problems a quiet-room test can miss entirely.

QuickSIN, WIN, and Other Common Tests

Several standardized tools exist for measuring this real-world listening difficulty, each with its own signal type, noise type, and testing protocol. The table below summarizes the tests most frequently used in clinical practice.

Notice that some of these tests use single words rather than full sentences specifically to strip out contextual guessing — a listener can often infer a missed word in a sentence from surrounding context, which inflates the apparent score, while an isolated word offers no such crutch. Sentence-based tests like QuickSIN, by contrast, more closely mirror natural, cognitively demanding listening, which is part of why some clinicians prefer them for older patients or anyone where cognitive processing speed is itself a variable worth understanding. Choosing between them is itself part of the clinical judgment involved in a full speech recognition testing workup, and the choice often comes down to how much test time is available and what specific question the referral is trying to answer.

How SNR Varies by Everyday Environment

One reason speech-in-noise scores matter so much in practice is that the SNR a person actually encounters swings wildly depending on where they are. In modern audiology, researchers have spent decades measuring real-world SNR across common settings, and the resulting picture explains a lot about why two patients with identical tone-test results report very different day-to-day struggles. A patient with hearing impairment who does fine in a quiet consultation room can still be functionally lost the moment ambient SNR drops:

• Quiet living room conversation: SNR of roughly +10 to +15 dB
• One-on-one conversation in a typical office: SNR around +5 to +10 dB
• Busy restaurant at a shared table: SNR near 0 to +5 dB
• Crowded classroom or open-plan workspace: SNR often below 0 dB
• Public transportation or a moving vehicle: SNR frequently negative, meaning the background noise is louder than the speech itself

Because so much of daily communication happens at an SNR well below what a quiet-room word list ever tests, a QuickSIN or WIN score often predicts real-world comprehension abilities far better than a quiet-room WRS does on its own. This is also why counseling matters as much as the raw numbers: a patient told their SNR loss is elevated benefits from concrete strategies — closer seating, reduced background noise at home, or directional hearing-aid microphones — rather than a number alone. Audiology as a field has increasingly leaned on this environmental framing precisely because it translates a clinical score into something a patient can actually act on.

The Anatomy Behind Every Hearing Test: How Sound Becomes Signal

To understand why volume and discrimination are measured separately, it helps to see how sound physically travels through your ear before your brain ever interprets it as speech.

Outer, Middle, and Inner Ear

Sound waves first enter the outer ear, funnel down the ear canal, and strike the eardrum, causing it to vibrate. Those vibrations pass through three tiny bones in the middle ear — collectively called the ossicles — the malleus, incus, and stapes, which mechanically amplify the signal before delivering it to the inner ear.

An evaluation that isolates a problem to this stage is generally describing conductive hearing loss: damage to the tympanic membrane or the ossicle chain that blocks sound from being efficiently transmitted, rather than a failure of the nerve itself. Because this pathway is mechanical rather than neural, conductive problems are frequently the most treatable category of hearing loss — sometimes resolved entirely with wax removal, a course of antibiotics for infection, or a minor surgical repair.

The Cochlea and Hair Cells

From the stapes, vibration reaches the cochlea, a fluid-filled spiral lined with microscopic hair cells. Their movement generates the nerve signals your brain ultimately interprets as sound and, further downstream, as recognizable words.

Damage at this level — or along the vestibulocochlear nerve that carries the signal onward — produces sensorineural hearing loss, the most common category seen in adult hearing clinics and the type most closely tied to poor word recognition scores even when pure tones are still detectable. Unlike similar sensory structures elsewhere in the body, these cells don't regenerate once damaged, which is exactly why prevention (limiting exposure to sustained loud noise) and early detection through routine testing matter so much for this category of loss.

What a Word Recognition Test Reveals About Hearing Loss Type

Because the two loss categories respond to very different treatments, distinguishing between them is one of the most practical outcomes of a full discrimination workup.

Sensorineural Hearing Loss vs. Conductive Hearing Loss

The severity scale audiologists use to describe either type generally follows the same bands:

• Normal: less than 25 dB HL
• Mild: 25–40 dB HL
• Moderate: 41–65 dB HL
• Severe: 66–90 dB HL
• Profound: greater than 90 dB HL

A patient with a conductive loss often shows a relatively flat, uniform reduction across frequencies and — critically — a WRS that improves substantially once volume is raised enough to overcome the mechanical blockage. A patient with a sensorineural loss frequently shows the opposite: even at a comfortably loud volume, discrimination remains poor, because the problem isn't insufficient volume but distorted transmission of the signal itself. This is exactly why testing discrimination outperforms a tone-only exam at distinguishing the two — turning the dial up alone doesn't fix a sensorineural problem the way it can partially compensate for a conductive one.

Cross-Checking for Non-Organic Hearing Loss

Occasionally, a patient's responses on tone testing don't match their behavior in conversation, or their results seem inconsistent from one session to the next. Audiologists use speech testing as an independent cross-check in exactly this situation — the principle being that no single test result should be accepted as final until it's confirmed by at least one other, unrelated measure. If speech recognition scores and pure-tone thresholds disagree in ways that don't fit any known pattern of conductive or sensorineural loss, it can point toward a non-organic component, meaning factors beyond the ear itself are influencing the results, and the clinic will typically repeat testing under different conditions before drawing conclusions.

Complementary Audiometry Tests: Middle-Ear and Reflex Exams

Speech-based testing is rarely used in isolation. A full audiometry evaluation typically pairs it with a handful of shorter tests that each rule in or rule out a specific mechanism of hearing loss.

Tympanometry and Acoustic Reflex Testing

Tympanometry checks for fluid or wax buildup, a perforated tympanic membrane, ossicle damage, or middle-ear tumors by measuring how it moves under varying air pressure in the ear canal — a soft plug placed in the ear changes pressure and tracks that movement. Acoustic reflex testing evaluates the involuntary muscle reflex triggered by loud sound, which in turn provides information about cranial nerve function along the brainstem pathway the signal travels. Tuning forks add a low-tech but genuinely useful cross-check: the Rinne test compares how a patient hears a vibrating fork through air versus through the bone behind the ear, while the Weber test checks whether a tone felt at the center of the forehead localizes to one ear — both help confirm whether a loss is conductive or sensorineural before more sophisticated testing begins. None of these tests carry any real risk; the biggest source of patient discomfort is usually just sitting still and quiet in a soundproof room for the duration of the exam, and most clinics complete the entire battery in a single visit.

What Your Discrimination Score Means for Hearing Aid Success

Beyond diagnosis, this kind of speech recognition test has a very practical downstream purpose: predicting how much benefit a patient is likely to get from a hearing device, and which type makes sense. Audiologists specifically point to speech-in-noise results — more than pure-tone thresholds alone — as the measurement most closely tied to whether a patient ends up satisfied with hearing aids months after the fitting.

Binaural Amplification and CROS Devices

When word recognition scores are similar between the two ears, binaural — both-ear — fitting is usually the straightforward recommendation. When one ear performs dramatically worse than the other, often due to a different underlying cause such as unilateral Meniere's disease, fitting both ears with standard hearing aid devices may not help and can even complicate things; audiologists instead consider monaural fitting or a contralateral routing of sound (CROS) system that routes sound from the poorer ear to the better one. Fitting decisions like these are exactly why the maximum WRS, not just the pure-tone thresholds, matters so much during a hearing aid evaluation — a device tuned to a threshold that doesn't reflect real discrimination ability can leave a patient technically "aided" but still functionally frustrated.

Cochlear Implant Candidacy

For patients who don't get enough benefit from conventional hearing aids, speech-in-noise and word recognition scores also factor into cochlear implant candidacy evaluations. Regulatory guidelines specify a minimum aided speech-perception score on an open-set sentence test before a patient is considered a candidate, and current recommendations call for testing to be completed both in quiet and in background noise so the evaluation reflects real-world listening conditions rather than an artificially easy best case. Post-implantation, the same tests are repeated to document the improvement and verify that the device is delivering measurable benefit, and comparing pre- and post-implant scores side by side gives both the patient and the clinical team a concrete way to track progress over time rather than relying on subjective impressions alone.

Realistic Expectations and Counseling

Numbers on a page rarely change behavior on their own, which is why the counseling conversation built around your results matters just as much as the testing itself. A low discrimination score doesn't necessarily mean a hearing aid will fail you — it means the device needs to be matched to a specific, known limitation rather than fitted on pure-tone thresholds alone. Patients who understand why their comprehension struggles in specific settings — a packed restaurant, a car with road noise, a video call with lag — tend to report higher satisfaction with their devices than patients handed a hearing aid with no explanation beyond "your hearing test showed a loss." Modern speech audiometry practice treats that explanation as part of the clinical deliverable, not an optional extra.

This is also where expectations get calibrated honestly. No amount of amplification restores discrimination ability that's been lost to sensorineural damage; what a well-fitted device does is make speech loud enough to reach whatever discrimination ability remains, and directional microphones or noise-reduction algorithms can meaningfully narrow the SNR gap in a noisy room without ever fully closing it. Framing the conversation this way — around a measured, specific score rather than a vague sense of "hearing loss" — is consistently what separates patients who stick with their devices from patients who leave them in a drawer within a year.

Preparing for a Speech Recognition Test: What to Expect and Who Needs One

If you're scheduled for a discrimination test, the preparation involved is minimal, and understanding the sequence in advance tends to make the whole appointment feel far less clinical and intimidating.

What Happens During the Test

There's no special preparation required — no fasting, no medication changes, nothing to bring except yourself and, ideally, a good night's sleep so you're not fighting fatigue on top of the listening task itself. An audiologist will seat you in a soundproof room, place earphones over or in your ears, and ask you to repeat words or raise a hand each time you detect a sound.

The whole process, tone testing and speech testing combined, usually takes well under an hour. You'll typically be asked to raise your left hand for sounds in your left ear and your right hand for sounds in your right ear, or press a response button, so the tester can record the softest level at which each ear detects and then correctly identifies each stimulus.

After the Test: Talking Through Your Results

Once testing wraps up, the results are typically reviewed with you the same day or shortly after, along with a copy sent to whichever physician — often an otolaryngologist or, less commonly, a neurosurgeon for cases involving a tumor near the auditory nerve, such as an acoustic neuroma — referred you for testing in the first place. This is the point where your audiologist walks through your audiogram, SRT, and WRS together and explains what pattern they form, rather than leaving you to interpret a page of numbers alone. It's a good moment to ask direct questions: whether your result is conductive or sensorineural, whether hearing aids are likely to help given your specific discrimination score, and whether any follow-up testing — speech-in-noise or an imaging referral — is recommended based on what showed up.

Special Populations and When to Ask for Speech-in-Noise Testing

Certain patients benefit from a modified approach. Young children and adults with cognitive impairment may need simpler word lists, since complex vocabulary can understate their true hearing ability rather than reflect it.

Patients who report real trouble following conversation despite normal-looking test results are increasingly recognized as a distinct population — sometimes described under the umbrella of auditory processing difficulty — and for them, standard quiet-room testing can miss the problem entirely, which is exactly why requesting speech-in-noise testing specifically, rather than settling for a basic tone test, can be the more revealing choice. It's also worth flagging tinnitus or vertigo to your provider before testing begins, since either can subtly affect how you respond during the exam and both may point toward related conditions worth investigating alongside your hearing itself.

How Often You Might Need Retesting

There's no single universal interval — how often you should repeat testing depends heavily on why you were tested in the first place, what the initial results showed, and whether anything about your daily listening environment or general health has changed since your last visit. Someone establishing a baseline with no known risk factors and no reported symptoms might not need another full evaluation for several years.

Someone monitoring a progressive condition, recovering from ear surgery, or tracking noise exposure on the job typically gets retested annually, sometimes more often if results have been trending in one direction. Anyone newly fitted with a device benefits from a follow-up evaluation a few weeks to a few months after the initial fitting appointment, once they've had time to adjust to daily use, so the clinic can confirm the fitting is actually delivering the benefit it was calculated to deliver rather than assuming it based on the initial numbers alone.

Equipment itself matters here too, even though it rarely gets mentioned to patients directly. Just as an audiometer's tone-generating components require periodic calibration against a known reference, the calibration behind speech-testing material — whether it's delivered live, from older recorded media, or from digitally stored files — has to be checked on a routine schedule so that a "60 dB" word list actually plays at 60 dB and not something quietly drifted from it.

A patient's results are only as trustworthy as the equipment that produced them, which is part of why testing at an accredited clinic with documented calibration practices is worth asking about if you're choosing between providers. It's a reasonable question to ask directly during scheduling, and a clinic confident in its own process won't hesitate to answer it in plain terms rather than deflecting.

Whatever your starting point, testing how you discriminate speech at different volumes remains the most direct way to find out not just whether you can hear sound, but whether you can actually understand it — which, in the end, is the entire point of hearing in the first place. The broader field of audiology continues to refine these tools, but the core logic your test relies on today has held up for decades because it measures the thing that actually matters to patients: comprehension, not just detection.