If you have moderate-to-severe hearing loss and music sounds squashed, metallic, or just *wrong* through your hearing aids—even though speech is reasonably clear—the culprit is almost certainly your device’s compression settings. The fundamental problem is this: a normal-hearing ear comfortably handles a dynamic range of about 120 decibels, from barely audible rustling to uncomfortably loud. Recruitment typically shrinks that range to 40 dB or less, and wide dynamic range compression (WDRC) attempts to squeeze the entire audible world into that narrow window. The physics are straightforward. The fitting decisions that follow are anything but.
I’ve sat across from hundreds of musicians, audio engineers, and music lovers over the past two decades, and the conversation usually starts the same way: “Speech is fine, but when I put on my favourite record, it’s like listening through a transistor radio.” They’re not imagining it, and it’s not because they bought the wrong brand. It’s because the default speech-optimised compression that makes consonants audible in conversation actively mangles the dynamic structure of music.
What Recruitment Actually Does (and Why WDRC Exists)
Loudness recruitment is a consequence of outer hair cell damage in the cochlea. In a normal ear, outer hair cells amplify soft sounds and compress loud ones—a biological automatic gain control. When those cells are damaged, soft sounds stay inaudible, but loud sounds rapidly reach normal (or even excessive) loudness. The practical upshot: your threshold might sit at 60 dB HL, but your uncomfortable loudness level (UCL) might be only 100 dB HL. That 40 dB range between “I can just hear it” and “turn it off” is the entire palette your hearing aid has to work with.
Wide dynamic range compression exists to map the 120 dB range of real-world sound into your reduced dynamic range without causing discomfort or losing audibility. The hearing aid monitors incoming sound level and applies gain inversely: soft sounds get a lot of amplification, moderate sounds get less, and loud sounds get very little or none. The goal is noble. The execution is where things get messy, especially for music.
The Four Knobs That Matter: Kneepoint, Ratio, Attack, and Release
WDRC is governed by four key parameters, and every audiologist adjusts them—whether they realise it or not—every time they accept or override a first-fit formula.
Kneepoint is the input level at which compression starts. A low kneepoint (say, 40 dB SPL) means compression is active almost all the time. A high kneepoint (65 dB SPL) leaves more of the signal uncompressed. Music benefits from higher kneepoints because it preserves dynamic contrast in the louder passages.
Compression ratio defines how aggressively the aid reduces level increases above the kneepoint. A 3:1 ratio means that for every 3 dB increase in input, you get only 1 dB increase in output. Ratios for speech typically sit between 2:1 and 3:1; ratios for severe recruitment can climb to 10:1 or more. High ratios compress the life out of music.
Attack time is how quickly the compressor responds to a sudden increase in level. Fast attack times (1–5 milliseconds) clamp down on transients instantly, preserving audibility of soft consonants in running speech. Slow attack times (20–100 ms) let the initial impact of a drum hit or piano note come through before gain reduction kicks in.
Release time dictates how quickly gain returns after the sound drops. Fast release restores gain quickly, which is great for the gaps between words. It also pumps and breathes audibly when music plays—classic “compressor distortion” that recording engineers spend careers trying to avoid.
Fast Versus Slow Compression: The Gatehouse Problem
Here’s where evidence and clinical intuition collide. A landmark 2006 study by Gatehouse, Naylor, and Elberling found that cognitive capacity—specifically working memory and processing speed—predicts whether a patient does better with fast-acting or slow-acting compression. Patients with *better* cognitive function performed best with slow compression, which preserves the natural amplitude envelope of speech and allows the brain to use those cues for segregating talkers and understanding in noise. Patients with reduced cognitive capacity did better with fast compression, which delivers a more consistent, preprocessed signal.
In clinical practice, this means one-size-fits-all programming is nonsense. The default NAL-NL2 or DSL prescription might choose fast compression because it optimises audibility across the broadest population. But for a 68-year-old jazz pianist with intact cognition and recruitment, that fast compression will destroy the dynamic phrasing that makes music intelligible. We have the tools to slow it down; the question is whether the fitting audiologist takes the time to ask what matters to the patient.
Music’s Unique Problem: Crest Factor and Input Clipping
Speech has a relatively low crest factor—the difference between average level and peak level—typically around 12 dB. Music, especially orchestral or percussive music, can have crest factors of 20 dB or more. When a cymbal crash hits the microphone, the instantaneous level can be 20 dB higher than the average level of the passage.
Older hearing aid analogue-to-digital converters clipped at around 90–95 dB SPL input. Marshall Chasin, a Canadian audiologist who has written extensively on music and hearing aids, documented this in detail throughout the 2000s. If a musician is playing live or sitting in front of a loudspeaker, peaks can easily exceed 100 dB SPL. The result isn’t just compression—it’s hard clipping, a form of distortion that adds harsh, non-musical harmonics. Patients describe it as “breaking up” or “scratchy.”
Newer chipsets from the major manufacturers (we’re talking post-2015 platforms) offer input dynamic ranges of 105–113 dB SPL, and some include dedicated music programs that raise the clipping point, widen the kneepoint, slow the attack and release, and reduce or disable noise reduction. These are not marketing gimmicks; the specs matter, and the programming flexibility matters even more.
What a Music Program Actually Changes
When we create a dedicated music program in the clinic, here’s what typically changes from the default speech program:
– Kneepoint raised to 55–65 dB SPL, so more of the signal stays in linear amplification.
– Compression ratio reduced to 1.5:1 or 2:1, preserving dynamic contrast.
– Attack and release times lengthened—often 30–50 ms attack, 200–500 ms release—to let transients through and avoid pumping.
– Noise reduction disabled or minimised, because algorithms designed to suppress babble will also suppress instrumental overtones, backup vocals, and reverberation.
– Microphone mode set to omnidirectional, because directional mics are optimised for speech directivity, not the diffuse soundfield of a concert hall or your living room.
These changes are not available on every device, and they’re certainly not automatic. If your hearing aids didn’t come with a music program, or if your audiologist didn’t discuss your listening habits during the fitting, you’re probably trying to listen to Coltrane through a speech processor.
What You Can Do (and What Your Audiologist Should Offer)
If music sounds wrong, book a follow-up appointment and be specific. “Everything sounds bad” gives us nothing to work with. “The piano sounds tinny, and when the whole band comes in, it distorts” tells us exactly where to look. Bring the track that sounds worst on a phone or tablet; we can play it in the clinic and adjust in real time.
Ask whether your device supports a dedicated music program, and whether the compression settings can be adjusted independently of the speech program. If you’re a performing musician or regularly attend live music, ask about devices with high input limits—spec sheets will list “input dynamic range” or “maximum input level,” and you want 105 dB SPL or higher.
If you’re considering new devices and music matters to you, **tell your audiologist during the needs assessment**, not after you’ve already chosen and paid for a model optimised exclusively for speech. Some manufacturers offer better music handling than others, and some product tiers within a manufacturer’s range have more flexible compression parameters. We’re brand-agnostic here, but we’re not parameter-agnostic.
Fitting Decisions Matter as Much as Technology Tier
I’ve seen premium, top-tier devices sound appalling for music because the audiologist accepted the default first fit and sent the patient home. I’ve also seen mid-tier devices sound remarkably good because someone took the time to measure recruitment properly with loudness scaling, programmed a thoughtful music program, and followed up. The technology ceiling matters, but most patients never get close to it because the fitting is lazy.
Wide dynamic range compression is an elegant solution to a hard problem. But elegance in theory doesn’t guarantee good sound in practice. The closer you look at the parameters—kneepoint, ratio, attack, release—the more you realise that *every* fitting is a series of compromises, and those compromises should be informed by your goals, your residual dynamic range, and your cognitive profile, not just by what the manufacturer’s algorithm chose in the moment you pressed “calculate.”
If your hearing aids make speech clear but music unbearable, you’re not asking too much. You’re just asking your audiologist to go past the first fit.
