Channel Vocoder

Hear the world
through a cochlear implant

A cochlear implant squeezes rich sound into just a handful of "channels." Drag the channel count and use your own ears to feel why fewer channels make speech blur.

How it works
01 / INTERACTIVE

Vocoder simulator

Pick a sound — a built-in sample, an uploaded file, or your live microphone. Toggle "original / processed" to compare, drag the channel count, and watch each frequency channel light up. Headphones are recommended, especially in microphone mode to avoid feedback.

Status: Ready
8 channels · noise carrier
Lowfrequency channels (apex → base)High
CH 8 BAND 80–8000 Hz ENV 160 Hz SPREAD 0% NOISE 0% SR
CLARITY
Intelligibility index
An illustrative "speech clarity" estimate based on the current settings; it updates live as you adjust them — the higher the value, the closer to understanding continuous speech. For teaching only, not a real measurement.
Spectral detail
Pitch cues
Noise margin
Effective channels
8
A typical cochlear implant has 12–22 electrodes, but current spread between channels means the "effective" number is often only around 8.
Each scenario automatically combines channel count, frequency range, noise, and crosstalk, then plays a matching sample.

Lower is closer to a real implant (pitch detail is lost); higher keeps more timing/pitch cues.
Simulates current leaking between neighboring electrodes: the more crosstalk, the fewer "effective" channels — even if the electrode count is unchanged.
Add ambient noise to feel why recognition gets harder in noisy settings.
Renders offline with all current settings and downloads · microphone source not supported
02 / PRINCIPLE

How it "hears"

A healthy ear has tens of thousands of hair cells arranged along the cochlea as a fine frequency map. A cochlear implant bypasses the damaged hair cells and electrically stimulates the auditory nerve with a row of electrodes — and the information it can carry is essentially what this simulator does: slice sound into a few frequency bands and keep only each band's "loudness envelope."

STEP 01

Split into bands

A microphone captures the sound and splits it by frequency into several bands, each mapped to one place on the cochlea and one electrode.

STEP 02

Extract envelope

Each band keeps only the "envelope" of how its energy rises and falls over time, discarding the fine waveform and pitch detail.

STEP 03

Resynthesize

The envelope modulates a carrier (noise or sine) to refill each band.

STEP 04

Stimulate / play back

A real device turns this into electrical pulses on the auditory nerve; the simulator sums the bands and plays them for you.

SIGNAL PIPELINE Live demo
Sound enters from the left, is split into bands, has its loudness envelope extracted band by band, then is resynthesized with a carrier — exactly what happens inside the simulator.
Base · high freq Apex · low freq Closer to the base→ higher frequency
The cochlea's "tonotopic map": different places handle different frequencies — which is why electrodes are arranged by frequency.
0Healthy cochlea

About 15,000 hair cells form an extremely fine frequency gradient along the cochlea, spreading sound out continuously.

0Cochlear implant

Usually just 12–22 electrodes, and current spread cuts the "effective" channels down to about ~8.

Tens of thousands of fine frequency points compressed into a dozen-odd electrodes: this is the core gap the "channel vocoder" recreates.
03 / JOURNEY

The journey of a single "hello"

From a sound wave in the air to the moment the brain "hears" it, sound passes through several stages. The pipeline below runs in real time — follow the dots and see where the implant takes over work that once belonged to the ear.

SOUND · SIGNAL · SPIKES · SENSE
🎙️ Sound air vibration 🎛️ Processor split · envelope Electrodes electrical pulses 🧠 Nerve → brain learns to decode
Green dots = sound information. At the "processor" it is compressed into a few channels, then sent to the auditory nerve as electrical pulses — once information is lost here, it can't be recovered downstream.
04 / LIMITS

What a cochlear implant can't do

The cochlear implant is a remarkable technology that lets many people follow conversation in quiet. But what it delivers is a "downsampled" version of sound — only a few channels, only loudness envelopes. These are its common weak spots. You can hear each one for yourself in the simulator above.

🎵

Music & pitch

Melody and harmony rely on fine pitch information, which the vocoder mostly throws away. Rhythm still comes through, but tunes sound flat and distorted.

Tonal languages

Mandarin and Cantonese distinguish meaning by pitch contour (mā/má/mǎ/mà). Implants encode pitch poorly, so tone recognition is a real challenge for children learning these languages.

🗣️

Noisy places

With background noise or many people talking at once, recognition drops sharply — too few channels make it hard to separate the target voice from the noise.

🎯

Sound localization

Telling which direction a sound comes from needs fine timing and level differences between the two ears. A single implant can barely localize, and even bilateral ones fall well short of normal hearing.

🧠

Relearning required

It doesn't "restore" the old hearing — it delivers a brand-new, simplified signal. The brain relies on plasticity and long-term training to gradually "read" it.

Electrode–nerve interface

Current from neighboring electrodes spreads and overlaps; combined with "dead regions" and differences in insertion depth, the truly independent channels are far fewer than the electrode count.

05 / CHALLENGE

Listen & guess: how many channels?

Listen first, then guess how many channels were used. Play and answer anonymously — no account needed.

This round · channels
?
Loading challenge…
Your guess
Press play on the left, listen carefully, then pick the channel count you think it was.

Score 0 / 0  · 
06 / READ-ALOUD

Read-aloud learning

Browse published read-aloud cards and play cochlear-simulated audio from the same content library as the mini program. Clips are already processed — the site plays them as-is, without running the vocoder again.

07 / CHILDREN

Deaf children: why "early" matters so much

A baby's brain is in a golden window for learning language. The earlier and richer the auditory input, the better later spoken-language and literacy development tends to be. That is why newborn hearing screening and early intervention are emphasized again and again.

~1–3 ‰
Roughly 1–3 in every 1,000 newborns have some degree of congenital hearing loss
1 · 3 · 6
A widely used principle: screen by 1 month, diagnose by 3 months, begin intervention by 6 months
< 12 mo
When appropriate, earlier implantation (often before age 1) tends to favor spoken-language development
Birth – 1 mo

Newborn hearing screening

Complete initial/repeat screening before discharge or by one month to catch possible hearing issues early.

By 3 months

Diagnostic evaluation

More thorough audiological testing clarifies the type and degree, informing the plan.

By 6 months

Begin intervention

Fit hearing aids and start auditory–verbal training; for severe/profound loss with limited hearing-aid benefit, evaluate for a cochlear implant.

Ongoing

Therapy & family involvement

Speech therapy, everyday family interaction, and educational support are the long, crucial part — the device is only the beginning.

More than one path

Families facing hearing loss can make different choices: some follow a spoken-language, auditory–verbal path, some choose sign language, and some use a bilingual-bimodal approach. A cochlear implant is not the only "right" answer, and the Deaf community has its own language and cultural identity. Respecting the values of the child and family, and communicating fully with a professional team, usually matters more than arguing over "which path is better."

Support beyond the device

Wireless remote microphones, classroom FM systems, captions, and visual cues all improve access in real settings (especially noisy classrooms) beyond the device itself. Peer acceptance and a strong sense of self matter just as much to development.

A note for parents: the simulator above can help family, teachers, and classmates get a sense of "what the child might be hearing," and bring a little more understanding and patience. But every child's real experience differs; the simulation is only an approximation and cannot replace professional assessment.