⚙️ Apps — how they are built

Three apps in three months — one engine underneath

“We built one thing properly, and three apps came out of it.” April, May and June of 2026, one release a month — and the reason was not overtime. The apps were sharing a heart.

⚙️The engine — the same sound, only clearer

The first thing to go as hearing dulls is the consonants: the f, s and th of English, and the sharp edges of syllables in every other language too. The engine plays uncompressed WAV at raised speed and adds processing that holds those consonants together, alongside elements taken as inspiration from Tomatis’ work on high-frequency stimulation and from the 40 Hz gamma-band research (MIT, Nature, 2016). No special hardware and no special file — the voice you already listen to, out of whatever speaker you have. The core of the engine is at patent-application stage.

🎛️Off by default — how you handle a technology still being tested

The 40 Hz element ships switched off in every one of the apps, and only somebody who wants it turns it on. We are not going to push a feature whose effect we cannot state plainly onto everyone who installs. BrainHz took its inspiration from work on hearing; it is not a medical device and it claims to prevent or treat nothing. What people actually report is closer to “fast speech sounds a good deal clearer” — and even that varies from one listener to the next.

🚪The hard part was never the code

Once the engine existed, making an app became a question of what you send through it: fast English for people who need fast English, a hearing course for an ear that has gone quiet, scripture for people who want the text. What actually ate the months sat outside the code — the paperwork of two app stores, review (Sound Power came back rejected five times before it passed), and payments, which we are still working through. When AI knocked down the wall of writing code, it became obvious where the real walls had been standing all along.

🤱 And then a fourth — MamaTalk

The fourth thing built on the engine runs in the opposite direction. Nothing is shaved and nothing is sped up; instead the server switches off the bands that should not reach a foetus or a newborn. MamaTalk plays English fairy tales and the court music of Korea’s National Gugak Center as uncompressed originals, for people who are expecting. One engine, four doors.

⚙️The four doors

Different material each time, the same engine underneath.

🎧
Speed EnglishKorean speed listening and shadowing from 2× to 4×, every file uncompressed. The app is in Korean.Get it on Google Play ↗
🎧
Sound PowerMeasure your hearing age first, then train the range that has gone quiet.What it does ↗
🎧
Speed BibleScripture at speed in three voices, with subtitles in seven languages.What it does ↗
🎧
MamaTalkThe same engine turned around — nothing sped up, some bands turned off.What it does ↗

Source

Adapted from parts four and five of Oh Seung-jong, The 30-Day Wave Speed-Listening Miracle (Purple, 2026). The gamma-band reference is Iaccarino et al., Nature 540 (2016), an animal study that the engine takes as inspiration and nothing more. On why the source file has to be uncompressed before any of this matters see [[lossless-audio-guide]] and [[cognitive-load-numbers]]; on the training the engine was built for, [[wav-speed-listening]]. — Not carried over from the Korean edition: its links to the Korean app pages, the Korean partnership enquiry page and the Korean-language book listing, and its pointers to the three individual app stories, which have no English editions yet.

Read the original →

※ This article summarizes and restructures the key points of the source above.

📥 30일 영어 귀뚫기 로드맵 — 무료 PDF

읽은 내용을 오늘부터 실행하는 훈련표 · 청각 나이 측정법 · 주차별 배속 가이드

The BrainHz engine — one engine behind four listening apps | BrainHz