/insights

Deep dives, not blog posts.

Long-form thinking on audio AI, mix feedback, accessibility, voice infrastructure, live workflows, and the engineering choices behind them. We publish when we have something to say, not when the calendar says to.

01 Featured · pillar

Pillar piece Brand 22 min read Jun 3, 2026

What is audio AI? A working map of the field in 2026

The honest, plain-language map of where AI is actually changing audio in 2026 — and where it is still being oversold. Covers mix feedback, voice infrastructure, accessibility, audio intelligence, live audio, and learning.

Read the full map

02 13 more

VoiceLab

AI voice infrastructure in 2026 — what actually works

The honest map of voice infrastructure in 2026 — TTS, ASR, voice cloning, dubbing, real-time conversion, QA tooling. What ships in production, what is overpromised, and where the unbuilt opportunities sit.

Jun 5, 2026

24 min
Engineering

Android audio development in 2026 — the engineer guide

A working engineer guide to building serious audio on Android in 2026 — AAudio, Oboe, JUCE, latency budgets, hearing-device routing, AudioWorklet-equivalents, and the parts of the platform that finally just work.

Jun 5, 2026

22 min
HearLab

Hearing accessibility on Android — the platform map in 2026

The platform-level state of hearing accessibility on Android in 2026 — ASHA, LE Audio, Auracast, Live Caption, hearing-aid routing, captions in the browser. What ships, what does not, and where the unbuilt opportunities sit.

Jun 5, 2026

20 min
MixLab

AI mastering tools, honestly compared (2026)

A working engineer's comparison of LANDR, eMastered, BandLab, iZotope Ozone AI, and CloudBounce. What they do well, where they fall down, and which to use for which job.

Jun 4, 2026

14 min
SignalLab

Evaluating audio AI models: a working guide to measurements that matter

How to measure whether an audio AI model is actually good — subjective listening tests (MUSHRA, ABX, MOS), objective metrics (PESQ, STOI, SI-SDR), and the gap between the two. The practical playbook for product teams.

Jun 4, 2026

19 min
Engineering

Real-time audio in the browser: what 2026 actually makes possible

The honest 2026 floor and ceiling for browser audio — AudioWorklet latency, WebGPU inference, MediaDevices capture, the gaps that still exist vs native, and the gaps that have finally closed.

Jun 4, 2026

20 min
VoiceLab

Voice clone detection in 2026: what works, what doesn’t, what’s coming

A working overview of how synthetic-speech detection has evolved against modern voice cloners. Spectral artefacts, prosody fingerprints, watermarking, the regulatory landscape, and the honest limits of current detectors.

Jun 4, 2026

21 min
MixLab

Building MixLab Analyzer — BS.1770-4 LUFS in the browser

A working engineer’s walkthrough of how MixLab Analyzer works in the browser — K-weighting biquads, FFT, true peak, stereo correlation, and the design choices behind keeping it all client-side.

Jun 2, 2026

18 min
HearLab

The accessibility audio market is bigger than people think

There are 1.5 billion people with measurable hearing loss. Most products in the audio space are not designed for them — and that is a strategic mistake.

May 25, 2026

9 min
Engineering

WebAudio vs native — where the line actually is in 2026

WebAudio has crossed the threshold from "demo only" to "production for many use cases". Here is where it still falls short and where it’s now the right tool.

May 20, 2026

11 min
MixLab

Why AI mastering plateaued — and what creators want next

The first wave of AI mastering proved the category. The second wave has to earn it. A look at what readable feedback could replace black-box "enhance".

May 12, 2026

8 min
HearLab

Designing hearing support without medicalising it

A non-medical companion app for hearing loss has to thread several needles at once: regulatory, social, technical. Here is the framing we use.

Apr 29, 2026

12 min
Engineering

Realtime DSP on Android — what AAudio gets right

AAudio + AudioWorklet is now a credible realtime audio stack. A short tour through the rough edges and the parts that just work.

Apr 8, 2026

15 min

Deep dives, not blog posts.

What is audio AI? A working map of the field in 2026

AI voice infrastructure in 2026 — what actually works

Android audio development in 2026 — the engineer guide

Hearing accessibility on Android — the platform map in 2026

AI mastering tools, honestly compared (2026)

Evaluating audio AI models: a working guide to measurements that matter

Real-time audio in the browser: what 2026 actually makes possible

Voice clone detection in 2026: what works, what doesn’t, what’s coming

Building MixLab Analyzer — BS.1770-4 LUFS in the browser

The accessibility audio market is bigger than people think

WebAudio vs native — where the line actually is in 2026

Why AI mastering plateaued — and what creators want next

Designing hearing support without medicalising it

Realtime DSP on Android — what AAudio gets right