Technical Guide

How VibeDowns Studio Works

The architecture of browser-based audio processing, technical flow of every tool, and security layers.

Architecture

System Architecture

All audio processing happens in the browser. The server is used only for pages and analytics.

User
Browser
Security Layer
MIME + Magic Bytes
Web Audio API
AudioContext
BPM Engine
Spectral Flux + Comb
Stem Separator
STFT + Wiener Filter
Parametric EQ
BiquadFilterNode
Export Engine
WAV / MP3 / OGG
Local Storage
Preferences
Server
No audio data
Data Flow
User──→Security LayerUpload file
Security Layer──→Web Audio APIValidated
Web Audio API──→BPM EngineAudioBuffer
Web Audio API──→Stem SeparatorAudioBuffer
Web Audio API──→Parametric EQAudioNode chain
BPM Engine- - →Export EngineBPM result
Stem Separator──→Export EngineStem buffer
Parametric EQ──→Export EngineProcessed audio
Export Engine- - →UserDownload
User- - →Local StoragePreferences
Security Layer- - →ServerAnalytics only
Tools

Technical Flow of Every Tool

Every tool runs entirely in the browser. Below is the algorithm and data flow for each one.

📡
BPM Detector
Analysis
Open →

Downmixes to mono, computes energy flux via Spectral Flux, scans candidate tempos with a Comb Filter Bank, and picks the strongest period.

1File → Mono mix
2Compute energy windows
3Comb filter scan
4Pick BPM + confidence
Spectral FluxComb Filter BankBeat TrackingOfflineAudioContext
🎤
Vocal Remover
Processing
Open →

Converts to time-frequency domain via STFT, estimates source masks per frequency bin, and applies Wiener soft masking to cleanly separate 5 stems.

1AudioBuffer → STFT
2Compute spectral masks
3Apply Wiener filter
4ISTFT → 5 stems
STFT (Cooley-Tukey FFT)Wiener FilterHann WindowOverlap-Add ISTFT
Audio Speed & Pitch
Processing
Open →

Changes speed via playbackRate, preserves pitch with preservesPitch. For independent pitch shift, combines OfflineAudioContext + playbackRate.

1Load file
2Set speed/pitch params
3OfflineAudioContext render
4WAV/MP3/OGG export
HTMLMediaElement.playbackRatepreservesPitchOfflineAudioContextWeb Audio API
🎛️
Audio Lab
DAW
Open →

A full browser DAW. Runs a BiquadFilter chain, ConvolverNode reverb, DelayNode echo, AnalyserNode meter, and OfflineAudioContext render pipeline on a single AudioContext.

1File → AudioBuffer
2EQ/Delay/Reverb chain
3Meter analysis
4Selection render → Export
BiquadFilterNodeConvolverNodeDelayNodeAnalyserNodeOfflineAudioContext
⏱️
BPM Converter
Timing
Open →

Converts BPM to milliseconds: ms = (60000 / BPM) × note_multiplier. 14 note values including dotted and triplet variants for delay and reverb pre-delay calculations.

1Enter BPM
2Compute 14 note values
3Display ms + Hz
4Copy
Pure mathReact stateNo audio processing
🥁
Kick/Snare Analyzer
Mixing
Open →

Compares kick and snare frequency bands, detects overlap zones, and visually shows which instrument should take priority at each frequency.

1Select preset
2Compute frequency bands
3Overlap analysis
4Show recommendations
Frequency band mathOverlap detectionSVG visualization
↔️
Stereo Width
Mixing
Open →

Performs Mid-Side decomposition: Mid = (L+R)/2, Side = (L-R)/2. Side/Mid ratio gives stereo width. Mono compatibility score is derived from Side energy vs total energy.

1Input L/R signal
2Compute M/S
3Width score
4Show mono compat
Mid-Side mathCanvas visualizationLissajous display
✍️
Lyric Toolkit
Creative
Open →

Runs entirely in-browser with no API. It uses syllable heuristics, sound-similarity rhyme checks, and rule-based writing notes.

1Write lyrics
2Syllable/rhyme analysis
3Style hints
4Generate snippets
Syllable heuristicsSound similarity scoringRhyme bank lookupGrammar rules
Security

Privacy & Security

Your audio files are safe. Here's how.

🔒
Audio never leaves your browser

Your audio file is processed only in browser memory. Not a single byte is sent to a server.

🛡️
MIME + Magic Bytes validation

Every file is checked for MIME type and magic bytes before processing. Foreign formats and malicious content are blocked.

Rate limiting

Client-side rate limiting prevents excessive processing in a short time window.

🍪
Consent-based cookies

Except for essential technical cookies, analytics and advertising cookies activate only with your consent. Audio file contents are never sent to the server.

Technology

Web Audio API

The foundation of every tool is the W3C standard Web Audio API.

AudioContext
Audio processing environment
AudioBuffer
Raw audio data
AudioNode
EQ, Delay, Reverb...
AnalyserNode
FFT analysis
AudioDestination
Output / Export

OfflineAudioContext renders audio independently of real-time playback. This makes export processing much faster than real-time.

Try it now

No install, no account. Open your browser, drag your file.