Production Masterclass

Understanding the Process

Mix & Master
Finale Master
Stem Separation
01

Phase 1: Stem Preparation

Before uploading, your audio must be prepared for the AI Engine. Think of this as giving the chef the best possible ingredients.

What You Do

Export each instrument (Kick, Vocal, Bass, etc.) as individual WAV files. Ensure no track is "Clipping" (hitting the red/0dB).

Best Practice

Export "Dry" stems (without reverb/delay) if you want the AI to manage the spatial depth, or "Wet" if you love your current effects.

Phase 1: Mix Submission

For Finale Mastering, we require your final stereo mixdown. This is the culmination of your entire creative process.

The Mixdown

Submit a high-resolution 24-bit or 32-bit WAV file. Avoid using limiters on your master bus to leave "Headroom" for the AI.

Headroom Target

Try to keep your peaks around -3dB to -6dB. This gives the AI surgical room to enhance your dynamics without distortion.

Phase 1: Source Selection

For Stem Separation, you need a high-quality stereo source. The AI needs to "hear" the individual instruments clearly within the mix.

Original File

Use lossless files (WAV/FLAC) whenever possible. MP3s contain compression artifacts that can confuse the neural network.

Mix Clarity

A cleaner mix always yields better separation. Heavily distorted or "muddy" mixes are harder for the AI to de-layer accurately.

What is going on? The engine expects uncompressed audio. High-resolution audio allows the AI to perform "surgical" processing moves without introducing noise.
02

Phase 2: AI Ingest & Analysis

Once you drop your files, the engine begins its most complex task: listening.

The Auto-Tagger

The system scans your filenames. If it sees "Vocal", it assigns a vocal-specific AI processing chain.

Spectral Profiling

The AI analyzes the "DNA" of each stem—finding where the frequency energy sits and identifying potential "masking" issues.

Phase 2: Loudness Optimization

Mastering is about translation—ensuring your song sounds incredible on everything from a iPhone to a festival stack.

Target LUFS

Choose -14 for Spotify/Apple Music compliance, or push to -8 for aggressive, competitive club loudness.

Dynamic Preservation

The AI uses "Intelligent Limiting" to reach your volume target without "squashing" the life out of your transients.

Phase 2: Demucs V4 Neural Engine

Our extraction engine uses a state-of-the-art U-Net architecture to isolate signals in the frequency domain.

Waveform Source

The AI doesn't just filter EQ; it predicts the original waveform of each instrument based on trained audio patterns.

Artifact Reduction

A second AI pass cleans up "Bleed" (e.g., drums in the vocal mic) to ensure studio-grade isolation.

What is going on? The AI is performing a FFT analysis to understand the harmonic content. It’s building a virtual 3D map of your sonic energy.
03

Phase 3: The Genre Matrix

Your choice of genre tells the AI what the "Gold Standard" for your record should be.

Sonic Blueprints

A "Hip Hop" master needs heavy sub-bass. A "Rock" master needs mid-range grit and drum weight.

Spectral Matching

The AI compares your record to thousands of top-tier masters in your genre to identify tonal gaps.

What is going on? The engine loads a "Reference Style Profile" containing data points from legendary records. It uses these as spectral targets for its EQ curves.
04

Phase 4: The Intelligent Mixdown

This is where the magic happens. The AI acts as your personal mixing engineer.

Spatial Placement

The AI places instruments in a virtual room. Kick stays centered; Synths move wide for immersion.

Vocal Anchoring

If you choose "Center", the vocal is locked to the front. "Wide" creates a cinematic, lush vocal spread.

Phase 4: Mastering Profile

Select the "Soul" of your master. Every record needs a different emotional weight.

Character Modes

"Vintage" adds analog warmth and tape saturation. "Modern" focus on clarity and high-end sparkle.

Stereo Imaging

The AI adjusts the "Width" of your master, ensuring the mono-compatibility while creating a vast, immersive soundstage.

What is going on? The engine is running thousands of simulations per second, adjusting faders and thresholds to find the "Perfect Balance."
05

Phase 5: The Final Render

The final polish that makes your song ready for the world.

Shawn Render Engine

Our proprietary engine performs a multi-pass render, ensuring phase-coherence and peak-precision.

Reference Matching

If provided, the AI performs a "Tone Transfer," matching your record's balance to your favorite artist's sound.

What is going on? The final stage uses "Look-Ahead" limiting to prevent distortion while maximizing commercial impact and emotional depth.