Quick Start Guide
Vanta is designed to take you from raw mix to final master in minutes – without sacrificing quality or transparency. Here's the complete path from drop to export.
Drop your mix
Drag and drop a WAV, FLAC, or AIFF file into the main drop zone. Vanta immediately begins a silent spectral health scan – you'll see the waveform load and analysis populate shortly after. Supported formats include 16-bit, 24-bit, and 32-bit float at any sample rate from 44.1 kHz to 192 kHz.
Select your genre or preset
Choose the genre that best represents your track's intent. This is not just a sound character preset – it sets the comparison profile the engine uses to evaluate your mix. A hip-hop track with heavy sub energy that might flag as "over-bearing" against a Classical profile is perfectly normal against the Hip-Hop profile's P50 values.
Review the source metrics
After analysis Vanta displays your source measurements: integrated loudness (LUFS), true peak, loudness range (LRA), crest factor, and stereo correlation. These are the exact values the engine feeds into its decision logic. The full tonal analysis – archetypes, spectral breakdown, confidence evidence – unlocks after you run the master.
Set your loudness target
Select your delivery profile from the dropdown. Vanta is optimised around Streaming Standard (−14 LUFS), Streaming Dynamic (−16 LUFS), and Broadcast (−23 LUFS). You can also enter a custom target when a downstream requirement explicitly calls for it. The engine uses an ISP-aware true-peak limiter with adaptive oversampling, typically 4× with 8× escalation when tighter ceilings or ISP risk justify it.
Master & Export
Hit Master. The chain runs – the exact number of active stages depends on your settings, as options like the De-esser and Bass Lift are off by default and only run when you've enabled them. When complete, A/B the master against your original using the waveform player, then export to WAV, FLAC, or MP3/M4A. A PDF Evidence Certificate can also be exported for audit and delivery traceability.
Already well-mixed? The engine will back off.
If Vanta's analysis finds that your mix is already hitting healthy genre targets, the Confidence System will hold back and apply only light safety processing. You're not forced into a house sound – the engine's job is to improve what needs improving and protect everything else.
Interface Tour
Vanta's interface is organised into a sidebar navigation and a main workspace. Here's what lives where.
Sidebar
The left sidebar contains the five main sections: Quick Master (fastest path), Master (full console), Album Mode (multi-track cohesion), Projects (save and reload sessions), and Profiles (custom processing profiles). The theme selector at the bottom lets you switch between five colour themes: Gold, Blue, Green, Orange, and Cyan.
Master View – Before a file is loaded
Master View – After analysis
Once a file is loaded and analysed, the workspace shows: the waveform player, spectral analysis charts, your genre and platform settings, the master/export controls, and the visual analytics panel below. The analytics panel expands to show the tonal balance, stereo field, archetype, and confidence evidence once a master has been run.
Visual Analytics Panel
After a master, the analytics panel unlocks a set of views – toggled by the tabs at the top of the panel:
Spectral Before / After
An overlay of your mix's spectral energy before and after mastering, per band. Shows exactly where the engine made changes.
Stereo Field
Lissajous-style stereo correlation view, plus per-band correlation bars showing how the engine managed width at the sub, bass, and mid layers.
Archetype Radar
A radar chart of the nine tonal archetype scores – how "muddy", "thin", "harsh", "club-ready" etc. the engine judged your mix to be.
Confidence Evidence
The breakdown of every sub-score that contributed to the engine's confidence level – and how that confidence scaled each processing stage.
Genre Profiles & Custom Profiles
Selecting a genre does far more than apply an EQ curve. It tells the engine what is normal for this kind of music – and what counts as a problem worth correcting.
What a Genre Profile controls
Each genre profile defines a processing posture: how assertively the engine should treat dynamics, how cautiously it should handle width, what low-end behaviour is normal for that genre, and which compound rules should be emphasised or suppressed. For example:
| Genre | Family | Exciter | Dynamics Posture | Notable Behaviour |
|---|---|---|---|---|
| Classical | Organic | Low (0.4×) | Highly conservative | Protects extreme dynamic range, keeps harmonic enhancement low, and avoids density-first behaviour. |
| Hip-Hop | Urban | Full (1.0×) | Controlled but assertive | Sub-dominant bass mode, genre-relative low-end tolerance, and selective rule suppression where dense modern masters are normal. |
| EDM | Electronic | Full (1.0×) | Controlled but assertive | Club-ready archetype weighting, stronger transient-to-loudness push, and rule suppression where dance material expects density. |
| Lo-Fi | Urban | Gentle (0.5×) | Gentle and texture-preserving | Suppresses some bright/harsh interpretations and preserves warmth rather than forcing a polished hi-fi finish. |
| Acoustic | Organic | Moderate (0.7×) | Highly conservative | Conservative dynamics correction, restrained enhancement, and stronger protection of room and natural transient space. |
| Rock | Hybrid | High (0.9×) | Balanced | Percussion-aware low-end handling with a bias toward retaining drive and openness rather than flattening the mix. |
Genre profiles use real statistical data
Vanta's genre logic is grounded in measured corpus statistics, expressed as P10/P50/P90 anchors, and combined with bounded engine policy. The engine compares your track to a real distribution of releases rather than to an imaginary "perfect" reference.
Custom Profiles
The Profiles section lets you save your own processing configurations: custom loudness targets, genre overrides, bass lift settings, and output format preferences. Profiles are stored locally and can be named, edited, and duplicated. When a custom profile is selected, the label in the main view reflects the profile name rather than only the raw genre.
How the Engine Works
Vanta's mastering engine is a 32-stage deterministic pipeline. Every stage either analyses, corrects, or protects – and every decision is traceable back to a measured property of your mix. There is no AI black-box, no style interpolation, and no hidden "taste" layer.
The Three Phases
The 32 stages are grouped into three functional phases that run in sequence:
The Chain Budget
Every INT-driven gain change must be approved by the Chain Budget Controller before it is applied. The budget keeps a running total of cumulative processing. When the total approaches the ceiling, proposed corrections are proportionally scaled back. This hard ceiling is what prevents Vanta from accidentally stacking a +2 dB EQ boost, +1.5 dB of harmonic saturation, +1 dB of compression makeup gain, and a limiter reduction into an over-processed master.
The ceiling adapts to your material:
Dynamic material → 3.5 dB ceiling
Well-preserved dynamics need less correction. The engine applies a tighter budget to avoid over-processing something that's already healthy.
Standard material → 5.0 dB ceiling
Normal commercial mix with moderate corrections needed. The standard operating range for most tracks.
Compressed material → 6.5 dB ceiling
Dense, heavily limited tracks need more corrective work to approach a natural result. The engine gets more budget to work with.
Compound Rules
Eleven compound rules fire when a combination of measured metrics indicates a specific problem. Unlike the single-metric archetype scores, these rules detect interactions – e.g. a track that is simultaneously harsh and over-compressed needs different treatment than one that is only harsh. Rules modify the processing plan directly: a fired rule might reduce a compressor ratio, add a targeted EQ dip, or cap stereo width before those stages even run.
Low End Density
Fires when the sub-bass is excessive and crest factor is already dense. Raises the mono floor and reduces sub compression ratio.
Phase Risk
Fires when stereo correlation is very low with strong sub energy. Caps width and raises the mono-below frequency to protect translation to mono systems.
Harshness Control
Fires on bright, high-presence material. Clamps exciter drive and adds a subtle 9 kHz rolloff to reduce listening fatigue.
Already Crushed
Fires when the track is already severely limited at the source. Pulls back compressor ratios and relaxes limiter release – the engine refuses to make it worse.
Dynamic Ballad Protection
Fires on material with wide intentional dynamic swings. Reduces compressor ratios on mid/high bands and protects the bass from dynamic processing.
Transient Preservation
Fires on rhythmically dense material. Scales per-band transient extension with the highest weight in the snare/attack range (200 Hz–1 kHz).
Spectral Imbalance
Fires when top and bottom spectral halves are significantly mismatched in energy. Applies targeted band corrections to restore tonal balance before archetype corrections run.
Midrange Coherence
Fires when the low-mid region is both elevated and spectrally diffuse. Reduces mid-band gain and tightens the multiband compressor ratio in that range.
Spectral Tilt Correction
Fires when energy rolls off or climbs too steeply across the spectrum. Applies a gentle broadband tilt correction before any per-band EQ to avoid overstacking narrow adjustments.
MS Mono Safety
Fires when side content is dangerously high relative to mid. Forces a mid/side rebalance before width processing runs, preventing mono fold collapse on consumer playback systems.
Multiple Moderate Problems
Fires when three or more metrics are mildly out of range but none is severe enough to trigger individual rules. Reduces the overall processing budget ceiling to avoid stacking corrections on an ambiguous mix.
Suppressed rules are not bugs
Some rules are intentionally suppressed for certain genres. Hip-Hop suppresses the ballad protection rule because heavily compressed, maximised masters are normal and expected. Lo-Fi suppresses harshness control because its character depends on brightness. The suppression is intentional and logged in the Evidence Certificate.
Track Archetypes
Before any processing begins, Vanta classifies your mix into one or more of nine tonal archetypes. Each archetype represents a specific signal character – and the engine tailors its corrections to match. All thresholds are genre-relative: the "muddy" threshold for Hip-Hop is different from the "muddy" threshold for Classical.
Problem Archetypes
Very low crest factor and LRA – the mix has been over-limited at source. Engine backs off compression and limiter aggression.
Very little sub-bass energy, elevated air band, and high stereo correlation. Triggers sub enhancement and air management.
Low-mid energy is elevated and presence is suppressed. Triggers a targeted low-mid cut with presence recovery.
Stereo correlation is dangerously low. The mix risks losing content in mono. Triggers the mono safety and width cap rules.
Presence band is aggressive, especially when combined with high crest smash. Exciter is clamped; high shelf rolloff applied.
Both air and presence are very low. The mix lacks HF detail and energy. Air, presence, and exciter stages are pushed harder.
Positive Archetypes
Strong sub, crest factor in the 8–12 dB range, good bass correlation. Signals that this track is intentionally loud and sub-heavy – engine protects it.
High crest factor, wide LRA, good stereo correlation. Ballad protection rule fires; compression is kept very light.
No problem archetype exceeds 0.20 severity. The engine defaults to minimal, preservative processing.
Archetypes can co-exist
A track can score on multiple archetypes simultaneously. A mix that is both "harsh and fatiguing" and "dense and loud" gets treatment for both – and the compound rule system detects this multi-problem scenario and applies an additional confidence reduction to prevent the engine from over-correcting in multiple directions at once.
Genre Landscape
For every mastered track, Vanta compares your source mix against a corpus of commercial releases in the same genre. The result is a set of Z-scores and percentile ranks that show exactly where your track sat — before mastering — relative to the genre norm. This is not generic loudness advice; it is genre-relative, metric-by-metric context.
Reading the Z-score lanes
Each metric is shown as a dot on a horizontal lane spanning the genre corpus. Three anchors are marked: P10 (bottom 10%), P50 (median), and P90 (top 10%). The grey dot shows the source value; the gold dot shows the mastered output. The percentile rank (P-value) is displayed as a badge on the right.
What metrics are measured
Integrated LUFS
LUFS-I value versus the genre's typical loudness range. Useful for checking whether a mix is already over-limited or too quiet for its genre.
Crest Factor
Peak-to-average ratio — a measure of dynamic punch. Genre-normalised so a high crest factor in Electronic is judged against Electronic norms, not Classical.
Sub Bass
Energy in the sub-bass region relative to the genre. Useful for checking whether the track is carrying the expected foundation or overloading the very bottom end.
Bass
Low-frequency energy above the sub band, where weight and note definition usually sit. Shows whether the mix is carrying enough body for the genre.
Presence
Upper-mid energy where forwardness and bite live. Useful for spotting mixes that may feel recessed or, conversely, too aggressive in the vocal range.
Air
High-frequency openness relative to the genre. Low air can indicate dull mixes; excessive air can indicate brittle high-end that exciter stages would amplify.
Z-scores explained
The Z-score tells you how many standard deviations your track's value sits from the genre mean. A Z-score of 0 means you're exactly average for your genre; +2 means you're unusually high; −2 means unusually low. In the Evidence Certificate, the bar spans −3σ to +3σ — gold bars are above the genre mean, grey bars below.
This is context, not a score
The genre landscape does not tell you whether your mix is good or bad — it tells you where it sits. A track with a crest factor at P15 for Pop might be intentional. A track at P98 for LRA in a Hard Techno session almost certainly isn't. Use it to confirm intent and spot unintended anomalies.
The Confidence System
The Confidence System is what separates Vanta's processing from a static preset chain. It asks: "How certain am I about this mix's signal character?" – and scales every correction accordingly. A track that can't be reliably fingerprinted gets safety-only processing, not a full treatment.
The Four Confidence Tiers
What raises or lowers confidence?
Confidence is a weighted mean of ten sub-scores. The four highest-weighted components are:
Archetype score (28%)
How strongly your mix matches a known tonal character. A clear fingerprint means the engine knows what it's dealing with.
Genre certainty (10%)
User-selected genre scores 1.0 (you know best). Detected genre uses a weighted formula of match confidence, margin, and entropy.
Analysis quality (14%)
Based on file duration. Tracks under 30 seconds score 0.2; 60 seconds scores 0.7; 120+ seconds scores 1.0. Very short clips get light treatment.
Archetype spread (18%)
How clearly separated the top archetype is from the second-place archetype. Ambiguous material (two equally strong archetypes) reduces confidence.
The remaining six sub-scores account for the final 34%: profile fit (8%), physics compressor reliability (6%), spectral clarity (6%), stereo reliability (5%), transient reliability (3%), and dynamic range (2%). All ten values are visible in the Confidence Evidence tab after mastering.
The Confidence Evidence tab shows all sub-scores
After a master, open the Visual Analytics panel and select the "Confidence Evidence" tab. Every sub-score is listed with its value and weight, so you can see exactly why the engine was confident – or cautious – about your specific mix.
Quick Master
Quick Master is the fastest path through Vanta. It runs the full 32-stage engine with smart defaults – you select a genre and a platform target, and the engine does the rest. No additional configuration required.
When to use Quick Master
Quick Master is ideal for:
- Rapid client turnarounds where you need a solid master fast
- Demos and reference masters to evaluate a mix
- Batch processing where consistent settings are preferred
- Teaching situations where you want students to hear the engine without the complexity of the full console
Quick Master uses the same engine as the full console
There is no "lite" mode. The same 32 stages, the same compound rules, and the same Confidence System run in Quick Master. The only difference is that fewer settings are exposed for manual adjustment.
Album Mode
Album Mode allows you to master an entire album or EP as a cohesive body of work. Each track is individually analysed and mastered, but with cross-track awareness – the engine ensures consistent loudness, tonal balance, and spectral character across all tracks.
Album cohesion system
After all tracks are analysed, Vanta computes a statistical profile of the album – the median spectral energy, loudness, and dynamic character. This profile is used as a reference when planning each individual master:
Album Median Reference
By default, each track's spectral plan targets the album median. Tracks that deviate from the median are gently pulled toward it – not forced to match exactly, but to cohere. The delta correction is 35% of the gap, preserving natural tonal variation between songs.
Hero Track Reference
Optionally, designate one track as the "hero" – the reference that represents the intended sound of the album. All other tracks are planned to match this hero's tonal character. Useful when one specific track is your sonic benchmark.
Per-Track LUFS Planning
Each track gets an individual loudness target derived from the album's shift toward the platform target, plus a subtle cohesion pull of ±1.2 LU maximum. This prevents one very loud track from making quieter tracks feel jarring in context.
Outlier flags
Album Mode flags tracks that are statistical outliers in the context of the album – significantly louder, much more sub-heavy, or with mono compatibility risk. These flags are shown per track in the album view so you can decide whether to apply the album plan, override it, or leave a track intentionally distinct.
Evidence Reports & Certificates
Vanta can export a Mastering Evidence Certificate, a structured PDF that documents every decision the engine made. This report is the foundation of Vanta's transparency guarantee.
What the certificate contains
Spectral Correction Log
Band-by-band: original energy, mastered energy, EQ intent, and correction state (held / soft-corrected / hard-corrected / protected).
Chain Budget Report
Cumulative processing budget used, per-stage contributions, and the adaptive ceiling policy that was applied to this track.
Archetype Evidence
Every archetype score, the compound rules that fired, and the severity that triggered each rule's actions.
Risk Translation
Four automated quality checks: mono fold risk, live codec survival (AAC/MP3 transcode), true-peak safety, and loudness target accuracy – each rated clean / watch / risk.
Genre Landscape
Z-score and percentile rank for each measured metric against the genre's corpus – shows where your track sat before mastering relative to commercial releases.
Confidence Evidence
All ten confidence sub-scores and per-stage confidence values, with the scaling factor that was applied to each processing stage.
Useful for clients and collaborators
The certificate provides an auditable record of the final outcome – loudness result, true-peak compliance, codec checks, warnings, and engine decisions. Engineers can include it in deliverable packages to clients or A&R contacts as evidence of what the engine did and why.
Loudness Targets & Platform Compliance
Vanta's final limiter is ITU-R BS.1770-4 compliant – the international standard for integrated loudness and inter-sample true-peak measurement. The delivery profiles below are the ones Vanta is optimised around. Custom remains available for specific downstream requirements, but it is not presented as the default workflow.
About streaming loudness normalisation
While early streaming services varied wildly in their playback levels, the industry has largely converged around the −14 LUFS standard. Rather than presenting a redundant list of platform-specific presets (Spotify, Apple, YouTube), Vanta groups targets by musical intent: "Streaming Standard" for modern density, and "Streaming Dynamic" for genres demanding wider headroom. In all modes, Vanta's strict True-Peak safety layer ensures your master translates flawlessly across every service without triggering clipping penalties.
Verified "Encode Survival" Checks
Most automated systems rely on a static -1.0dBTP ceiling and hope it survives the transcode to lossy formats. Vanta is the only local engine that performs a live "Encode Survival" check: we run a hidden AAC/MP3 transcode of your master and measure the resulting peaks before you save. This ensures your master translates to streaming without Clipping distortion, no matter how aggressive the encoder's psychoacoustic model.
Keyboard Shortcuts
Vanta is built for keyboard-driven workflows. These shortcuts work in the main Master view.
| Action | Shortcut |
|---|---|
| Play / Pause playback | Space |
| Toggle A/B (Original ↔ Master) | O |
| Toggle Mono Audition | M |
| Toggle Reference Track | R |
| Save Custom Profile (Profiles view) | ⌘S |
| Duplicate Profile (Profiles view) | ⌘D |
Education & Instructors
Vanta can be used in higher education audio engineering programmes as a teaching tool for mastering science. Because the engine is fully deterministic and local, students can iterate without cloud costs, internet dependencies, or opaque results.
What makes Vanta suitable for teaching
Auditable decisions
Every correction is explained in the Evidence Certificate. Instructors can walk students through exactly why a compressor ratio was reduced or why a high-shelf was applied – backed by measured data, not opinion.
Deterministic output
The same file processed with the same settings always produces the identical master. This is essential for controlled A/B comparisons and reliable classroom demonstrations.
Fully offline
No internet connection required for the mastering engine. Secure for campus lab environments, air-gapped networks, and exam settings.
Physics-based explanations
The archetype, confidence, and compound rule systems all map directly to textbook DSP and psychoacoustics concepts – crest factor, LRA, Bark-scale masking, stereo correlation, and spectral tilt.
Curriculum integration
Vanta maps naturally to standard audio engineering curriculum topics:
| Topic | Vanta Feature |
|---|---|
| Loudness & dynamics (EBU R128, ITU BS.1770) | LUFS Analysis stage, platform targets, Evidence Certificate loudness section |
| Psychoacoustics – critical band masking | Bark-Scale Masking Suppressor (Stage 17) |
| Stereo imaging & mono compatibility | Stereo Width, M/S EQ, mono-below-Hz, Phase Risk rule, stereo field visualiser |
| Dynamic range & compression | Multiband Compressor, Transient Shaper, crest factor analysis, LRA measurement |
| Harmonic distortion & saturation | Parallel Harmonic VCA (Stage 11), Harmonic Exciter (Stage 19) |
| Spectral analysis & tonal balance | Spectral Fingerprint, archetype radar, genre landscape Z-scores |
| True peak & inter-sample limiting | Adaptive-oversampling True-Peak Limiter, Clipper A + B, Transient Guard |
Technical Architecture Documentation
For instructors, programme leads, and institutional evaluators requiring deep technical documentation: the full Engine Whitepaper (v2.0) covers every stage's algorithm, all threshold values, the Confidence System's weighted components, and the Compound Rules activation logic in full. Available on request – contact support@vantaaudio.cc.