Master with clarity — analyse locally and export with confidence.
Complete user guide for SPIRAL by GLENDER — local mastering, spectral analysis and secure export in a premium browser workflow.
What this app does
SPIRAL by GLENDER applies a professional mastering chain to your exported mix, also known as your bounced mix. It is not a DAW plugin. It is a complete processing engine that runs directly in the browser using the Web Audio API.
When you load a track, the app automatically performs three tasks: it analyzes the frequency spectrum, calculates optimal starting settings for each module by comparing your mix with a training of the best tracks for every genre, and prepares the real-time processing chain so you can play, compare and adjust immediately.
Recommended workflow
Load your mix
Drag the file into the upload area or click to select it. Accepted formats: MP3, WAV, FLAC, AAC, OGG and M4A. Uncompressed WAV gives the best results.
Read the spectral analysis
The analysis shows your track in gold bars against the target spectrum. Bands that are far from the target are highlighted — focus your adjustments on those. The metrics panel shows LUFS-I, True Peak, Dynamic Range, and Stereo Correlation to guide you.
Press ▶ Mastered and listen
The chain already contains auto-calculated settings. Listen before changing anything. Compare with ▶ Original to understand the starting point.
Adjust module by module, top to bottom
Follow the chain order: HPF → Corrective EQ → Saturation → Bus Comp → Tonal EQ → M/S → Limiter. Every module prepares the signal for the next one. Changes are applied instantly while you listen.
Use A/B constantly
Switch between Original and Mastered every 30 seconds. The ear adapts quickly and loses perspective. A/B comparison forces a more objective decision. If the mastered version is not clearly better, something needs adjusting.
Save the preset and export
Save the preset before exporting in case you need to return to the same settings. Then choose the bit depth — 24-bit for professional delivery, 16-bit for direct distribution — and press WAV. Output Gain controls the final level before rendering.
Spectral analysis & Metrics
When you load a track, SPIRAL performs a comprehensive spectral analysis. The results appear in two places: the Mastering Intelligence panel (right sidebar) and the Real-Time Analyzer (center canvas).
Mastering Intelligence Panel
Shows your track's key metrics: LUFS-I (loudness), True Peak (inter-sample peaks), Dynamic Range, and Stereo Correlation. Each metric is color-coded — green for good, gold for caution, red for attention. Off-target frequency bands are highlighted with specific dB values so you know exactly where to focus.
Real-Time Analyzer
The center canvas displays the live frequency spectrum during playback. EQ response curves are overlaid directly on the spectrum — corrective EQ in gold and tonal EQ in blue. You can drag EQ nodes directly on the canvas to adjust frequency and gain visually. When playback is paused, the original analysis spectrum is shown for reference.
The analysis uses 15 averaged FFT snapshots across the first 40 seconds of your track for an accurate spectral profile. What you see in the analyzer always reflects the actual processing in real time.
The seven frequency bands
| Band | Range | What it contains |
|---|---|---|
| Sub | 20–80 Hz | Subsonic frequencies, deep kick and very low bass. Powerful but dangerous if excessive. |
| Low | 80–250 Hz | Bass body, kick body and the warmth of the mix. |
| Lo-Mid | 250–500 Hz | The mud zone. Guitar and vocal body. Easily becomes problematic. |
| Mid | 500–2000 Hz | Presence, vocal intelligibility and instrument body. |
| Hi-Mid | 2–8 kHz | Attack, definition and articulation. The area that cuts through the mix. |
| High | 8–16 kHz | Brightness, sibilance and cymbal presence. |
| Air | 16–22 kHz | Air, openness and room sensation. The ingredient of a professional sound. |
Analysis metrics
Five metrics are calculated when the file loads. After mastering, the processed value appears beside the original value.
| Metric | What it measures | Optimal area |
|---|---|---|
| LUFS-I | Integrated loudness according to ITU-R BS.1770-4 using K-weighting and absolute/relative gating. This is the practical loudness number used by Spotify, Apple Music and YouTube. | −16 to −9 LUFS |
| True Peak | Inter-sample peak calculated with 4× cubic interpolation. It detects clipping that can appear after MP3/AAC encoding even when the WAV looks safe. | ≤ −1.0 dBTP |
| Dyn. Range | The difference between the loudest and quietest levels, estimated as DR. Low values indicate heavily compressed masters. | 7–14 dB |
| Correlation | Stereo phase. +1 means perfect mono, 0 means no correlation, and negative values suggest phase problems. | 0.5–0.95 |
| Duration | Original file duration and sample rate. | — |
The modules are processed in a fixed left-to-right order. Every stage prepares the signal for the next one. You cannot change the order, and you should not want to. The order is the recipe.
The HPF cuts everything below its cutoff frequency. Subsonic content is inaudible but destructive: it consumes limiter headroom and makes the compressor pump in reaction to energy you cannot hear. Removing it frees real dynamic space.
The cutoff frequency is automatically adjusted according to the spectral analysis. If the Sub band has relative excess, the app raises the HPF up to 55 Hz. If the Sub band is balanced, it stays at the default 22 Hz.
| Slider | Range | Direction | How to use it |
|---|---|---|---|
| Cutoff frequency | 15–80 Hz | Raise to cut more low end Lower to preserve more low end | The auto-analysis sets it around 22–55 Hz depending on sub content. Lower it if the bass loses body; raise it up to 40–60 Hz if the low end still feels muddy. Do not go beyond 80 Hz or you will remove bass weight. |
Its mission is to remove what is excessive before the compressor amplifies it. A compressor does not know the difference between an annoying resonance and useful musical content. If there is a spike at 300 Hz, the compressor will press everything down together. Cutting first makes the compressor respond to musical dynamics instead of mud.
The module has four independent bands centered around the mud area, each with draggable frequency, gain and Q controls. The auto-analysis already sets the gain according to the relative excess detected in each region.
| Slider | Typical range | Direction | How to use it |
|---|---|---|---|
| Gain | −12 to 0 dB | Always downward | This EQ is for cutting only. You may temporarily boost to locate a resonance, then reduce the gain to cut it. Typical cuts are −2 to −6 dB. |
| Frequency | 50–6000 Hz | Sweep to find | Move it while listening. Default points are 200 Hz for low/kick body, 300 Hz for classic mud, 500 Hz for boxiness and 1 kHz for hardness or metallic tone. You can also drag the nodes directly on the real-time analyzer. |
| Q — precision | 0.5–10 | Raise for narrower cuts Lower for wider cuts | For specific resonances, use Q 3–8. For smoothing a wide area, use Q 0.8–1.5. A very high Q affects almost only that exact frequency. |
Saturation introduces gentle harmonic distortion using a mathematical tanh curve that imitates analog tape behavior. Pure digital audio can feel too perfect, cold or thin. Saturation adds spectral density without sounding like obvious distortion; it simply feels fuller and more present.
It is placed before the compressor on purpose. The compressor then settles the new harmonics, just like in an analog hardware chain: tape into compressor.
| Slider | Range | Direction | Value reference |
|---|---|---|---|
| Amount — tape warmth | 0.0–0.8 | Raise for more warmth and density Lower for cleaner digital transparency | 0.05–0.15 = almost invisible warmth, noticed when removed 0.15–0.30 = soft musical analog color, the usual area 0.30–0.50 = clearly audible saturation 0.50+ = intentional heavy saturation effect |
When 20 separately recorded instruments pass through gentle stereo-bus compression, they all move down slightly at the same time. This creates the feeling that they are performing together in the same room. That is glue. It is not about making the track louder; it is about making it feel cohesive.
| Slider | Typical range | Direction | How to use it |
|---|---|---|---|
| Threshold | −28 to 0 dB | Lower for more compression Raise for less compression | The point where compression starts. Begin around −18 dB and lower it until the mix settles and the elements integrate better. Aim for 2–4 dB of gain reduction. |
| Ratio | 1.5:1–8:1 | Lower for gentle glue Raise for obvious compression | 2:1 is classic musical glue. 3:1–4:1 becomes audible. Above 4:1 it clearly sounds compressed. For pop-rock mastering, you should rarely go beyond 3:1. |
| Attack | 5–300 ms | Slower = more punch Faster = more control, less punch | 30–50 ms is the classic glue zone. It lets the first hit of the kick and snare through, creating punch, while still controlling overall dynamics. Avoid going below 10 ms unless you have a clear reason. |
| Release | 50–600 ms | Slower = more natural Faster = more energy, pumping risk | 150–250 ms is the usual range. If the mix breathes or pumps obviously, increase the release time. |
| Makeup Gain | 0–12 dB | Raise to recover lost level | Compression reduces level. Makeup Gain restores it. Adjust it until the perceived level with the compressor active is similar to the bypassed level, so the comparison stays fair. |
This EQ is placed after the compressor because you are shaping a signal that is already compressed and cohesive. Here you are not fixing problems — that was the Corrective EQ. Here you are giving character, like broad strokes of color on a clean canvas.
Its seven bands match the seven zones of the initial spectral analysis. That gives direct correspondence: if the analysis shows too much Lo-Mid, the Lo-Mid node will already be adjusted automatically. The three center bands — Lo-Mid, Mid and Hi-Mid — are draggable peak filters. The outer four bands are fixed shelves.
| Band | Frequency | Direction | How to use it |
|---|---|---|---|
| Sub | 40 Hz shelf | Raise for physical weight Lower if sub clouds the mix | Only obvious on systems with a subwoofer or good monitors. Move carefully, usually no more than ±1–2 dB. |
| Low | 120 Hz shelf | Raise for body and warmth Lower if there is too much weight | Adds body and warmth. Move in 0.5 dB steps. More than +3 dB often sounds cloudy. If you cut this area in the Corrective EQ, do not undo that work here. |
| Lo-Mid | 350 Hz peak | Lower to clean mud Raise for more guitar body | The musical mud band. −1 to −2 dB cleans the mix without hollowing it out. A Lo-Mid cut around −1.5 dB is often the difference between an amateur master and a professional one. |
| Mid | 1 kHz peak | Raise for presence and clarity Lower if voice or guitars feel hard | The range of vocal and guitar presence. +1 to +2 dB helps the mix cut through small speakers. −1 to −2 dB softens aggressive or fatiguing mixes. |
| Hi-Mid | 3 kHz peak | Raise for definition and attack Lower if harsh or sibilant | Definition and impact. +1 dB adds clarity and cut. Be careful: the ear is very sensitive here, and more than +2 dB can become tiring. |
| High | 8 kHz shelf | Raise for clarity and shine Lower if there is sibilance or harshness | Brightness and cymbal attack. +1 to +2 dB adds clarity. If vocal sibilance is already a problem, do not boost here. |
| Air | 16 kHz shelf | Usually raise Lower only if there are digital whistles | The secret ingredient of a modern professional sound. A boost of +1.5 to +3 dB opens the space above the mix without adding artificial brightness. |
This module uses a professional Mid/Side matrix to give you independent control over the center and the sides of your stereo image — the same technique used in mastering studios worldwide.
How Mid/Side works
The stereo signal is decoded into Mid (L+R, the center — vocals, kick, bass, snare) and Side (L−R, the difference — reverbs, cymbals, ambience). Each is processed independently, then re-encoded back to stereo. Because the Side cancels out when summed to mono, M/S processing is 100% mono-compatible — unlike Haas delays or phase tricks.
Controls (3D Knobs)
| Knob | Range | Default | How to use it |
|---|---|---|---|
| Mid Gain | -6 to +6 dB | 0 dB | Controls the center. Boost to bring vocals and kick forward. Cut to clean up a cluttered center. |
| Side Gain | -12 to +12 dB | 0 dB | Controls stereo width. +2 to +4 dB opens up the mix. -2 to -4 dB narrows if there are phase issues. |
| Bass Mono | 0 (Off) to 300 Hz | 120 Hz | Critical for professional masters. Forces everything below this frequency to mono by filtering the Side channel. 120 Hz is the industry standard — keeps bass punchy on club PAs, vinyl, and Bluetooth speakers. Set to 0 to disable. |
Solo Monitoring
Three buttons below the knobs let you audition: Mid (hear only the center), Side (hear only the sides), or Full (normal stereo). Essential for checking what you're adjusting.
The limiter is a compressor with a very high ratio and ultra-fast attack. No sample in the resulting audio will exceed the configured Ceiling. It is the absolute ceiling of the master. It also controls final loudness: the more level reaches it, the harder it works.
| Slider | Range | Direction | How to use it |
|---|---|---|---|
| Ceiling | −6 to −0.1 dBFS | Closer to 0 = louder, more risk Lower = more codec headroom | −1.0 dBFS = streaming standard −1.5 dBFS = when original True Peak was high −0.3 dBFS = physical CD mastering −2.0 dBFS or lower = if the song will be processed again later |
Streaming platforms and LUFS
Streaming platforms normalize loudness automatically. If your master is louder than their target, they will turn it down. Knowing these values helps you decide how hard to push the limiter.
| Platform | Target LUFS | Recommended ceiling |
|---|---|---|
| Spotify | −14 LUFS integrated | −1.0 dBFS |
| Apple Music | −16 LUFS integrated | −1.0 dBFS |
| YouTube | −14 LUFS integrated | −1.0 dBFS |
| Tidal | −14 LUFS integrated | −1.0 dBFS |
| CD / Physical | No normalization | −0.3 dBFS |
Built-in Meters
Opening the Limiter card shows a VU Meter (horizontal bar with LUFS-M) and a Gain Reduction meter showing how hard the limiter is working in real time. The Compressor card also includes its own Gain Reduction meter.
3D Knobs & Controls
Every module now uses professional 3D rotary knobs instead of flat sliders. They look and feel like hardware mastering gear.
Knob interactions
| Action | Result |
|---|---|
| Drag up/down | Adjust value — 240px = full range |
| Scroll wheel | Fine adjustment — ±1 step |
| Double-click | Reset to default value |
| Ctrl+Click | Reset to default value (any control) |
Module Card
Each module opens as a premium card with a colored strip, LED indicator, ON/OFF toggle, reset button, and close button. The 3D knobs are arranged in a row with values in JetBrains Mono and labels in the serif font.
Mix Analysis
When you load a track, SPIRAL analyzes your raw mix and shows the results in a card at the top of the center panel. This tells you what needs attention in your mix before mastering — LUFS level, dynamic range, clipping, frequency balance, stereo correlation, and more.
What it shows
| Feature | Description |
|---|---|
| Mix Score (0–100) | Overall track health — circular gauge with color-coded grade (Excellent, Good, Fair, Poor) |
| Key Metrics | LUFS-I, Dynamic Range (dB), True Peak, Stereo Correlation, Clipping detection |
| Mix Issues | Expandable list showing every check — both PASSED and flagged — with severity tags and fix suggestions using SPIRAL modules |
Output Gain is the final control of the whole chain, after the limiter. Unlike compressor Makeup Gain, which changes how hard the limiter works, Output Gain adjusts the level after limiting has already happened. This lets you raise or lower the final delivery level without changing compression or limiting behavior.
| Slider | Range | Direction | When to use it |
|---|---|---|---|
| Output Gain | −12 to +6 dB | Raise for more delivery level Lower for headroom or fair A/B | 0 dB = default, no change −1 to −3 dB = if True Peak exceeds −1 dBTP +1 to +2 dB = to match perceived level with the original during A/B −6 to −12 dB = if the master will be processed again later |
A/B comparison & Level Match
The app has three ways to compare audio, and they work together.
Level Match (NEW)
The Match button in the footer player uses LUFS-based loudness matching. When enabled:
Match ON + A: Plays the fully mastered audio through the complete chain.
Match ON + B: Plays the raw original with no processing at all, boosted to exactly match the mastered LUFS level. This is the fairest possible comparison — same loudness, zero processing on B.
Match OFF + A/B: Two independent mastering slots for comparing different settings.
▶ Mastered / ◉ Original
Direct play buttons for processed vs raw audio. ◉ Orig plays the unprocessed file. When Match is ON, it's automatically gain-matched to the master. Playback position is preserved when switching.
Slots A / B
When Match is OFF, A and B hold two independent mastering configurations. When Match is ON, A = mastered, B = original (matched).
- Slot A is automatically loaded with the auto-analysis settings when the file opens.
- Slot B activates when you press B. The app asks whether you want to copy A’s settings or start from the original analysis values. Once initialized, B keeps its own settings independently.
- When you switch between A and B, the audio chain is rebuilt immediately. If playback is running, the master changes without losing position.
- Each slot stores its settings in memory. When you close or reload the page, both slots are lost unless you save a preset.
Loop
The Loop button activates continuous repetition between two handles that appear on the timeline. When enabled, the waveform shows the loop region with a purple tint and two draggable triangular flags.
| Element | Function |
|---|---|
| Left handle | Defines the loop start point. Drag it to adjust where the loop begins. |
| Right handle | Defines the loop end point. When playback reaches it, it jumps back to the start. |
| Darkened area | Audio outside the loop is visually dimmed to show that it is not being repeated. |
The loop works in both Original and Mastered modes. A/B switching keeps the position inside the loop, which allows you to compare the same fragment indefinitely.
Module reset
Inside each module you will find two specific buttons beside the activation toggle.
- Help button — opens a modal explaining that module, just like the question-mark button in the card header.
- Module reset button — restores only that module to the values automatically calculated by the AI when the song was loaded. This is useful when you have experimented too much and want to return only that module to the starting point without touching the rest of the chain.
The bottom action bar also includes a global Reset button. It restores all modules to the initial settings calculated by the automatic analysis. Slot A is also updated if you are currently working in Slot A.
Presets
The save and load buttons store all module settings as an .mpreset file in JSON format. This allows you to:
- Save a configuration that works before experimenting.
- Reuse the same chain on tracks from the same project or artist.
- Share configurations with other users of the app.
- Build a preset library by genre, such as pop, rock, jazz or electronic music.
Export
The Export button renders the full chain with the current settings and downloads the result. Select the output format from the dropdown next to the button:
| Format | When to use it | Dithering |
|---|---|---|
| 24-bit WAV (default) | Professional delivery to studios, digital distributors, and streaming platforms. | Not needed — 24-bit headroom makes dithering irrelevant. |
| 16-bit WAV | Physical CD, platforms requiring 16-bit, or smaller files for distribution. | Triangular TPDF dithering applied automatically. |
| MP3 320 (NEW) | Quick sharing, reference listens, or platforms that prefer compressed formats. | LAME encoder at 320kbps CBR — highest quality MP3. |
M/S Module — Analog Processor
The M/S module uses a professional analog Mid/Side processor with a Linkwitz-Riley 2nd-order crossover for transparent bass mono-izing. This is superior to the simple highpass filter used in most web-based mastering tools.
Seven quick presets are available directly in the module card:
| Preset | Mid Gain | Side Gain | Width | Best for |
|---|---|---|---|---|
| Clean | 0 dB | 0 dB | 100% | Transparent M/S — no coloration |
| Console | +1.5 dB | +1 dB | 105% | Subtle console-style presence |
| Analog | +2 dB | +1.5 dB | 110% | Rich, wide soundstage |
| Tube | +3 dB | +2 dB | 115% | Maximum width and density |
| Tight Low | +1 dB | −1 dB | 100% | Clean up muddy bass |
| Wide | 0 dB | +2.5 dB | 120% | Open, airy stereo field |
| Vocal Up | +2.5 dB | 0 dB | 95% | Vocals forward, tight bass |
Solo monitoring (Mid / Side / Full) lets you audition individual components. Works identically on both analog and fallback modes.
Master Bypass
The BYPASS ALL button in the left panel (below the module list) instantly disables every module. Click again to re-enable them all.
This gives you a before/after comparison of your entire chain — hear what the processing is really doing without toggling each module individually. The button glows gold when active modules are bypassed, grey when all are on.
Compressor Presets
The compressor module includes 6 quick presets accessible directly in the module card. Each preset configures all 5 compressor parameters with values calibrated for specific mastering styles.
| Preset | Ratio | Threshold | Attack | Release | Makeup |
|---|---|---|---|---|---|
| Gentle Glue | 2:1 | -18 dB | 30ms | 200ms | +2 dB |
| Light Touch | 1.5:1 | -14 dB | 50ms | 300ms | +1 dB |
| Modern Punch | 3:1 | -16 dB | 20ms | 150ms | +3 dB |
| Vintage Warm | 4:1 | -12 dB | 10ms | 100ms | +4 dB |
| Transparent | 1.5:1 | -20 dB | 40ms | 250ms | +1.5 dB |
| Aggressive | 6:1 | -10 dB | 5ms | 80ms | +5 dB |
EQ Presets
Both EQ modules include quick presets for common mastering tasks.
EQ− (Corrective) Presets
Five presets for common resonance problems: Flat (reset), Mud Cleanup (cut at 300Hz), Harshness (cut at 1kHz), Boxiness (cut at 500Hz), Low Boom (cut at 200Hz).
EQ+ (Tonal) Presets
Five presets for tonal shaping: Flat (reset), Air Boost (+2dB High, +1.5dB Air), Warmth (+1.5dB Low), Presence (+2dB Mid), Bright (+1.5dB High + Air).
Solo Monitoring
The M/S module has three solo buttons: Mid (hear only the center), Side (hear only the sides), and Full (normal stereo). These let you audition individual stereo components. Solo only works during playback — press Play first, then click a solo button.
Mastering Intelligence
The Mastering Intelligence panel in the right panel provides real-time, analysis-aware feedback on your mastering session.
Track Analysis
When you load a track, the panel immediately shows LUFS-I, Dynamic Range, True Peak, and Stereo Correlation — with colour coding (green = good, gold = okay, red = problem). Bands that are off-target are highlighted.
Smart Recommendations
Based on the original track analysis and your current settings, the panel suggests specific improvements — like adjusting the HPF for excess sub, or lowering the compressor ratio for already-compressed tracks.
The Ask Spiral input lets you ask questions about your track. Spiral knows your exact analysis data, current settings, and recent changes. It gives specific, actionable advice with exact numbers. When Spiral suggests a preset or parameter change, action buttons appear — click to apply instantly.
Spiral remembers your conversations and learns from past interactions to give better advice over time. All data stays in your browser.
Level Match
The Match button uses LUFS-based loudness matching to let you compare the mastered version against the original unprocessed audio at the same perceived loudness. This is the most honest way to evaluate your mastering.
Click Match to toggle ON. The button glows gold.
Match ON + A: Plays the fully mastered audio.
Match ON + B: Plays the raw original boosted to match the mastered LUFS — zero processing, same loudness.
Match OFF: A and B become independent mastering slots for comparing two different settings.
Real-time analyzer
The central analyzer canvas shows three layers at the same time on top of the frequency spectrum.
FFT spectrum
Real-time frequency analysis of the output signal. During mastered playback it animates at 60 fps. When paused, it shows the original analysis spectrum in a dim static color.
EQ Response Curves
The real-time analyzer overlays your EQ curves on the live spectrum. The corrective EQ curve (gold) and tonal EQ curve (blue) show exactly how your filters are shaping the sound. Drag the nodes directly on the analyzer for visual control.
Tonal EQ curve
The response curve of the seven Tonal EQ filters: Sub shelf, Low shelf, Lo-Mid peak, Mid peak, Hi-Mid peak, High shelf and Air shelf.
Draggable EQ nodes
Corrective EQ nodes and the three center Tonal EQ nodes can be dragged directly on the analyzer. Horizontal movement changes frequency. Vertical movement changes gain. This makes the visual feedback match the audible decision in real time.
Timeline
The timeline shows the full waveform of the loaded track and the current playback position. It is also where the loop region is displayed and edited.
| Element | Meaning |
|---|---|
| Waveform | Visual overview of the track’s amplitude over time. Dense areas usually correspond to loud sections such as choruses. |
| Playhead | The vertical marker that shows the current playback position. You can jump to a point by clicking the timeline. |
| Loop region | The selected area that repeats while Loop is active. It helps compare the same musical section while adjusting modules. |
Output Gain and VU Meter
Output Gain
Output Gain is the final trim after all processing and limiting. It changes the rendered level without changing how hard the compressor or limiter works.
VU Meter L/R
The VU meter shows left and right output level in real time. It helps you catch clipping, excessive limiting and stereo imbalance.
| Element | Meaning |
|---|---|
| Level bars — RMS | Average energy level updated in real time. The level moves from safe low values to warning values as it approaches 0 dBFS. |
| Peak-hold tick | A horizontal mark showing the recent maximum peak. It stays visible briefly before updating, helping you detect repeated hot peaks. |
| 0 dBFS line | The absolute digital ceiling. If the bars consistently reach this line, the limiter is working too hard. Reduce Makeup Gain or Output Gain. |
Reference Track
The Reference Track system lets you load a separate finished track and use it as a target while mastering. It is a listening, measurement and matching tool. It does not replace your master and it is not exported with your final file.
What happens when you load a reference
SPIRAL decodes the reference locally in the browser and analyzes its spectrum, LUFS-I, Dynamic Range, True Peak, Stereo Correlation, average level and stereo image. The app then compares those values against your current track so you can see what is different before applying anything.
Reference analysis
The reference is measured as its own audio source. This gives you a target for tonal balance, loudness, dynamics, peak safety and stereo width instead of only guessing by ear.
Safe comparison
The reference can be played from the footer player without changing your mastering chain. Use it to check whether your master is too dark, bright, narrow, loud, quiet, compressed or dynamic.
Reference player
| Control | What it does |
|---|---|
| + Ref / Load Reference | Loads the reference audio file and runs the full reference analysis. |
| Play Ref | Plays the reference track directly through its own monitoring path. The output meters, correlation meter and real-time spectrum respond to the reference while it is playing. |
| Reference waveform | The footer waveform switches to the reference waveform when Ref is selected. It uses a different color so you can see that you are looking at the reference, not the master or original. |
| Reference seek | Click the reference waveform to jump to any point in the reference track. The reference keeps its own playback position instead of snapping back to the master position. |
| Reference loop | The reference has its own independent loop region. You can loop a chorus or drop in the reference while keeping a different loop on the master/original track. |
| Level Match with Ref | Use loudness matching when you want a fair comparison against the master. Turn it off when you want to hear the reference at its own file level. |
Reference Match Strength
The Reference Match Strength slider controls how strongly SPIRAL previews the reference match. At 100%, the app applies the maximum safe transfer, which is intentionally limited to a partial match instead of copying the reference aggressively. At 50%, it applies half of that safe movement.
Moving the slider updates the matching preview live, including the module settings and the visual curves where applicable. This lets you decide whether the reference direction is helping before committing it.
Apply Reference
Apply Reference is not only an EQ match. It can transfer the reference direction across the mastering chain:
| Module | What can be adjusted from the reference |
|---|---|
| HPF | Low-end cleanup can move toward the reference low-frequency behavior. |
| Corrective EQ | Problem bands can be pushed closer to the reference tonal balance. |
| Saturation | Density and harmonic weight can move toward the reference character. |
| Bus Compressor | Compression direction can follow the reference dynamic range. |
| Tonal EQ | Sub, low, mid, high and air balance can be shaped toward the reference spectrum. |
| M/S | Stereo width and side level can move toward the reference stereo image and correlation. |
| Limiter / Output Gain | Ceiling and final gain can move toward the reference loudness and peak behavior. |
Integrated LUFS-I
LUFS-I, Loudness Units Full Scale Integrated, measures the perceived loudness of a recording according to the ITU-R BS.1770-4 standard. It is the number streaming platforms use to normalize playback volume.
How it is calculated
K-weighting
Two biquad filters are applied in series: a high-shelf pre-filter around 1682 Hz that emphasizes the frequencies where hearing is more sensitive, and a high-pass filter around 38 Hz that removes subsonic content. The result is an ear-weighted signal.
Block power
The K-weighted signal is split into 400 ms blocks with 75% overlap, using a 100 ms hop. For a three-minute track, this creates about 1800 blocks. The mean-square power of each block is calculated.
Absolute gating
Blocks below −70 LUFS are discarded. This removes start and end silence that would otherwise pull the measurement down artificially.
Relative gating
The average of the remaining blocks is calculated, and blocks 10 LU below that average are discarded. Quiet intros and outros are reduced in influence. The final average is the LUFS-I value.
| Platform | Target LUFS-I | What happens above target |
|---|---|---|
| Spotify | −14 LUFS | Playback level is reduced automatically. |
| Apple Music | −16 LUFS | Playback level is reduced automatically. |
| YouTube | −14 LUFS | Playback level is reduced automatically. |
| Tidal | −14 LUFS | Playback level is reduced automatically. |
| Amazon Music | −14 LUFS | Playback level is reduced automatically. |
True Peak
True Peak measures the maximum level the reconstructed signal can reach between digital samples — peaks that are not visible as sample values in the file but appear in the analog reconstruction.
Why it matters
A WAV file can show a maximum sample of −0.1 dBFS and look safe. But when it is encoded to MP3 or AAC, inter-sample peaks can exceed 0 dBFS and clip. Streaming platforms generally expect safe True Peak values, commonly around −1 dBTP for delivery.
How it is calculated
The app uses 4× Catmull-Rom cubic interpolation. For every pair of consecutive samples, it evaluates the cubic polynomial at t = 0.25, 0.5 and 0.75, creating three extra points between samples. The maximum across all these points is the True Peak.