Phantom Bass Enhancer + Haas Stereo Widener — User Guide
Psychoacoustic bass enhancement through harmonic generation combined with spatial widening using Haas effect and mid-side processing.
What this does
This script implements psychoacoustic bass enhancement combined with stereo spatial widening. It creates the perception of deep bass on systems with limited low-frequency response (small speakers, headphones) by generating harmonic overtones that trick the ear into "hearing" missing fundamentals. Simultaneously, it applies stereo enhancement using both the Haas (precedence) effect and mid-side processing to create a wider, more immersive soundstage.
Key Features:
- Phantom Bass: Waveshaping generates real harmonics from extracted bass content
- Smart Filtering: Original bass removed via high-pass, harmonics band-limited for clarity
- Haas Effect: Delays one channel slightly to create spaciousness
- Mid-Side Processing: Adjustable stereo width via side signal manipulation
- Mono Compatibility: Optional safety mode minimizes phase cancellation
- Peak Protection: Multiple normalization passes prevent clipping
- 6 Presets: From subtle enhancement to aggressive MaxxBass-style effects
Psychoacoustic Bass Enhancement: The human auditory system has a phenomenon called "the missing fundamental" where if harmonics of a fundamental frequency are present, we perceive the fundamental even if it's physically absent. This script:
- Extracts bass frequencies (30-120 Hz typical)
- Generates harmonics via waveshaping distortion (tanh function)
- Filters original audio to remove deep bass
- Mixes harmonics back in, creating perception of strong bass
Result: Audio appears to have more powerful bass without actually boosting low frequencies that small speakers can't reproduce.
Technical Implementation: (1) Stereo preparation: Converts mono to stereo if needed. (2) Bass extraction: Bandpass filters each channel for low frequencies. (3) Harmonic generation: Applies tanh(x·drive) waveshaping to create harmonics, then bandpass filters to desired brightness range. (4) Original filtering: High-passes original to remove conflicting bass. (5) Mixing: Combines filtered original with harmonics at controlled mix level. (6) Haas effect: Delays one channel slightly (±15ms typical). (7) Mid-side widening: Extracts mid/side components, boosts side signal. (8) Safety features: Multiple peak detection passes, mono compatibility controls. (9) Preset system: 6 carefully tuned starting points.
Quick start
- In Praat, select exactly one Sound object (mono or stereo).
- Run script… →
phantom_bass_haas_widener.praat. - Choose preset (6 options):
- Custom (use manual settings)
- Subtle Enhancement (gentle)
- Moderate Effect (balanced)
- Aggressive MaxxBass Style (strong)
- Mono-Safe Widening (broadcast safe)
- Wide Stereo (not mono-safe)
- Adjust Phantom Bass parameters if using Custom preset:
bass_low_freq,bass_high_freq: Bass extraction rangedrive: Harmonic generation intensity (0.1-10)harmonic_mix: Dry/wet blend (0-1)highpass_freq: Original bass removal cutoffharmonic_lowpass: Harmonic brightness control
- Set Haas Effect options (enable/disable, delay, mix, level difference).
- Set Mid-Side Widening options (enable/disable, stereo width).
- Enable preserve_mono_compatibility for broadcast/safe use.
- Enable play_result to auto-play output.
- Click OK — processing applied, output named "originalname_phantom_haas".
Quick tip: Start with Moderate Effect preset to hear balanced enhancement. For headphones/small speakers: Try Aggressive MaxxBass Style. For broadcast/streaming: Use Mono-Safe Widening. Always test in mono (collapse to mono in your DAW or sum L+R) to check for phase cancellation. The script includes multiple peak protection passes, but extreme settings can still cause clipping — listen for distortion. For speech/content: Use Subtle Enhancement or disable bass enhancement entirely (set harmonic_mix=0). The Haas effect works best at 10-30ms delays — longer delays create distinct echoes.
Critical: Always test in mono! The Haas effect and mid-side widening can cause severe phase cancellation when collapsed to mono. Enable preserve_mono_compatibility for broadcast, streaming, or club systems. Without this protection, your audio may disappear or become thin in mono playback. The script warns but cannot prevent all issues — your ears are the final test. Haas delays >25ms become audible as distinct echoes. High
drive values (>5) create strong distortion — use judiciously. Removing too much original bass (highpass_freq >150 Hz) can make audio sound thin — balance with harmonic mix.
Psychoacoustic Theory
The Missing Fundamental Phenomenon
👂 How We "Hear" Phantom Bass
Scientific basis: When a complex tone contains harmonics at frequencies f, 2f, 3f, 4f... the auditory system often perceives a fundamental frequency f even if it's physically absent.
Example: Small speaker playing 100 Hz tone
Physical limitation: Speaker can't reproduce 100 Hz effectively
Perceptual trick: Generate harmonics at 200 Hz, 300 Hz, 400 Hz...
Brain interprets: "There must be a 100 Hz fundamental"
Result: Perception of deep bass without actual low frequencies
Physical limitation: Speaker can't reproduce 100 Hz effectively
Perceptual trick: Generate harmonics at 200 Hz, 300 Hz, 400 Hz...
Brain interprets: "There must be a 100 Hz fundamental"
Result: Perception of deep bass without actual low frequencies
Practical application: Extract bass frequencies (30-120 Hz) → Generate harmonics (120-800 Hz) → Remove original bass → Brain fills in the "missing" low end.
Waveshaping function: tanh(drive × x)
Where:
x = extracted bass signal (normalized)
drive = distortion amount (higher = more harmonics)
tanh = hyperbolic tangent (soft clipping)
Harmonic series generated:
For fundamental f, tanh distortion creates:
f, 2f, 3f, 4f, 5f... (odd and even harmonics)
Advantages over simple distortion:
- Soft clipping (no harsh edges)
- Natural harmonic decay
- Controllable via drive parameter
Haas (Precedence) Effect
🎧 Creating Spatial Width with Time Differences
Discovery: Helmut Haas (1949) found that when two identical sounds arrive within 1-30ms of each other, we perceive a single sound located at the first arriving source.
Haas implementation:
L(t) = original_left(t)
R(t) = (1 - mix) × original_right(t) + mix × original_right(t - Δt)
Where:
Δt = delay time (10-30ms typical)
mix = wet/dry balance (0-1)
Perceptual effects:
Δt < 10ms: Image widens without localization shift
Δt = 10-30ms: Clear spatial separation
Δt > 40ms: Distinct echo perceived
Mono compatibility issue: When L and R are summed to mono:
Mono sum = L(t) + R(t)
= original_left(t) + (1-mix)×original_right(t) + mix×original_right(t-Δt)
If original_left ≈ original_right (stereo pair):
Comb filtering occurs due to time difference
Frequency-dependent cancellation
Can cause "phasiness" or complete cancellation
Mid-Side Processing
🎚️ Stereo Width Control
Concept: Any stereo signal can be represented as Mid (mono-compatible) and Side (stereo-difference) components:
Encoding:
M = (L + R) / 2 (Mid = center/mono content)
S = (L - R) / 2 (Side = stereo difference)
Decoding:
L = M + S
R = M - S
Width control:
L' = M + S × width
R' = M - S × width
Where width = 0.5-1.5:
width < 1.0: Narrower than original
width = 1.0: Same as original
width > 1.0: Wider than original
Mono compatibility: When collapsed to mono:
Mono = (L' + R') / 2
= (M+S×width + M-S×width) / 2
= (2M) / 2
= M
Perfect mono compatibility: Side signal completely cancels
But: If S contains low frequencies with phase issues → partial cancellation
Solution: High-pass side signal (200Hz cutoff in safety mode)
Phantom Bass Enhancement
Stage 1: Bass Extraction
🔊 Isolating Low Frequencies
Bandpass filtering: Extract only the frequency range that will generate harmonics
| Parameter | Typical Range | Effect | Recommendation |
|---|---|---|---|
| bass_low_freq | 20-60 Hz | Lowest frequency extracted | 30 Hz for most music, 40 Hz for speech |
| bass_high_freq | 80-150 Hz | Highest frequency extracted | 120 Hz balances harmonic generation |
Process per channel:
bass_L = Filter (pass Hann band): bass_low_freq, bass_high_freq, 100
bass_R = Filter (pass Hann band): bass_low_freq, bass_high_freq, 100
Filter characteristics:
Hann window: Smooth frequency response
Bandwidth: 100 Hz transition bands
Purpose: Clean extraction without excessive ringing
Design rationale: Extract only frequencies that benefit from harmonic generation. Too low (<30Hz) may not generate audible harmonics. Too high (>150Hz) may conflict with existing midrange.
Stage 2: Harmonic Generation via Waveshaping
Harmonic generation formula:
harmonics = tanh(drive × bass_signal)
Where:
tanh(x) = (e^x - e^{-x}) / (e^x + e^{-x}) (hyperbolic tangent)
drive = intensity control (0.1-10)
Mathematical properties:
tanh(x) ≈ x for small x (low drive = minimal distortion)
tanh(x) → ±1 as x → ±∞ (soft clipping)
Generates both odd and even harmonics
Frequency domain effect:
Input: pure sine at frequency f
Output: f + 2f + 3f + 4f + ... (harmonic series)
Amplitude decreases with harmonic number
Stage 3: Harmonic Filtering
Controlling harmonic brightness: Raw waveshaping generates harmonics far beyond the audible range. A lowpass filter controls how "bright" or "warm" the harmonics sound.
| Parameter | Typical Range | Sonic Character |
|---|---|---|
| harmonic_lowpass = 400 Hz | 400-600 Hz | Warm, subtle harmonics |
| harmonic_lowpass = 800 Hz | 700-900 Hz | Balanced, natural |
| harmonic_lowpass = 1200 Hz | 1000-1500 Hz | Bright, pronounced |
| harmonic_lowpass = 2000 Hz | 1500-2500 Hz | Aggressive, distorted |
Filter implementation:
harmonics_filtered = Filter (pass Hann band): bass_high_freq, harmonic_lowpass, 100
Why bandpass, not just lowpass?
- Removes any residual fundamental below bass_high_freq
- Clean transition from original to harmonics
- Prevents low-frequency "mud"
Stage 4: Original High-Pass Filtering
Critical step: Remove conflicting bass frequencies from original signal
- Purpose: Prevent original bass from masking generated harmonics
- Without this: Bass becomes boomy/muddy, harmonics less effective
- Parameter:
highpass_freqtypically 80-120 Hz - Rule: Should be near or slightly below
bass_high_freq
original_highpassed = Filter (stop Hann band): 0, highpass_freq, 100
Filter type: High-pass (stop band 0 to highpass_freq)
Transition: 100 Hz Hann window (smooth rolloff)
Result: Original with bass removed, mids/highs intact
Common mistake: Setting highpass_freq too high (>150Hz) removes too much body from audio, making it sound thin.
Stage 5: Harmonic Mixing
Final mix per channel:
result = original_highpassed + harmonics_filtered × harmonic_mix
Where:
harmonic_mix = 0.0-1.0 (dry/wet balance)
Typical values:
0.0-0.3: Very subtle enhancement
0.4-0.6: Balanced, natural effect (default)
0.7-0.9: Strong, obvious enhancement
1.0: Full wet (rarely used)
Perceptual effect:
Low mix: Bass feels "fuller" but not distorted
High mix: Obvious harmonic enhancement, "MaxxBass" style
Parameter Interactions
Parameter synergy examples:
Warm, subtle bass: bass_low_freq=40, bass_high_freq=100 drive=2.0, harmonic_mix=0.4 highpass_freq=90, harmonic_lowpass=500 Result: Gentle fullness
Aggressive modern bass: bass_low_freq=30, bass_high_freq=120 drive=5.0, harmonic_mix=0.8 highpass_freq=120, harmonic_lowpass=1000 Result: Pronounced "hyperreal" bass
Speech/clarity focused: bass_low_freq=60, bass_high_freq=150 drive=1.5, harmonic_mix=0.3 highpass_freq=120, harmonic_lowpass=400 Result: Vocal presence without boominess
Warm, subtle bass: bass_low_freq=40, bass_high_freq=100 drive=2.0, harmonic_mix=0.4 highpass_freq=90, harmonic_lowpass=500 Result: Gentle fullness
Aggressive modern bass: bass_low_freq=30, bass_high_freq=120 drive=5.0, harmonic_mix=0.8 highpass_freq=120, harmonic_lowpass=1000 Result: Pronounced "hyperreal" bass
Speech/clarity focused: bass_low_freq=60, bass_high_freq=150 drive=1.5, harmonic_mix=0.3 highpass_freq=120, harmonic_lowpass=400 Result: Vocal presence without boominess
Haas Effect Implementation
Basic Haas Algorithm
⏱️ Channel Delay Technique
Standard implementation: Delay one channel slightly relative to the other
Right channel processing (typical):
delayed_right[n] = right[n - D] × level_factor
Where:
D = delay_samples = round(haas_delay_ms / 1000 × sampleRate)
level_factor = 10^(level_difference_db / 20)
Mix with original:
right_result = (1 - haas_mix) × right + haas_mix × delayed_right
left_result = left (unchanged)
Why usually delay right channel? Convention, not requirement. Can swap L/R for different spatial effects.
Haas Parameters
| Parameter | Range | Effect | Mono-Safe Limits |
|---|---|---|---|
| haas_delay_ms | 1-50 ms | Delay time; longer = wider but risk echo | ≤12 ms |
| haas_mix | 0.0-1.0 | Wet/dry balance; higher = stronger effect | ≤0.25 |
| level_difference_db | -6 to +6 dB | Delayed channel level relative to original | -3 to 0 dB |
Mono Compatibility Issues
⚠️ Phase Cancellation Problem
What happens in mono? Left and right channels sum:
Mono_sum = left + [(1-mix)×right + mix×right_delayed]
= left + right + mix×(right_delayed - right)
If left ≈ right (typical stereo material):
Mono_sum ≈ 2×right + mix×(right_delayed - right)
Frequency cancellation:
For frequency f with period T:
right_delayed(t) = right(t - Δt)
Difference creates comb filtering with notches at:
f = k / Δt (k = 1, 2, 3...)
Example: Δt = 15ms → notches at 67Hz, 133Hz, 200Hz...
Safety mode implementation:
if preserve_mono_compatibility = 1:
haas_delay_actual = min(haas_delay_ms, 12)
haas_mix_actual = min(haas_mix, 0.25)
Rationale:
- Shorter delays (<12ms) push cancellation notches higher
- Lower mix (<25%) reduces cancellation depth
- Combined: Minimizes audible phasing in mono
Practical Haas Guidelines
Recommended settings by application:
- Headphones only: 15-30ms delay, 0.4-0.7 mix
- Stereo speakers: 10-20ms delay, 0.3-0.5 mix
- Broadcast/streaming: 5-12ms delay, 0.1-0.25 mix (with safety mode)
- Club/PA systems: Avoid or use ≤5ms with safety mode
- Music production: 10-20ms delay, automate mix for effect
When to avoid Haas:
- Material that will definitely be played in mono
- Vocals or center-panned elements (can cause "phasiness")
- Already wide stereo mixes (can become exaggerated)
- Material with strong transients (can blur timing)
Mid-Side Widening
M/S Encoding/Decoding
🎛️ Stereo Representation
Mid (M) and Side (S) components:
Encoding (L,R → M,S):
M = (L + R) / 2
S = (L - R) / 2
Decoding (M,S → L,R):
L = M + S
R = M - S
Properties:
- M contains all mono-compatible information
- S contains all stereo difference information
- In mono: L+R = 2M (S cancels completely)
Width control: Amplify or attenuate the side component:
Width adjustment:
L' = M + S × width
R' = M - S × width
Where:
width = 0.0: mono (S completely removed)
width = 0.5: narrower than original
width = 1.0: original width
width = 1.5: 50% wider
width = 2.0: stereo extremes
Implementation Details
Script implementation:
# Calculate mid
mid = (left_mixed + right_mixed) / 2
# Calculate side with width control
side = (left_mixed - right_mixed) / 2 × stereo_width
# Mono safety: high-pass side signal
if preserve_mono_compatibility = 1:
side_filtered = Filter (stop Hann band): 0, 200, 100
# Removes low-frequency phase issues
# Reconstruct
left_result = mid + side_filtered
right_result = mid - side_filtered
Why High-Pass the Side Signal?
Low-frequency phase issues: In real-world recordings, low frequencies (<200Hz) often have random phase relationships between channels due to:
- Room acoustics (standing waves, reflections)
- Microphone placement differences
- Bass management in mixing
- Proximity effect variations
Problem: These random phase differences create unpredictable cancellation when:
Mono_sum = L' + R' = 2M + (S - S) = 2M (theoretically)
But with phase differences at low frequencies:
S_low ≠ (L_low - R_low)/2 in phase-coherent way
Result: Partial cancellation, thin bass in mono
Solution: High-pass side signal at 200Hz removes problematic low-frequency phase issues while preserving stereo width at mid/high frequencies where phase coherence is better.
Width Parameter Guidelines
| stereo_width | Perceived Effect | Recommended Use | Mono Safety |
|---|---|---|---|
| 0.0-0.3 | Narrower than original | Focusing center, reducing width | Safe |
| 0.4-0.6 | Moderate widening | General enhancement, most music | Safe with HPF |
| 0.7-0.9 | Significant widening | Electronic music, creative effects | Risky, test in mono |
| 1.0-1.3 | Extreme widening | Special effects, sound design | Not safe |
| 1.4+ | Hyper-wide, possibly inverted | Experimental only | Definitely not safe |
Parameters & Presets
Preset System
Preset 1: Custom
Manual parameter control only
Preset 2: Subtle Enhancement
drive=2.0, harmonic_mix=0.4, highpass_freq=90, harmonic_lowpass=600
haas: 10ms delay, 30% mix
MS width: 0.3, mono-safe: ON
Use: Gentle enhancement for natural material
Preset 3: Moderate Effect (Default)
drive=3.0, harmonic_mix=0.6, highpass_freq=100, harmonic_lowpass=800
haas: 15ms delay, 50% mix
MS width: 0.5, mono-safe: ON
Use: Balanced enhancement for most music
Preset 4: Aggressive MaxxBass Style
drive=5.0, harmonic_mix=0.8, highpass_freq=120, harmonic_lowpass=1000
haas: 20ms delay, 60% mix
MS width: 0.7, mono-safe: OFF
Use: Strong effect for EDM, hip-hop, headphones
Preset 5: Mono-Safe Widening
drive=2.5, harmonic_mix=0.5, highpass_freq=100, harmonic_lowpass=700
haas: 8ms delay, 20% mix
MS width: 0.4, mono-safe: ON
Use: Broadcast, streaming, club systems
Preset 6: Wide Stereo (not mono-safe)
drive=3.5, harmonic_mix=0.65, highpass_freq=100, harmonic_lowpass=900
haas: 25ms delay, 70% mix
MS width: 0.8, mono-safe: OFF
Use: Headphones, stereo-only playback, immersive effects
Manual parameter control only
Preset 2: Subtle Enhancement
drive=2.0, harmonic_mix=0.4, highpass_freq=90, harmonic_lowpass=600
haas: 10ms delay, 30% mix
MS width: 0.3, mono-safe: ON
Use: Gentle enhancement for natural material
Preset 3: Moderate Effect (Default)
drive=3.0, harmonic_mix=0.6, highpass_freq=100, harmonic_lowpass=800
haas: 15ms delay, 50% mix
MS width: 0.5, mono-safe: ON
Use: Balanced enhancement for most music
Preset 4: Aggressive MaxxBass Style
drive=5.0, harmonic_mix=0.8, highpass_freq=120, harmonic_lowpass=1000
haas: 20ms delay, 60% mix
MS width: 0.7, mono-safe: OFF
Use: Strong effect for EDM, hip-hop, headphones
Preset 5: Mono-Safe Widening
drive=2.5, harmonic_mix=0.5, highpass_freq=100, harmonic_lowpass=700
haas: 8ms delay, 20% mix
MS width: 0.4, mono-safe: ON
Use: Broadcast, streaming, club systems
Preset 6: Wide Stereo (not mono-safe)
drive=3.5, harmonic_mix=0.65, highpass_freq=100, harmonic_lowpass=900
haas: 25ms delay, 70% mix
MS width: 0.8, mono-safe: OFF
Use: Headphones, stereo-only playback, immersive effects
Phantom Bass Parameters
| Parameter | Type | Default | Range | Description |
|---|---|---|---|---|
| bass_low_freq | positive | 30 | 20-100 Hz | Lowest frequency for harmonic generation |
| bass_high_freq | positive | 120 | 60-200 Hz | Highest frequency for harmonic generation |
| drive | positive | 3.0 | 0.1-10.0 | Harmonic generation intensity |
| harmonic_mix | real | 0.6 | 0.0-1.0 | Dry/wet mix of harmonics |
| highpass_freq | positive | 100 | 50-200 Hz | Original bass removal cutoff |
| harmonic_lowpass | positive | 800 | 400-2000 Hz | Harmonic brightness control |
Haas Effect Parameters
| Parameter | Type | Default | Range | Description |
|---|---|---|---|---|
| apply_haas | boolean | 1 | 0/1 | Enable/disable Haas effect |
| haas_delay_ms | positive | 15 | 1-50 ms | Delay time for one channel |
| haas_mix | real | 0.5 | 0.0-1.0 | Wet/dry balance for Haas |
| level_difference_db | real | -3 | -6 to +6 dB | Delayed channel level relative to original |
Mid-Side Parameters
| Parameter | Type | Default | Range | Description |
|---|---|---|---|---|
| apply_ms_widening | boolean | 1 | 0/1 | Enable/disable M/S widening |
| stereo_width | real | 0.5 | 0.0-2.0 | Stereo width multiplier |
Safety Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
| preserve_mono_compatibility | boolean | 1 | Enable safety features for mono playback |
| play_result | boolean | 1 | Auto-play processed sound |
Processing Stages
Complete Processing Pipeline
1. INPUT PREPARATION
Store original sound
Convert mono to stereo if needed
Extract left/right channels
2. PHANTOM BASS PROCESSING (per channel)
Bass extraction: bandpass(bass_low_freq, bass_high_freq)
Harmonic generation: tanh(drive × bass)
Harmonic filtering: bandpass(bass_high_freq, harmonic_lowpass)
Original high-pass: highpass(highpass_freq)
Mix: original_highpassed + harmonics × harmonic_mix
3. PEAK SAFETY PASS 1
Check max peak after harmonic mix
Scale if >0.95 to prevent clipping
4. HAAS EFFECT (if enabled)
Create delayed version of right channel
Mix: right_result = (1-mix)×right + mix×right_delayed
Apply mono safety limits if preserve_mono_compatibility=1
5. PEAK SAFETY PASS 2
Check max peak after Haas
Scale if >0.95
6. MID-SIDE WIDENING (if enabled)
Calculate: mid = (L+R)/2, side = (L-R)/2 × width
If mono-safe: high-pass side at 200Hz
Reconstruct: L' = mid + side, R' = mid - side
7. PEAK SAFETY PASS 3
Check max peak after M/S
Scale if >0.95
8. FINAL ASSEMBLY
Combine L'/R' to stereo sound
Scale peak to 0.99
Rename: originalname_phantom_haas
Play if requested
Print processing summary
Peak Protection System
Three-stage peak monitoring: The script checks and normalizes peaks at three critical points:
- After harmonic mixing: Harmonic addition can increase peaks
- After Haas processing: Delay-and-mix can create constructive interference
- After M/S widening: Side signal boosting can increase levels
Algorithm:
For each safety pass:
max_left = Get maximum of left channel
min_left = Get minimum of left channel
max_right = Get maximum of right channel
min_right = Get minimum of right channel
max_peak = max(abs(max_left), abs(min_left),
abs(max_right), abs(min_right))
if max_peak > 0.95:
scale_factor = 0.95 / max_peak
left = left × scale_factor
right = right × scale_factor
Final normalization: After all processing, scale peak to 0.99 (leaves 0.01 headroom for downstream processing).
Practical Applications
🎧 Headphone Listening Enhancement
Goal: Create immersive, bass-rich experience for headphones
Preset: Aggressive MaxxBass Style or Wide Stereo
Rationale: Headphones have no room acoustics/crossfeed, benefit from strong spatial enhancement and bass harmonics.
📻 Podcast/Voice Enhancement
Goal: Warm, present voice without boominess
Settings:
- bass_low_freq=80, bass_high_freq=200 (voice fundamentals)
- drive=1.5, harmonic_mix=0.3 (subtle)
- Haas: 5ms delay, 15% mix (minimal)
- MS width: 0.2 (slight widening)
- mono-safe: ON (broadcast essential)
🎵 Music Mastering for Streaming
Goal: Consistent impact across all playback systems
Preset: Mono-Safe Widening (modified)
Additional steps: Always A/B with reference tracks, test on phone speakers, car systems, and collapsed to mono.
Troubleshooting
Problem: Thin/weak bass in mono
Cause: Phase cancellation from Haas or M/S processing
Solution: Enable preserve_mono_compatibility, reduce Haas delay/mix, test in mono
Cause: Phase cancellation from Haas or M/S processing
Solution: Enable preserve_mono_compatibility, reduce Haas delay/mix, test in mono
Problem: Distorted/unpleasant harmonics
Cause: Drive too high, harmonic_lowpass too high
Solution: Reduce drive to 2.0-3.0, lower harmonic_lowpass to 600-800Hz
Cause: Drive too high, harmonic_lowpass too high
Solution: Reduce drive to 2.0-3.0, lower harmonic_lowpass to 600-800Hz
Problem: Echo/comb filtering artifacts
Cause: Haas delay too long (>30ms)
Solution: Reduce haas_delay_ms to 10-20ms, lower haas_mix
Cause: Haas delay too long (>30ms)
Solution: Reduce haas_delay_ms to 10-20ms, lower haas_mix
Problem: Audio sounds "hollow" or phasey
Cause: Too much side signal, especially at low frequencies
Solution: Reduce stereo_width to 0.3-0.5, ensure mono-safe mode enabled
Cause: Too much side signal, especially at low frequencies
Solution: Reduce stereo_width to 0.3-0.5, ensure mono-safe mode enabled
Problem: Clipping despite peak protection
Cause: Extreme settings, inter-sample peaks
Solution: Reduce all levels, use more conservative settings, add external limiter
Cause: Extreme settings, inter-sample peaks
Solution: Reduce all levels, use more conservative settings, add external limiter
Best Practices
Workflow recommendations:
- Start with presets: Use Moderate Effect as baseline
- A/B constantly: Compare processed vs original
- Test in mono: Collapse to mono after processing
- Test on different systems: Headphones, laptop speakers, car
- Use reference material: Compare with professionally processed tracks
- Less is more: Subtle enhancement often works better than extreme
- Consider downstream processing: Leave headroom for further mastering
When to use (and when to avoid):
Good candidates:
- Mixes lacking low-end impact
- Material for headphone listening
- Audio for small speaker playback
- Stereo enhancement for narrow mixes
- Broadcast audio needing consistent bass
Avoid or use cautiously:
- Already bass-heavy material (can become muddy)
- Acoustic recordings needing naturalness
- Material with strong stereo effects already
- Audio destined primarily for mono playback
- Mastering for vinyl (phase issues critical)