Advanced Bass Architecture and Low-Frequency Psychoacoustics in Generative Audio

Full report

On this page

1. Introduction to Generative Low-Frequency Architecture

The engineering of low-frequency sound within generative audio systems occupies a uniquely challenging intersection of digital signal processing (DSP), physical electroacoustics, and human psychoacoustics. Unlike linear audio media—where a mastering engineer can surgically automate problematic sub-bass notes, correct phase rotations, and limit transient peaks ahead of time—generative audio is algorithmic, unbounded, and unpredictable. A generative system might autonomously combine a [Figure omitted from source export] sine wave drone with a deeply pitched kick drum and a low-frequency granular synthesis texture. Without a deterministic architectural framework, this combination will inevitably cause catastrophic digital clipping, induce severe upward spread of masking, and potentially destroy physical reproduction equipment through thermal overload or mechanical over-excursion.

Consequently, managing bass in a generative context cannot rely on reactive human intervention. It requires an autonomous, mathematically rigorous topology that anticipates the limitations of the playback medium while manipulating human auditory perception to maintain the artistic impact of immense scale. This engineering study provides an exhaustive examination of bass and low-frequency design for mesmerizing generative audio. It investigates the physiological mechanisms of auditory and tactile perception, the hard acoustic constraints of transducers and listening environments, and the strict mathematical boundaries required to manage sub-bass. By synthesizing principles from psychoacoustic masking models, non-linear harmonic generation, and deterministic DSP limiters, this document establishes a comprehensive strategy to ensure low-end complexity translates gracefully across any device, from cellular phone speakers to dedicated cinematic subwoofers.

2. Psychoacoustics and the Perception of Low Frequencies

To engineer generative bass effectively, one must discard the assumption that the human auditory system acts as a flat-response microphone. The ear is highly non-linear, particularly at the extreme low end of the frequency spectrum ([Figure omitted from source export] to [Figure omitted from source export]). The perception of low-frequency sound is a composite experience formed by cochlear frequency analysis, neural autocorrelation, and vibrotactile somatosensory feedback.

2.1 ISO 226 and Equal-Loudness Contours

The human ear's sensitivity varies drastically depending on both the frequency and the sound pressure level (SPL). The ISO 226:2003 standard (and subsequent 2023 revisions) defines "Normal Equal-Loudness-Level Contours," which map the physical SPL required for pure continuous tones to be perceived as equally loud across the audible spectrum by listeners with normal hearing1. The perceptual unit of loudness is the phon, anchored specifically so that at [Figure omitted from source export], the phon value exactly equals the SPL in decibels4.

At low frequencies, the auditory system exhibits a severe sensitivity penalty. The ISO 226 mathematical model provides a non-linear transfer function to compute the required SPL ([Figure omitted from source export]) for a given frequency ([Figure omitted from source export]) and target loudness level ([Figure omitted from source export] in phons). The function utilizes frequency-dependent coefficients, including the threshold of hearing ([Figure omitted from source export]) and the normalized magnitude transfer function ([Figure omitted from source export]):

[Figure omitted from source export]

Where [Figure omitted from source export] is an intermediate quantity defined as:

[Figure omitted from source export]

In practice, this non-linear scaling means that to perceive a [Figure omitted from source export] tone at a quiet loudness of [Figure omitted from source export], the physical SPL must be approximately [Figure omitted from source export]—nearly [Figure omitted from source export] louder than a [Figure omitted from source export] tone at the exact same perceived loudness6. Conversely, as SPL increases to [Figure omitted from source export] or [Figure omitted from source export], the contours flatten, meaning the disparity between low-frequency and mid-frequency sensitivity decreases6.

This phenomenon dictates that any generative audio application must implement equal-loudness compensation algorithms. If a user reduces the master playback volume, the generative sub-bass will perceptually vanish much faster than the midrange, dismantling the illusion of scale. The generative engine must dynamically boost low frequencies as the overall SPL decreases to preserve the intended tonal balance3.

2.2 The Upward Spread of Masking and Critical Bands

When multiple frequencies are generated simultaneously, they interact within the cochlea. The human ear processes sound in discrete frequency channels known as critical bands, which can be mathematically approximated by the Equivalent Rectangular Bandwidth (ERB) scale10. The bandwidth of an auditory filter at a given center frequency ([Figure omitted from source export], in Hz) is modeled by the Glasberg and Moore formula:

[Figure omitted from source export]

At [Figure omitted from source export], the ERB is exceedingly narrow (approximately [Figure omitted from source export]), but it widens significantly at higher frequencies11. Because of the asymmetrical shape of the traveling wave on the basilar membrane, low frequencies mask high frequencies much more effectively than the reverse—a phenomenon termed the "upward spread of masking"11. As the amplitude of a low-frequency signal increases, its masking slope becomes shallower, obscuring a wider range of high-frequency information11.

If a generative system produces continuous, dense low-mid or sub-bass drones without strict attenuation, it will catastrophically mask the intelligibility of higher-frequency melodic, harmonic, or transient elements14. Consequently, low-frequency content must be deeply sculpted, primarily transient-driven, or spectrally sparse in its arrangement to prevent the upward spread of masking from collapsing the clarity of the generative mix.

3. The Missing Fundamental and Psychoacoustic Bass Synthesis

Translating deep generative bass to devices entirely incapable of reproducing sub-sonic frequencies (such as laptop speakers, mobile phones, or smart devices) requires the exploitation of auditory illusions. The most critical of these is the "missing fundamental" phenomenon.

3.1 Neural Pitch Perception and the Frequency-Following Response

If a sound system cannot physically reproduce a [Figure omitted from source export] fundamental, the human brain will still perceive a [Figure omitted from source export] pitch if presented with a phase-coherent series of its upper harmonics (e.g., [Figure omitted from source export])17. This illusion, also known as "virtual pitch" or "residue pitch," is driven by complex neural mechanisms rather than simple spectral place-theory in the cochlea17.

Licklider's duplex theory of pitch, alongside subsequent autocorrelation models by Meddis and Hewitt, posits that pitch is determined by time-domain processing within the auditory nerve20. The auditory system evaluates the time intervals between neural spikes (Inter-Spike Intervals) across multiple critical bands, identifies the common temporal periodicity, and reconstructs the fundamental pitch even when the physical energy is absent22.

Recent neuroimaging using high-density electroencephalography (EEG) and magnetoencephalography (MEG) to measure the Frequency-Following Response (FFR) has confirmed these mechanisms. The FFR is an electrophysiological potential that phase-locks to the periodic fluctuations of sound waves24. While traditionally viewed as a purely subcortical response generated in the brainstem (e.g., the inferior colliculus), modern neuroimaging reveals significant right-lateralized contributions from the auditory cortex26. This distributed neural network actively integrates harmonic partials to generate the perception of the missing fundamental26.

3.2 Psychoacoustic Bass Enhancement Algorithms

Generative audio systems must integrate psychoacoustic bass synthesis algorithms to ensure translation. Pioneered by Meir Shashoua (who patented the MaxxBass algorithm in 1999\) and further formalized by Larsen and Aarts in Audio Bandwidth Extension, these algorithms actively isolate un-reproducible sub-bass and replace it with synthesized harmonics18.

There are two primary DSP topologies for generating these harmonics, each with distinct advantages and drawbacks for generative audio:

 

To optimize generative audio, the system should employ a hybrid approach. The low-frequency signal is processed through median filters to separate it into transient and tonal streams. Transients are routed through an NLD (soft-clipping) to preserve impact, while continuous tonal drones are processed through a phase vocoder to maintain harmonic purity. The streams are then scaled based on equal-loudness contours and blended back into the main signal above the physical cutoff frequency of the target speaker31.

4. Vibrotactile Perception and Headphone Behavior

True sub-bass ([Figure omitted from source export] to [Figure omitted from source export]) traverses the boundary between auditory sensation and physical feeling. In environments capable of reproducing deep low frequencies, the generative experience becomes somatosensory.

4.1 Mechanoreceptors and Cross-Modal Enhancement

Vibrotactile perception is mediated by specific mechanoreceptors in the human skin and fascia, which respond to distinct frequency bands:

• Meissner corpuscles: Located just beneath the epidermis, these are highly sensitive to low-frequency vibrations ranging from [Figure omitted from source export] to [Figure omitted from source export], with peak sensitivity around [Figure omitted from source export] to [Figure omitted from source export]37.
• Pacinian corpuscles: Larger receptors located deeper in the dermis, sensitive to higher-frequency vibrations from [Figure omitted from source export] up to [Figure omitted from source export]37.

When acoustic energy interacts with the body, it triggers these mechanoreceptors, leading to a phenomenon known as "cross-modal enhancement." Research demonstrates that the presence of congruent tactile vibration can bias auditory loudness judgments, effectively increasing the perceived auditory dynamic range and apparent loudness by up to [Figure omitted from source export]41.

In a generative system, this implies that immense scale can be perceived without sustaining continuous, high-SPL [Figure omitted from source export] drones. By utilizing targeted, transient low-frequency bursts that trigger Meissner corpuscles, the brain synthesizes a massive physical impact. Consequently, the generative engine can lower the average sustained airborne sub-bass energy by [Figure omitted from source export] to [Figure omitted from source export], preserving digital headroom and minimizing neighbor disturbance while maintaining the psychological illusion of scale44.

4.2 Headphone Constraints and Bone Conduction

Headphones present a unique environment for generative bass. Because they decouple the acoustic energy from the room and the listener's torso, they eliminate the broad tactile stimulation of the chest cavity. However, they introduce bone conduction directly into the skull41.

When low-frequency actuators or transducer housing vibrate against the temporal bone, the energy bypasses the outer ear and directly stimulates the cochlea. Stenfelt's research indicates that for every [Figure omitted from source export] increase in airborne auditory stimuli, there is a corresponding rise of [Figure omitted from source export] in bone conduction to achieve similar loudness41. Generative systems must account for this by understanding that sub-bass in closed-back headphones may feel physically intense due to bone conduction, but will lack the holistic full-body impact of a room environment. Psychoacoustic harmonics (virtual pitch) become critical in headphones to trick the brain into perceiving the spatial scale that the body cannot feel41.

5. Physical Acoustics and Reproduction Constraints

Translating generative sub-bass into the physical world introduces severe mechanical, thermal, and environmental limitations. Failing to account for these constraints will result in destroyed loudspeakers and degraded sonic experiences.

5.1 Transducer Limitations: Excursion and Thermal Compression

To reproduce low frequencies at high sound pressure levels, a loudspeaker cone must displace a massive volume of air. This pushes the electroacoustic transducer toward two primary failure states: mechanical over-excursion and thermal overload.

Mechanical Excursion Limits ([Figure omitted from source export]): [Figure omitted from source export] defines the maximum linear peak excursion of the driver's voice coil, typically evaluated at a threshold of 10% Total Harmonic Distortion (THD) using a sinusoidal signal46. If generative sub-bass forces the driver beyond [Figure omitted from source export], the voice coil leaves the linear magnetic gap. The driver's force factor ([Figure omitted from source export]) and suspension compliance ([Figure omitted from source export]) drop to less than 50% of their small-signal values46. This nonlinear behavior results in severe intermodulation distortion (IMD), specifically Doppler distortion, where the massive, rapid movement of the cone modulates the frequency of any higher-pitched sounds playing simultaneously46.

Thermal Power Compression: Continuous sub-bass (e.g., a sustained generative drone) demands massive continuous current. As electrical energy flows through the voice coil, it heats up rapidly, often exceeding [Figure omitted from source export] to [Figure omitted from source export] under heavy load50. Because the electrical resistance of conductive metals (copper, aluminum) increases with temperature, the impedance of the voice coil rises, reducing current flow51. This results in "thermal power compression." A [Figure omitted from source export] increase in resistance yields a [Figure omitted from source export] drop in acoustic output—a massive distortion of the intended macro-dynamics51. Furthermore, in passive loudspeaker systems, this shift in impedance alters the crossover frequencies, completely ruining the phase alignment between drivers51.

Furthermore, modern active subwoofers overwhelmingly rely on Class D amplifiers. While highly efficient (often \>90%) and generating less heat than Class AB designs, Class D amplifiers must deal with reactive loads53. The massive back-EMF (electromotive force) generated by a heavy subwoofer cone returning to its resting position can pump energy back into the amplifier's power supply. Unmanaged continuous generative drones can destabilize these supply rails55.

5.2 Room Modes and the Schroeder Frequency

Once acoustic energy leaves the transducer, it is heavily modified by the geometry of the listening environment. In enclosed spaces, the behavior of sound is bifurcated by the Schroeder Frequency ([Figure omitted from source export]), which marks the transition between discrete modal resonances and a statistically uniform reverberant field57.

The Schroeder Frequency is calculated as:

[Figure omitted from source export]

(where [Figure omitted from source export] is the reverberation time in seconds and [Figure omitted from source export] is the room volume in cubic meters).

Below [Figure omitted from source export] (typically [Figure omitted from source export] to [Figure omitted from source export] in domestic rooms), the modal density is low58. Continuous sub-bass tones will inevitably align with the room's axial, tangential, or oblique modes, establishing standing waves58. This results in massive spatial variance; one listener may be seated in an anti-node and experience a [Figure omitted from source export] resonant boost, while another listener a meter away sits in a node and experiences total phase cancellation61.

Because of this, generative systems should avoid long, static sub-bass drones. The resonance of a room mode takes time to build up to its steady-state maximum. If the generative engine utilizes transient low-frequency impacts or constantly evolving pitches, the acoustic energy will not sit still long enough to fully excite the standing waves, resulting in a much smoother perceived frequency response across the listening area.

6. Digital Signal Processing Architecture and Headroom

To format low-frequency energy safely for physical reproduction, the generative engine must employ strict DSP rulesets governing headroom, phase, dynamic range, and mono compatibility.

6.1 High-Pass Filtering, DC Offset, and Phase Rotation

Protecting digital headroom and preventing mechanical over-excursion requires strict high-pass filtering (HPF). The generative engine must aggressively cull inaudible subsonic frequencies ([Figure omitted from source export]) and DC offset. DC offset acts as a silent, constant voltage that forces speaker cones away from their resting position, causing asymmetric waveform distortion, reducing maximum SPL, and consuming vital limiter headroom63.

However, filtering low frequencies introduces significant phase challenges. The choice of filter topology is critical:

• Minimum Phase Filters (IIR): Standard infinite impulse response filters alter the phase of the signal. Because low frequencies oscillate slower than high frequencies, minimum phase filters delay frequencies unequally (phase rotation)68. At the cutoff frequency (e.g., [Figure omitted from source export]), maximum phase rotation occurs. This can cause the peaks of disparate frequencies to align constructively, inadvertently increasing the True Peak of the audio signal by several decibels without increasing perceived loudness69.
• Linear Phase Filters (FIR): These finite impulse response filters maintain perfect phase alignment by delaying all frequencies equally70. However, to accomplish this symmetry, they introduce "pre-ringing"—a temporal smear or ripple that precedes the transient. At low frequencies (e.g., a [Figure omitted from source export] cutoff with steep [Figure omitted from source export] slopes), this pre-ringing can last [Figure omitted from source export], severely softening kick drums and creating an audible "swoosh" or brown-noise artifact before the initial impact69.

Architectural Rule: Generative systems must avoid linear phase for sub-bass high-pass filtering. The pre-ringing is too destructive to transient physicality69. Instead, utilize minimum-phase IIR filters with gentle slopes ([Figure omitted from source export] to [Figure omitted from source export]) to minimize extreme phase rotation68. If specific phase alignment between internal generative sub-systems is required, utilize dedicated all-pass filters to manage group delay without altering magnitude75. When separating bass for internal multi-band processing, employ 4th-order Linkwitz-Riley crossovers, which sum to flat magnitude and keep the split signals in phase with one another, despite introducing a global phase shift77.

6.2 Transient Management and Compressor Interaction

Applying dynamic range compression to low frequencies requires specialized temporal configurations to prevent digital distortion. The physical period of a low-frequency wave is long; a [Figure omitted from source export] sine wave takes [Figure omitted from source export] to complete a single cycle79.

If a compressor's attack time is set too fast (e.g., [Figure omitted from source export]), the Voltage Controlled Amplifier (VCA) will react within a single cycle of the waveform. This effectively clips the sound wave, morphing a smooth sine wave into a squared-off wave and generating aggressive, audible harmonic distortion79. To preserve the physicality of low-frequency transients in generative audio:

1. Compressor attack times on bass busses must generally exceed [Figure omitted from source export] to allow the full transient waveform to pass unclipped79.

2. If strict peak limitation is required, the compressor must utilize a digital "lookahead" buffer of [Figure omitted from source export]. This allows the DSP to evaluate the signal in advance and smoothly ramp down the gain envelope before the transient arrives, thereby avoiding digital clicking79.

3. Sidechain compression (ducking) is mandatory to separate transient impacts (e.g., kicks) from sustained sub-bass drones. When a kick transient fires, the continuous sub-bass must duck gracefully via a synchronized release time, preventing low-mid masking and preventing the combined signals from exceeding [Figure omitted from source export]80.

6.3 Mono Compatibility and Elliptical Equalization

Low frequencies exhibit exceedingly large acoustic wavelengths and radiate omnidirectionally44. Hard-panning sub-bass creates severe phase cancellation in the playback environment, stresses left and right stereo amplifiers unequally, and causes physical needle-jumping on vinyl reproduction82.

The generative system must employ an Elliptical EQ (effectively a high-pass filter applied exclusively to the Side channel in a Mid/Side matrix) set between [Figure omitted from source export] and [Figure omitted from source export]66. This algorithm mathematically forces all sub-bass and mid-bass frequencies into dead-center, phase-accurate mono while allowing the upper midrange and high-frequency layers to remain infinitely wide82.

7. The Five-Layer Internal Bass Architecture

To achieve a mesmerizing, perceived scale without relying on continuous, fatigue-inducing, high-level sub-bass, the generative audio system must abandon monolithic bass synthesis. The engine must adopt a deeply segmented, multi-layered approach to low-end generation, creating distinct phase-aligned internal layers, each serving a specific psychoacoustic function.

 

7.1 Broad Low-Frequency Breathing vs. Continuous Drones

To create the psychological perception of immense scale, the engine must leverage macroscopic dynamics rather than pure volume. Human sensory systems quickly habituate to continuous stimuli. If a generative [Figure omitted from source export] drone plays continuously at [Figure omitted from source export], the listener's brain filters it out as background noise within minutes, yet the physical speakers continue to suffer thermal power compression51.

By modulating the Subharmonic Motion layer via broad, generative "breathing" envelopes (e.g., massive swells lasting 8 to 16 seconds, followed by near-silence), the system creates an acoustic vacuum. When the low-frequency energy inevitably returns, the stark dynamic contrast triggers the autonomic nervous system, resulting in a profound perception of physical scale and gravity, achieved with a fraction of the total average RMS energy.

8. Deliverables: BassPolicy, Metrics, and Translation

To finalize the integration of this research into a functional generative audio engine, the following deterministic rulesets, translation strategies, and safety protocols must be implemented as the ultimate BassPolicy.

8.1 Deterministic Low-Frequency Metrics

Most modern streaming and broadcast standards rely on ITU-R BS.1770 LUFS (Loudness Units relative to Full Scale) to normalize audio90. However, the BS.1770 algorithm utilizes a K-weighting filter that intentionally rolls off low frequencies91.

If a generative system relies solely on LUFS to normalize its internal output, a track with massive sub-bass will read as computationally "quiet" to the algorithm. The system might autonomously turn the track up, resulting in catastrophic clipping and speaker damage.

• Dual-Metric Strategy: The generative engine must measure loudness using a dual-metric approach. It must calculate standard LUFS for mid-range compliance, but concurrently monitor Zwicker Loudness (ISO 532B) or ERB-based specific loudness89. Because Zwicker models critical bands and excitation patterns directly, it accurately quantifies the psychoacoustic weight of the sub-bass, allowing the generative engine to safely throttle low-end output before clipping occurs89.

8.2 Strict Digital Headroom Rules

• DC Offset Rule: A first-order ([Figure omitted from source export]) minimum-phase high-pass filter must sit permanently at [Figure omitted from source export] on the master bus to remove DC offset and subsonic asymmetry65.
• ISP / True Peak Limiting: Standard digital limiters only measure sample peaks. Because Digital-to-Analog (D/A) converters reconstruct the continuous analog waveform, "inter-sample peaks" (ISPs) can overshoot digital zero, causing severe low-end distortion93. The final output stage must feature an oversampled ([Figure omitted from source export] or [Figure omitted from source export]) True Peak limiter clamping absolute output to [Figure omitted from source export]96.

8.3 Public Controls and Presets

To empower end-users without allowing them to inadvertently destroy their sound systems, user-facing controls must act as macro-wrappers for complex DSP parameters.

1. "Sub Weight" (0-100%): Does not merely turn up a bass EQ (which would consume headroom). Instead, it increases the drive into the hyperbolic tangent ([Figure omitted from source export]) saturation circuit (the Harmonic Bass Layer), creating perceived loudness via harmonic density without raising the peak SPL33.

2. "Physicality" (0-100%): Modulates the release time of the sidechain compression and the depth of the True Sub layer. Higher values create sharp, pumping transients (triggering tactile Meissner corpuscles) rather than sustained drones39.

3. "Translation" (Presets):

• Mobile: Aggressive virtual pitch via Phase Vocoder, sharp HPF at [Figure omitted from source export]31.
• Club: Moderate virtual pitch, Elliptical EQ to mono at [Figure omitted from source export], minimum phase filtering66.
• Cinematic: Maximum True Sub depth, lowest HPF cutoff ([Figure omitted from source export]), slow breathing macrodynamics.

8.4 Hardware Listening Checklist

Before deploying any iteration of the generative bass algorithm, the system output must be empirically verified against the following cross-reference matrix:

 

8.5 Recommendations for Safely Avoiding Accidental Extreme Output

In generative environments, logic loops can compound, creating rogue resonant frequencies, DC buildup, or runaway gain stages resulting from internal feedback loops100.

1. Jump Detection: Implement a slew-rate limit on the master VCA. If the algorithm generates a sudden instantaneous DC jump or a massive low-frequency transient beyond human capability, the DSP must clamp the signal to zero to protect physical transducers102.

2. Hard Ceiling Safety: The DSP chain must terminate in a brickwall hard-clipper placed after the True Peak limiter. While digital clipping introduces harsh harmonic distortion, it protects physical transducer cones from mechanical over-excursion by absolutely limiting the maximum voltage swing sent to the D/A converter102.

9. Conclusion

Designing mesmerizing, generative low-frequency audio requires abandoning the brute-force approach of simply boosting sub-bass frequencies. As this study demonstrates, the hard physics of transducer excursion, the thermodynamics of voice coils, and the acoustic chaos of room modes necessitate a highly sculpted, transient-aware DSP architecture.

By exploiting the brain's reliance on temporal autocorrelation for virtual pitch, and by leveraging the somatosensory network's sensitivity to tactile transients, a generative engine can create the illusion of massive, planetary scale while operating well within the strict headroom limits of digital audio. Implementing minimum-phase high-pass filters to preserve transients, true-peak lookahead limiting to prevent inter-sample clipping, and Zwicker-based loudness metering to prevent algorithmic over-amplification ensures that the generated audio translates gracefully. Ultimately, the meticulous orchestration of synthesized harmonics, macro-dynamic breathing, and phase-aligned layering is the definitive engineering strategy to evoke deep emotional and physical resonance across any playback medium.

Works cited

1. Equal-loudness contour \- Wikipedia, https://en.wikipedia.org/wiki/Equal-loudness\_contour

2. ISO 226:2023 Equal-Loudness Contours | PDF | Sound \- Scribd, https://www.scribd.com/document/802697448/ISO-226-2023

3. Low-Volume Mixing and the Case for ISO 226:2023, https://giovannicordova.com/blog/low-volume-mixing-iso-226-2023-equal-loudness-compensation

4. Equal-Loudness Contours (ISO 226\) Calculator \- MetricGate, https://metricgate.com/docs/equal-loudness-contours-iso226/

5. The Mysterious Loudness Control: What Does It Do? \- Extron, https://www.extron.es/article/loudnesscontrol\_ts

6. ISO 226 Normal Equal-Loudness-Level Contour Calculator, https://williamssoundstudio.com/tools/iso-226-equal-loudness-calculator-fletcher-munson.php

7. Understanding Isophonic Curves And Equal-Loudness Contours, https://www.tecnare.com/article/isophonic-curves-equal-loudness-contours-human-loudness-perception/

8. Equal loudness contours (ISO 226-2003) \- Lawrence Yule, https://lawrenceyule.com/2024/06/07/equal-loudness-contours-iso-226-2003/

9. What is the Fletcher Munson curve? Equal loudness ... \- iZotope, https://www.izotope.com/community/blog/what-is-fletcher-munson-curve-equal-loudness-curves

10. Equivalent rectangular bandwidth \- Wikipedia, https://en.wikipedia.org/wiki/Equivalent\_rectangular\_bandwidth

11. Critical band \- Wikipedia, https://en.wikipedia.org/wiki/Critical\_band

12. ERB (Glasberg-Moore) Calculator | MetricGate, https://metricgate.com/docs/erb-equivalent-rectangular-bandwidth/

13. The Role of Suppression in the Upward Spread of Masking, https://www.researchgate.net/publication/7506318\_The\_Role\_of\_Suppression\_in\_the\_Upward\_Spread\_of\_Masking

14. Frequency Masking — The Producer's Bible | MusicProductionWiki, https://musicproductionwiki.com/bible/frequency-masking

15. Psychoacoustics & Hearing \- CJS Labs, http://cjs-labs.com/sitebuildercontent/sitebuilderfiles/PsychoacousticsAndHearing.pdf

16. A psychoacoustic-masking model to predict the perception of, https://www.ee.ucla.edu/\~spapl/paper/hant\_specom03.pdf

17. Low Complexity Virtual Bass Enhancement Algorithm For Portable, https://cseweb.ucsd.edu/\~marora/files/papers/13868.pdf

18. Missing fundamental \- Wikipedia, https://en.wikipedia.org/wiki/Missing\_fundamental

19. The Application of Illusions and Psychoacoustics To Small ... \- Scribd, https://www.scribd.com/document/1021357561/The-Application-of-Illusions-and-Psychoacoustics-to-Small-Loudspeaker-Configurations

20. Correlation Network Model of Auditory Processing \! ", https://www.ee.columbia.edu/\~dpwe/workshops/crac/papers/cheveigne.pdf

21. speech f0 extraction based on licklider's, https://www.coli.uni-saarland.de/groups/FK/speech\_science/icphs/ICPhS1991/12\_ICPhS\_1991\_Vol\_4/p12.4\_218.pdf

22. (PDF) A unitary model of pitch perception \- ResearchGate, https://www.researchgate.net/publication/13920503\_A\_unitary\_model\_of\_pitch\_perception

23. Pitch Representations in the Auditory Nerve: Two Concurrent ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC2544468/

24. A high-density EEG and structural MRI source analysis of ... \- PubMed, https://pubmed.ncbi.nlm.nih.gov/37598815/

25. Frequency-Following Responses to Complex Tones at Different, https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2019.00130/full

26. Cortical Correlates of the Auditory Frequency-Following and Onset, https://pmc.ncbi.nlm.nih.gov/articles/PMC6597017/

27. Cortical correlates of the auditory frequency-following and onset, http://biorxiv.org/cgi/reprint/076448v1

28. Evolving perspectives on the sources of the frequency-following, https://brainvolts.northwestern.edu/wp-content/uploads/boxtrx/Coffey\_etal\_2019NC.pdf

29. (PDF) Audio Bandwidth Extension: Application of Psychoacoustics, https://www.researchgate.net/publication/235435103\_Audio\_bandwidth\_extension\_application\_of\_psychoacoustics\_signal\_processing\_and\_loudspeaker\_design

30. Amphitheatres \- Songstuff, https://www.songstuff.com/recording/article/amphitheatres/

31. Synthesis and Implementation of Virtual Bass System with a Phase, https://www.researchgate.net/publication/228362764\_Synthesis\_and\_Implementation\_of\_Virtual\_Bass\_System\_with\_a\_Phase-Vocoder\_Approach

32. Virtual Bass System With Fuzzy Separation of Tones and Transients, http://research.spa.aalto.fi/publications/papers/dafx20-vbs/media/DAFx2020\_vbs.pdf

33. Distortion, Saturation, and Wave Shaping \- Elementary Audio, https://www.elementary.audio/docs/tutorials/distortion-saturation-wave-shaping

34. transfer function \- Can someone explain waveshaping to me?, https://dsp.stackexchange.com/questions/46461/can-someone-explain-waveshaping-to-me

35. A VARIABLE COEFFICIENTS ADAPTIVE IIR NOTCH FILTER FOR, http://english.ioa.cas.cn/as/201409/W020140905537044221255.pdf

36. Convention Paper 10648 \- Audio Science Review (ASR) Forum, https://www.audiosciencereview.com/forum/index.php?attachments/22055-pdf.285963/

37. Vibrotactile information improves proprioceptive reaching target, https://pmc.ncbi.nlm.nih.gov/articles/PMC6034815/

38. Physiology, Vibratory Sense \- StatPearls \- NCBI Bookshelf, https://www.ncbi.nlm.nih.gov/books/NBK542288/

39. Emergence of a somatosensory tonotopic map for substrate, http://www.biorxiv.org/lookup/content/full/2023.09.26.559502v1

40. Supracutaneous Vibrotactile Perception Threshold at Various Non, https://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1432\&context=psychfacpub

41. Reducing preferred listening levels in headphones through coherent, https://pubs.aip.org/asa/jel/article/4/1/014401/3132668/Reducing-preferred-listening-levels-in-headphones

42. Perceived auditory dynamic range is enhanced with wrist-based, https://pubs.aip.org/asa/jasa/article/156/4/2759/3317723/Perceived-auditory-dynamic-range-is-enhanced-with

43. Perceived auditory dynamic range is enhanced with wrist-based, https://www.researchgate.net/publication/385142156\_Perceived\_auditory\_dynamic\_range\_is\_enhanced\_with\_wrist-based\_tactile\_stimulation

44. What Is a Bass Shaker? How It Works & When You Actually Need One, https://electronics.alibaba.com/question/bass-shaker-guide-what-it-is-how-to-use-it-right

45. Headphone Guide Pro \-, https://headphoneguidepro.com/

46. Excursion Limits \- B\&C Speakers, https://www.bcspeakers.com/en/products-overview/excursion-limits

47. X-Max and Mechanical limits of speakers \- TalkBass.com, https://www.talkbass.com/threads/x-max-and-mechanical-limits-of-speakers.920342/

48. Wolfgang KLIPPEL | Research profile \- ResearchGate, https://www.researchgate.net/profile/Wolfgang-Klippel

49. Rane Audio Reference | PDF | Analog To Digital Converter \- Scribd, https://www.scribd.com/document/105800662/Rane-Audio-Reference

50. Voice Coils: Types, Materials, Applications and Benefits \- IQS Directory, https://www.iqsdirectory.com/articles/electric-coil/voice-coils.html

51. Low Sensitivity Speakers Compression Issue., https://forum.audiogon.com/discussions/low-sensitivity-speakers-compression-issue?highlight=true%2Baudio\&page=2

52. The Power of Multiple Woofers \- QSC Cinema Blog \- Q-SYS, https://blogs.qsc.com/cinema/2020/08/18/116/

53. A Breakdown of Class D Amplifiers \- Technical Articles, https://www.allaboutcircuits.com/technical-articles/a-breakdown-of-class-d-amplifiers/

54. Class D Power Amplifier \- Power Electronics News, https://www.powerelectronicsnews.com/class-d-power-amplifier/

55. A Comprehensive Study of Class D Amplifier Technology \- TI E2E, https://e2e.ti.com/cfs-file/\_\_key/communityserver-discussions-components-files/6/A-comprehensive-study-of-class-D-amplifier-technology.pdf

56. Class D Amplifier | Seven Transistor Labs, LLC, https://www.seventransistorlabs.com/ClassD1/

57. The Schroeder Frequency: Where Room Acoustics Gets Weird, https://sonavyx.com/en/insights/schroeder-frequency-explained

58. Room Acoustics \- Linkwitz Lab, https://www.linkwitzlab.com/rooms.htm

59. Efficient synthesis of large room impulse responses in the modal, https://www.researchgate.net/publication/392707031\_Efficient\_synthesis\_of\_large\_room\_impulse\_responses\_in\_the\_modal\_domain

60. A scalable algorithm for physically motivated and sparse, https://vbn.aau.dk/files/295939172/15\_122\_FINAL2.pdf

61. Wide-Area Psychoacoustic Correction for Problematic Room-Modes, https://adamjhill.com/AJHILL/wp-content/uploads/2017/05/Hill-Hawksford-JAES-Nov-2011.pdf

62. (PDF) Low-Frequency Noise Attenuation in a Closed Space Using, https://www.researchgate.net/publication/331200329\_Low-Frequency\_Noise\_Attenuation\_in\_a\_Closed\_Space\_Using\_Adaptive\_Directivity\_Control\_Sources\_of\_a\_Quadrupole\_Type

63. Fourier — User Manual \- FluffyAudio, https://fluffyaudio.com/wp-content/uploads/2026/07/Fourier-Manual-1.pdf

64. DC-Offset:DAC \- Audio Science Review (ASR) Forum, https://www.audiosciencereview.com/forum/index.php?threads/dc-offset-dac.6440/

65. The Importance of AM Audio Preprocessing \- Radio Blacksmith Knoll, https://rbsk6120.radio.am/2025/07/05/the-importance-of-am-audio-preprocessing/

66. GOLD V2: Master & Capture. Vst3, Au,Standalone. Win/Osx, https://thecyclekit.com/product/gold-master-capture-vst3-austandalone-win-osx/

67. RCA Signal Summing: Why a Y-Cable Overdrives Op-Amps and, https://help.lf-audio.com/academy/rca-signal-summing-why-a-y-cable-overdrives-op-amps-and-how-to-sum-channels-correctly

68. Audio 101.7 \- The High-Pass Filter Controversy, https://korneffaudio.com/the-high-pass-filter-controversy/

69. Why does adding a lowcut and highcut make my 'mastered' track clip?, https://www.reddit.com/r/audioengineering/comments/un4due/why\_does\_adding\_a\_lowcut\_and\_highcut\_make\_my/

70. Linear Phase EQ 101: Enhance Your Mixes & Avoid Phase Issues, https://unison.audio/linear-phase-eq/

71. Increase (electrical) Damping of a Driver | Page 3, https://www.audiosciencereview.com/forum/index.php?threads/increase-electrical-damping-of-a-driver.46189/page-3

72. Understanding the state of the DSP market | Page 14, https://www.audiosciencereview.com/forum/index.php?threads/understanding-the-state-of-the-dsp-market.62323/page-14

73. The Ultimate Guide to Using an Audio High Pass Filter, https://isolate.audio/articles/audio-high-pass-filter

74. Should You Use Highpass Filters when Mixing? \- PreSonus, https://www.presonus.com/blogs/home/should-you-use-high-pass-filters-when-mixing

75. Allpass Filter: All You Need To Know \- WolfSound, https://thewolfsound.com/allpass-filter/

76. DSP function to directly change phase or group delay \- KVR Audio, https://www.kvraudio.com/forum/viewtopic.php?t=132001

77. Butterworth, Linkwitz-Riley, & Bessel Filters Explained, https://leaprofessional.com/blog/butterworth-linkwitz-riley-bessel/

78. Op Amps: Linkwitz-Riley Active Crossover \- YouTube, https://www.youtube.com/watch?v=eypCME6J8BA

79. How to Choose the Perfect Compressor Settings Every Time, https://flypaper.soundfly.com/produce/how-to-choose-compressor-settings/

80. Sidechain compression: the complete mixing guide \- ACE Studio, https://acestudio.ai/blog/sidechain-compression-the-ultimate-guide/

81. Apply audio effects in Premiere \- Adobe Help Center, https://helpx.adobe.com/premiere/desktop/add-audio-effects/apply-audio-effects/audio-effects-library.html

82. TK Audio M/S Station \- KMR Audio, https://kmraudio.com/products/tk-audio-m-s-station

83. The Mixing Engineer's Handbook 4th Edition 9780998503349, https://dokumen.pub/the-mixing-engineers-handbook-4th-edition-9780998503349-0998503347.html

84. The Mastering Engineers Handbook | PDF \- Scribd, https://www.scribd.com/document/540464123/The-Mastering-Engineers-Handbook-1

85. Audio Mastering: Separating the Science from Fiction \1, [https://dokumen.pub/audio-mastering-separating-the-science-from-fiction-1nbsped-9781032359038-9781032359021-9781003329251.html

86. Wavefolding | Drey Andersson, https://dreyandersson.com/music-production-terms/wavefolding/

87. 6 ways to use a high-pass filter when mixing \- iZotope, https://www.izotope.com/community/blog/6-ways-to-use-a-high-pass-filter-when-mixing

88. 9 | High-Pass & Low-Pass Filters Explained | Sound Training 101, https://www.youtube.com/watch?v=3h25QTynets

89. Zwicker's Loudness Model Explained | PDF | Hearing | Sound \- Scribd, https://www.scribd.com/document/84070482/Zwicker-Method

90. Loudness \- Wikipedia, https://en.wikipedia.org/wiki/Loudness

91. Convention Paper 10483 \- Josh Reiss, https://joshreiss.github.io/documents/2021/21076.pdf

92. Loudness Metering, An Overview | Tokyo Dawn Knowledge Base, https://docs.tokyodawn.net/loudness-metering-an-overview/

93. User Manual Bus PEAK \- downloads.arturia.com, https://downloads.arturia.net/products/bus-peak/manual/bus-peak\_Manual\_1\_0\_0\_EN.pdf

94. Plugins \- M Media Audio, https://www.mmediaaudio.com/plugins/

95. XMLimiter 2 \- The World's Most Transparent and Loudest Mastering, https://www.kvraudio.com/forum/viewtopic.php?t=631571

96. Best Master Chain Plugins in 2026: 9 Pro-Sounding Tools, https://pluginreviewlab.com/best-master-chain-plugins/

97. Best Mastering Plugins: What Goes on the Master Bus | Hyperbits, https://hyperbits.com/blog/best-mastering-plugins/

98. oxideav-audio-filter \- GitHub, https://github.com/OxideAV/oxideav-audio-filter

99. Fruity Parametric EQ 2 \- Effect Plugin \- FL Studio, https://www.image-line.com/fl-studio-learning/fl-studio-online-manual/html/plugins/Fruity%20Parametric%20EQ%202.htm

100. How To : Sound Design With Feedback Loops \- YouTube, https://www.youtube.com/watch?v=-TSCYUG97h4

101. How to create intentional audio feedback loops with level volume, https://sound.stackexchange.com/questions/40741/how-to-create-intentional-audio-feedback-loops-with-level-volume

102. JDSherbert/Audio-DSP-Limiter \- GitHub, https://github.com/JDSherbert/Audio-DSP-Limiter

103. Designing a straightforward limiter : Blog \- Signalsmith Audio, https://signalsmith-audio.co.uk/writing/2022/limiter/