The conventional wisdom -- established by Toole 1, Thiele 2, and Linkwitz 3 -- holds that subwoofer enclosure topology barely matters for transient quality because room modal ringing (50-200 ms) buries enclosure group delay differences (10-40 ms). That consensus was formed when real-time FIR room correction was either nonexistent or prohibitively expensive. Affordable FIR correction -- running 16,384-tap convolution on a Raspberry Pi -- changes the calculus: it removes roughly 97% of modal ringing energy, elevating enclosure group delay from negligible to one of several comparable time-domain factors. This document asks two questions: (1) which enclosure topology produces the crispest transients, and (2) does our FIR pipeline handle all topologies? The answers: sealed and true transmission line designs win on transients (10-20 ms group delay at 50 Hz vs 25-40 ms for ported), and the pipeline handles any topology -- the practical differences reduce to subsonic protection (D-010) and time alignment for designs with internal acoustic paths.
Psytrance lives and dies by its kick drums. The sub-bass fundamental (30-60 Hz) needs to arrive as a tight punch, not a smeared thud. When we chose combined minimum-phase FIR filters for this system -- merging crossover and room correction into a single filter per output channel -- the goal was transient fidelity. That raises an obvious question: does the subwoofer enclosure itself help or hinder that goal?
The answer depends on whether the room is corrected. This document works through the physics of four enclosure topologies to answer the primary question -- which topology produces the crispest transients? -- and then examines whether our specific FIR pipeline handles all of them equally well.
Several design decisions constrain the analysis:
- Cut-only correction (D-009): All correction filters have gain at or below -0.5 dB at every frequency. Room peaks are attenuated; nulls are left uncorrected. The pipeline cannot boost output below the enclosure's natural rolloff.
- Per-sub independence (D-004): Each subwoofer has its own FIR correction filter, delay value, and gain trim, because different physical placement produces different room interaction.
- Per-venue measurement (D-008): Room correction filters and time alignment values are regenerated at each venue setup. The measurement pipeline captures the actual impulse response of whatever subwoofer is connected, including any internal acoustic path delay.
Before examining the topologies themselves, we need to establish when group delay differences are actually audible. The topology analysis that follows relies on psychoacoustic research that sets audibility thresholds -- and on a critical caveat about how that research was conducted. This section consolidates all relevant findings so that subsequent sections can reference them without repetition.
The most widely cited study on group delay audibility 4. Their measured findings:
- Group delay audibility was tested at frequencies from 500 Hz to 8 kHz.
- At 500 Hz, the audibility threshold was approximately 3.2 ms (corresponding to roughly 1.6 cycles).
- At higher frequencies, the threshold decreased (shorter group delays became audible).
- The general pattern: the threshold is approximately 1.5-2 periods of the signal frequency.
Blauert and Laws did not measure group delay audibility below 500 Hz. The sub-bass thresholds commonly cited in audio engineering literature (for example, "~30 ms at 50 Hz" or "~50 ms at 30 Hz") are extrapolations of the ~1.5-cycle rule to frequencies the study never tested. These extrapolated values appear throughout the literature as if they are measured data, but they are not.
The extrapolation assumes that the "cycles of the signal frequency" relationship holds at sub-bass frequencies. This is plausible but unverified:
- Arguments for: The psychoacoustic mechanism (temporal pattern recognition requiring multiple cycles) should be frequency-independent.
- Arguments against: Sub-bass frequencies are perceived partly through vibrotactile (body feel) mechanisms, not just auditory processing. The temporal resolution of tactile perception differs from auditory perception. Group delay audibility at 40 Hz may be governed by different perceptual mechanisms than at 500 Hz.
Extrapolation caveat: All sub-bass group delay thresholds cited in this document (for example, ~30 ms at 50 Hz) are extrapolated from the Blauert and Laws data, not directly measured. The actual audibility threshold at sub-bass frequencies is unknown. This caveat applies throughout the topology analysis and comparison sections that follow.
Lipshitz, Pocock, and Vanderkooy (1982) 5 studied audibility of linear-phase vs minimum-phase crossovers. Found that group delay differences were inaudible in most listening conditions but could be detected with specific test signals (clicks, impulses) at moderate SPL. This suggests that for complex musical signals, group delay differences below the Blauert and Laws threshold are reliably inaudible.
Linkwitz 3 (co-inventor of the Linkwitz-Riley crossover) has stated in his published writings that LR4 crossover phase distortion is inaudible in his personal listening tests. This is notable because Linkwitz is arguably the person most motivated and qualified to detect phase artifacts from his own crossover design. The specific publication should be identified for proper citation before relying on this claim.
Thiele 2 (cited by Small) analyzed the transient behavior of correctly designed ported and sealed enclosures and concluded that the differences between them are "likely to be inaudible." This is consistent with the document's broader conclusion that room effects dominate the sub-bass time response. The specific Thiele paper should be identified and verified before publication.
Toole (2008), "Sound Reproduction" 1 discusses group delay audibility in the context of subwoofer systems. The general conclusion: group delay from loudspeaker/room systems is almost always inaudible because room reflections dominate the time-domain behavior. The exception is direct-radiator subwoofers at very close range in acoustically treated rooms.
A psytrance kick typically has a broadband attack transient (2-10 kHz, under 5 ms) above the crossover frequency, and a sub-bass fundamental and tail (30-60 Hz, extending 100-300 ms) below it. The attack passes through the mains with minimal group delay regardless of enclosure type. The sub-bass content passes through the subwoofer, where enclosure group delay applies.
At 50 Hz, the extrapolated Blauert and Laws threshold is approximately 30 ms. A sealed sub with Qtc = 0.707 at Fc = 40 Hz has roughly 15-20 ms of group delay at 50 Hz -- below the extrapolated threshold. A ported sub tuned to Fb = 35 Hz (B4 alignment) has roughly 25-35 ms -- at or above it. A horn-loaded sub falls in the 20-40 ms range depending on design. A true TL is comparable to sealed (~10-20 ms), while a quarter-wave resonator is comparable to ported (~25-40 ms).
Does the difference matter in a real venue? The answer depends on whether the system applies room correction. The full analysis of corrected vs uncorrected systems is in Section 5.4.
This section addresses the primary question -- which enclosure topology produces the crispest transients? -- by examining each of the four topologies in turn. Each is covered with the same subsections -- operating principle, group delay, cone excursion, transient behavior, and PA suitability -- to enable direct comparison.
The driver operates against a closed air volume that acts as a spring. The system is a second-order highpass filter (12 dB/octave rolloff below resonance). The trapped air adds to the driver's own suspension compliance, raising the system resonance frequency (Fc) above the driver's free-air resonance (Fs). Sealed boxes trade low-frequency extension for predictable, well-damped transient behavior.
Group delay follows a second-order model. It peaks at the system resonance frequency Fc and is determined by the system Q factor (Qtc):
| Alignment | Qtc | Group delay at Fc | At Fc = 50 Hz |
|---|---|---|---|
| Bessel (maximally flat delay) | 0.577 | ~0.6 / Fc | ~12 ms |
| Butterworth (maximally flat magnitude) | 0.707 | ~0.9 / Fc | ~18 ms |
| Underdamped | 1.0 | ~1.3 / Fc | ~26 ms |
These relationships derive directly from the second-order highpass transfer function.
Cone excursion increases monotonically as frequency decreases below Fc. The restoring force (air spring plus suspension) keeps the driver under control at all frequencies. Excursion is limited only by the driver's mechanical limits (Xmax). There is no unloading transition -- the air spring provides a restoring force at all frequencies. Subsonic protection (D-010) is not needed.
A sealed box's cone motion is always well-damped by the air spring. Musical content with significant energy below Fc (sub-bass drops, 808 kicks reaching 30 Hz) produces less excursion than in a ported box because the air spring provides restoring force, keeping the driver in its linear range longer. For transient linearity, sealed is the reference standard among the four topologies.
Sealed subwoofers are rare in professional PA. Their lower efficiency (no resonant output boost) and earlier rolloff (12 dB/octave vs 24 dB/octave) make them less competitive where maximum SPL per watt and per cubic meter are priorities. They are, however, the simplest to design and the most predictable in behavior.
The group delay table in 3.1.2 lists three canonical sealed alignments. Which is optimal for a system that applies cut-only (D-009) minimum-phase FIR correction targeting psytrance transient fidelity? The answer is not obvious, because the cut-only constraint interacts with each alignment differently.
| Alignment | Qtc | -3 dB re Fc | Group delay at Fc | Impulse overshoot |
|---|---|---|---|---|
| Bessel | 0.577 | ~1.55 x Fc | ~0.6 / Fc (flattest) | 0% |
| Butterworth | 0.707 | 1.00 x Fc | ~0.9 / Fc | ~4.3% |
| Underdamped | 1.0 | ~0.79 x Fc | ~1.3 / Fc | Visible ringing |
Bessel has the flattest group delay and cleanest impulse response but rolls off earliest. Butterworth extends 1.55x lower in frequency for the same Fc. Underdamped extends further still but with ringing that degrades transients. These are textbook second-order highpass properties 6 7.
In a system that allows both boost and cut, the standard argument is: choose Bessel for its superior transient behavior, then boost the low-frequency rolloff to recover bass extension. The FIR filter adds only the minimum additional group delay associated with that magnitude correction.
Our system cannot do this. D-009 mandates cut-only correction because psytrance source material is mastered to within a hair's breadth of digital maximum -- any boost risks clipping on a PA system. The consequence for alignment choice:
- Bessel at Fc = 40 Hz: -3 dB at ~62 Hz, -6 dB at 40 Hz, -12 dB at 30 Hz.
- Butterworth at Fc = 40 Hz: -3 dB at 40 Hz, -6 dB at ~30 Hz.
The Bessel box is 6 dB quieter at 40 Hz -- in the heart of the psytrance kick fundamental range -- with no recovery path. This is a permanent, irrecoverable loss of low-frequency output under D-009.
The Butterworth group delay peak at Fc is the minimum-phase delay associated with the magnitude response knee. To reduce it, the FIR correction would need to make the transition from passband to rolloff more gradual -- effectively cutting frequencies near Fc to lower the effective Q. This works, but trades bandwidth for group delay:
- Butterworth at Fc = 40 Hz: group delay ~22.5 ms at Fc.
- Apply ~2 dB shelf cut at Fc (reducing effective Qtc to ~0.63): group delay drops to ~18.8 ms (~17% reduction).
- Cost: -2 dB at 40 Hz (irrecoverable under D-009), ~-1 dB at 50 Hz.
The fundamental constraint: within a minimum-phase system, you cannot reduce group delay without either extending the passband (boost, prohibited) or narrowing it (cut, which sacrifices extension). The correction effectively slides the system from Butterworth toward Bessel in post-processing 8.
Instead of Bessel at the same Fc, use a larger box to push Fc low enough that the -3 dB point matches Butterworth. For a typical PA driver (Fs = 30 Hz, Qts = 0.4, Vas = 120 L):
- Butterworth: ~53 L box, Fc ~47 Hz, -3 dB at 47 Hz.
- Bessel: ~102 L box, Fc ~38 Hz, -3 dB at ~59 Hz.
The Bessel box is nearly 2x the volume but its -3 dB point is still 12 Hz higher. For drivers with Qts < 0.577 (most PA drivers), matching the Butterworth -3 dB point with a Bessel alignment requires impractically large boxes.
The Bessel and Butterworth points sit on a continuous Qtc spectrum:
| Qtc | Group delay (re Bessel) | -3 dB re Fc | Overshoot |
|---|---|---|---|
| 0.577 | 1.00x (reference) | ~1.55 x Fc | 0% |
| 0.63 | ~1.15x | ~1.35 x Fc | ~1.5% |
| 0.65 | ~1.25x | ~1.25 x Fc | ~2.5% |
| 0.707 | 1.50x | 1.00 x Fc | 4.3% |
At Fc = 40 Hz, the Bessel-to-Butterworth spread is 15 ms to 22.5 ms -- a 7.5 ms difference. A Qtc of 0.65 gives ~18.8 ms (75% of Bessel's advantage) while keeping the -3 dB point within ~25% of Butterworth.
Butterworth (Qtc = 0.707) is the better default for our system.
-
Extension matters more than group delay margin. Psytrance kick fundamentals at 30-50 Hz. Bessel sacrifices 6+ dB in this range with no recovery path. The group delay improvement (7.5 ms) is in a perceptual regime where audibility is unverified (below Blauert's measured range).
-
The group delay penalty is bounded. Butterworth at Fc = 40 Hz: ~22.5 ms -- below the extrapolated Blauert threshold (~25 ms at 40 Hz) and comparable to early reflections (5-30 ms) that remain after room correction.
-
FIR provides a tunable compromise. If transient testing reveals Butterworth group delay is audibly problematic, the FIR correction can gently reshape the knee to reduce group delay at the cost of some extension -- sliding the system toward Bessel in post-processing.
-
Underdamped (Qtc > 0.707) is not preferred when box volume is unconstrained. The ringing produces audible time-domain artifacts; the FIR correction can reduce them only by cutting the magnitude peak, which yields less usable bandwidth than Butterworth with no transient advantage. Exception: volume-constrained builds -- see "Compact builds" below.
-
For custom builds, Qtc = 0.65 in a somewhat larger box is worth considering -- 75% of Bessel's group delay advantage with a -3 dB point within ~25% of Butterworth. But this is a refinement, not a requirement.
When box volume is constrained -- portable rigs, compact flight cases, or builds using drivers with high Qts in small enclosures -- the resulting Qtc lands above 0.707 (underdamped). The magnitude response has a hump near Fc and the group delay peak is higher than Butterworth. The standard advice (point 4 above) is to avoid this alignment when possible.
However, the FIR correction can flatten the magnitude hump by cutting the peak region. This is D-009 compliant (cut-only). Because the system is minimum-phase, reducing the magnitude peak also reduces the associated group delay peak -- the corrected system approximates a Butterworth frequency response and approaches Butterworth group delay.
The limitation is physical, not electrical. FIR correction reshapes the signal but cannot change the air spring stiffness of the smaller box or reduce cone excursion demands. At PA levels near Xmax:
- A Qtc = 0.9 box in ~30 L requires roughly 3x more cone excursion at 30 Hz than the same driver in a true Butterworth box (~53 L) for the same acoustic output.
- The smaller box loses 3-5 dB of native extension below 30 Hz that D-009 cannot recover (no boost allowed).
- Thermal and mechanical stress on the driver increase at equivalent SPL.
Recommendation: When box volume is unconstrained, the true Butterworth alignment is strictly preferred -- more extension, more SPL headroom, lower excursion stress, no FIR dependency for flat response. The underdamped + FIR strategy is a valid engineering tradeoff for compact sealed builds only, and the D-010 speaker profile should note both the native Qtc and the target Qtc after correction so the pipeline can verify the correction is within safe limits.
For sealed enclosures, the D-010 speaker profile should record approximate Qtc. The pipeline can compute optimal correction from the measured impulse response alone, but recording Qtc helps the operator interpret results and understand why the measured response rolls off where it does. It also informs whether FIR knee reshaping (trading extension for group delay) is available or already exhausted.
The enclosure has a tuned port -- a tube or slot that resonates at a specific frequency (Fb). At Fb, the port's air mass acts as a second radiator, in phase with the cone above Fb and out of phase below Fb. The system is a fourth-order highpass filter (24 dB/octave rolloff below Fb). This gives more output near Fb but introduces more complex phase behavior and reduced cone damping below the port tuning frequency.
Group delay is more complex due to the fourth-order nature. It depends on both the box tuning frequency (Fb) and the alignment type:
- QB3 (quasi-Butterworth 3rd order): Moderately damped, smooth rolloff. Group delay peaks below Fb, typically 1.5-2x the sealed equivalent at the same -3 dB frequency.
- B4 (4th-order Butterworth): Maximally flat magnitude. Group delay peaks at approximately 1.4 / Fb. At Fb = 35 Hz: ~40 ms -- significantly more than a sealed box tuned to the same -3 dB point.
- C4 (4th-order Chebyshev): Maximally flat group delay for a fourth-order system. Lower group delay peak than B4 but with magnitude ripple in the passband.
The general relationships are well-established in electroacoustic theory 6 7. Exact values depend on the specific alignment parameters and are best computed from the transfer function for each case.
Cone excursion is at its minimum at Fb (the port does the work). Below Fb, excursion increases rapidly because the port no longer provides loading -- the driver is effectively operating in free air. This is the "unloading" phenomenon that can cause mechanical damage, and the reason D-010 requires subsonic protection for ported enclosures. D-031 extends this requirement to ALL enclosure types when small drivers are involved -- sealed enclosures with limited-excursion drivers (e.g., 5.25" isobaric) also need subsonic HPF protection to prevent over-excursion from subsonic content.
At the port tuning frequency Fb, the air in the port resonates and radiates in phase with the cone. Below Fb, the port radiation inverts phase relative to the cone:
- Above Fb: Cone and port radiate in phase, summing constructively. This is where the ported box gains its efficiency advantage.
- At Fb: The port does most of the radiating; the cone is nearly stationary (minimum excursion).
- Below Fb: Cone and port cancel each other. Output drops at 24 dB/octave (vs 12 dB/octave for sealed).
The phase inversion has a direct effect on the impulse response: a ported box shows a longer tail than a sealed box at the same -3 dB frequency, because the port continues to radiate energy after the cone has stopped. This tail is the time-domain manifestation of the higher group delay.
A ported box's cone is well-damped near and above Fb but poorly damped below Fb. Musical content with significant energy below Fb (sub-bass drops, 808 kicks reaching 30 Hz) can push the ported driver into its nonlinear excursion range, producing distortion on transients. The same content through a sealed box produces less excursion because the air spring provides restoring force, keeping the driver in its linear range longer.
Ported subwoofers are the standard in professional PA. Their higher efficiency near Fb, extended low-frequency output, and straightforward construction make them the default choice. Subsonic protection (D-010) is mandatory.
A horn-loaded subwoofer uses a flared acoustic path between the driver and the listening environment. The horn transforms the driver's high-impedance, low-velocity acoustic output into a low-impedance, high-velocity output -- analogous to an electrical transformer. This provides:
- Higher efficiency: The horn couples the driver more effectively to the room air. A horn-loaded sub can produce 6-10 dB more output than a direct-radiating (sealed or ported) design with the same driver and input power.
- Controlled directivity at low frequencies: Long horns (longer than 1/4 wavelength at the lowest operating frequency) begin to exhibit directional behavior, which can be useful for steering bass away from stage areas.
Horn acoustics are well-established 9 10 11.
In practice, most horn-loaded subwoofers use folded horn paths to reduce physical dimensions. The horn is folded one or more times within the enclosure, maintaining the required path length in a more compact form factor. Folded horns have path lengths of 2-4 meters depending on tuning frequency.
The horn path introduces its own group delay. Sound must travel the physical length of the horn before exiting. For a horn tuned to 40 Hz with a quarter-wave path:
- Quarter wavelength at 40 Hz: 343 / (4 x 40) = 2.14 meters
- Sound travel time through 2.14 m of horn path: 2.14 / 343 = 6.24 ms
- Folded horns have path lengths of 2-4 meters depending on tuning frequency
This is a minimum group delay -- the actual delay is higher because the horn also has resonant behavior that adds time-domain ringing.
The horn's group delay is harder to predict analytically than sealed or ported because it depends on the specific horn flare profile (exponential, conical, hyperbolic), the path length, and the throat and mouth dimensions. Approximate magnitude at 50 Hz: ~20-40 ms (design-dependent).
Horn loading reduces cone excursion at frequencies where the horn is effective (near and above the horn's cutoff frequency). This is the primary advantage for high-SPL applications: the driver operates in a smaller excursion range, staying in its linear region and producing less distortion.
Below the horn's cutoff frequency, the horn ceases to provide loading and the driver behaves similarly to a direct radiator. Excursion increases, and the same unloading concern as ported designs applies -- albeit typically at a lower frequency because the horn's resonant reinforcement extends the usable range. Subsonic protection (D-010) is advisable.
For psytrance at high SPL, horn-loaded subs have a genuine advantage: they can produce higher output with less distortion in the 30-80 Hz range that carries the kick drum's energy. The tradeoff is size (folded horns are physically large) and group delay (the horn path adds delay).
Horn-loaded subs have an additional time alignment consideration: the acoustic center of a horn sub is not at the mouth of the horn -- it is somewhere inside the horn path, at a point that varies with frequency. This means:
- Physical measurement of driver-to-listener distance does not accurately predict the acoustic delay.
- The measurement pipeline's impulse response detection (arrival time from the onset of energy) correctly captures the effective acoustic delay regardless of where the acoustic center is.
- Time alignment values for horn subs may be significantly different from what physical distance suggests.
This is another reason why per-venue measurement (D-008) is essential rather than relying on calculated delays from physical measurements. The frequency-dependent acoustic center of horns is well-documented in horn loudspeaker design literature.
Horn-loaded subwoofers are common in professional PA, particularly for high-SPL applications. Their higher efficiency makes them the preferred choice when maximum output is required. The tradeoff is physical size -- folded horns are large and heavy. For psytrance at high SPL, horn-loaded subs offer a genuine advantage in the 30-80 Hz kick drum range due to reduced cone excursion and lower distortion at high output levels.
A transmission line (TL) enclosure is a long, typically tapered acoustic pipe with the driver mounted at one end (the closed end) and the other end open to the room. The pipe is stuffed with damping material (acoustic wadding, long-fiber wool, polyester fill) along part or all of its length.
- The driver radiates into the pipe. The sound wave travels down the pipe toward the open end.
- The damping material progressively absorbs mid and high frequencies as the wave travels through the pipe. By the time it reaches the open end, only low frequencies remain.
- At the open end, the low-frequency wave radiates into the room, supplementing the driver's direct front radiation.
- The pipe length is chosen so that the path length equals one quarter of the wavelength at the target reinforcement frequency. At this frequency, the wave arriving at the open end is in phase with the driver's direct radiation (the quarter-wave path introduces a 90-degree phase shift, and the pressure-to-velocity inversion at the open end adds another 90 degrees, totaling 180 degrees -- but the rear radiation from the driver is already 180 degrees out of phase, so the net result is in-phase reinforcement).
The pipe may be straight, folded, or tapered. Tapering changes the impedance transformation and affects the frequency response shape.
Transmission line theory is well-documented 12 13 14.
This distinction is the source of widespread confusion in loudspeaker design.
A true transmission line uses heavy damping throughout the pipe to absorb reflections and resonances. The ideal TL has no internal standing waves -- the pipe acts as an acoustic termination that absorbs the driver's rear radiation. The driver "sees" an infinitely long pipe and behaves as if mounted in an infinite baffle. The system provides resistive rather than reactive loading, resulting in approximately second-order behavior similar to sealed but with lower system Q. The low-frequency output from the open end comes from the progressive wave that survives the damping, not from resonance.
A quarter-wave resonator (sometimes called a mass-loaded transmission line, or MLTL) uses minimal damping and relies on the pipe's standing-wave resonance to boost output at the tuning frequency. This is functionally similar to a ported enclosure, with the pipe's air mass playing the role of the port's air mass.
In practice, most "transmission line" designs are hybrids -- they use enough damping to suppress upper harmonics of the pipe resonance but not enough to eliminate the fundamental quarter-wave resonance. The degree of damping determines where the design sits on the spectrum between true TL and quarter-wave resonator.
| Design type | Damping | Resonance behavior | Closest analogy |
|---|---|---|---|
| True TL (heavily damped) | Heavy throughout | Minimal resonance, smooth rolloff | Sealed box with lower Fc |
| Hybrid TL (moderate damping) | Moderate, often concentrated near driver | Attenuated fundamental, suppressed harmonics | Between sealed and ported |
| Quarter-wave resonator / MLTL | Light or localized | Strong fundamental resonance | Ported box with pipe instead of port |
The TL vs quarter-wave resonator distinction is extensively discussed in the loudspeaker design literature. Martin King's work (2001-2010, published on quarter-wave.com and in peer-reviewed journals) provides detailed analysis with simulation and measurement data 15.
Quarter-wave resonators exhibit analogous behavior to ported enclosures: the open end of the pipe acts as the radiating element (like a port), with similar phase inversion below the tuning frequency.
The HOQS Paraflex designs are community-developed folded quarter-wave subwoofers shared through the diyAudio community. Multiple variants exist (single-fold, double-fold, different tuning frequencies), designed for high efficiency and high SPL in pro audio and DJ applications. The "Paraflex" name refers to the specific folding geometry. Given their light damping and reliance on quarter-wave resonance, they sit at the quarter-wave resonator end of the spectrum described above.
No peer-reviewed measurements of Paraflex transient behavior or group delay have been published 16. The diyAudio community has noted the lack of published measurement data from HOQS. Individual community members have shared measurements, but these are not standardized or independently verified.
True TL (heavily damped): Group delay is dominated by two mechanisms:
- Propagation delay through the pipe: Sound travels the physical length of the pipe at ~343 m/s. For a TL tuned to 40 Hz (quarter wavelength = 2.14 m), propagation delay is ~6.2 ms.
- System resonance: In a heavily damped TL, the quarter-wave resonance is suppressed. The system behaves closer to a sealed box with an effective Qtc determined by the driver parameters and the damping. Group delay from resonance is lower than ported or lightly-damped designs.
The net group delay of a true TL at 50 Hz is estimated at 10-20 ms -- lower than a ported B4 alignment (~25-35 ms) and comparable to a sealed box (~15-20 ms). The heavy damping trades away the resonant output boost in exchange for time-domain behavior closer to sealed.
Intermediate designs (moderate damping): Group delay falls between the true TL and quarter-wave resonator bounds, typically 15-25 ms at 50 Hz, depending on the damping density and taper profile.
Lightly damped TL / quarter-wave resonator: With light damping, the quarter-wave resonance is strong, and the group delay behavior approaches that of a ported enclosure. Group delay peaks near the tuning frequency, with magnitude comparable to a ported B4 alignment (~25-40 ms at 50 Hz depending on tuning).
| TL damping level | Approximate behavior | Group delay at 50 Hz |
|---|---|---|
| Heavy (true TL) | ~Second order (sealed-like) | ~10-20 ms |
| Moderate (practical TL) | Between 2nd and 4th order | ~15-25 ms |
| Light (quarter-wave resonator) | ~Fourth order (ported-like) | ~25-40 ms |
The propagation delay component is straightforward physics. The resonance-related group delay depends heavily on the specific damping profile and is harder to predict without simulation or measurement.
True TL: The resistive pipe loading provides damping at all frequencies, including below the quarter-wave frequency. Cone excursion increases below the quarter-wave frequency but more gradually than in a ported box because the pipe still provides some resistive loading -- there is no sharp unloading transition. Subsonic protection is still advisable (D-010) to prevent mechanical limits being reached on sub-bass content, but the failure mode is less abrupt than with a port and the protection can be less aggressive (for example, 12 dB/octave vs 24 dB/octave).
Quarter-wave resonator / MLTL: With light damping, cone excursion behavior is closer to ported -- unloading occurs below the tuning frequency. Subsonic protection comparable to a ported enclosure is appropriate.
A true TL falls between sealed and ported for cone control -- better than ported, though not as predictable as sealed.
TL advocates frequently claim that transmission lines produce "tighter" or "faster" bass than ported designs.
There is a physical basis for this claim, but it applies only to true (heavily damped) TLs, not to lightly-damped quarter-wave resonators.
The physical basis:
- Damping absorbs reflections. In a true TL, the wadding absorbs the sound wave as it travels down the pipe. There are no strong standing waves bouncing back and forth, so the impulse response decays faster than in a ported box (where the port resonance stores and re-radiates energy). This produces a shorter, cleaner impulse response tail.
- No cone unloading below tuning. Unlike a ported enclosure where cone excursion increases dramatically below Fb, a heavily damped TL provides relatively uniform loading across frequency. The damping material acts as an acoustic resistance that loads the driver at all frequencies, keeping it better-controlled on transients with sub-bass content.
- No port noise. Ported enclosures can produce audible turbulence noise ("chuffing") from the port at high excursions. TLs have no port -- the open end has a much larger cross-section, reducing air velocity and eliminating chuffing.
However:
- The "tighter bass" perception is partly attributable to the TL's reduced output near the tuning frequency compared to a ported box. Less bass boost creates a perception of "tighter" bass, which is not the same as faster transient response.
- Lightly-damped TLs (which many commercial "TL" designs actually are) have comparable group delay to ported designs, because they rely on the same quarter-wave resonance mechanism.
- The Blauert and Laws threshold caveat applies here too (see Section 2): at sub-bass frequencies, group delay differences of 5-15 ms between enclosure types are likely below the audibility threshold (with the caveat that sub-bass thresholds are extrapolated, not measured).
- Toole's observation 1 that room effects dominate applies equally: any TL transient advantage is swamped by the room's modal behavior in practice.
The "tighter bass" reputation likely originates from comparisons between well-designed TLs (moderate to heavy damping) and poorly aligned ported boxes (underdamped, high-Qtc driver producing a response hump). A properly aligned ported box (QB3 or B4 with an appropriate driver) has transient behavior that is difficult to distinguish from a moderately damped TL in a real listening environment, per Toole's broader conclusion about room dominance 1.
Assessment: The "tighter bass" claim for true TLs has a valid physical basis in reduced impulse response ringing and better cone control below tuning. For lightly-damped TLs and quarter-wave resonators, the claim is not supported -- they behave more like ported enclosures. In real rooms at PA levels, the difference is unlikely to be audible.
Transmission lines are primarily a hi-fi topology. They are rare in professional PA for several reasons:
- Size. A TL tuned to 40 Hz needs a pipe at least 2.14 m long. Even folded, the enclosure is large relative to its output. PA applications prioritize output-per-cubic-meter, where ported and horn-loaded designs are more efficient.
- Efficiency. A true TL absorbs significant energy in its damping material. This energy is converted to heat rather than sound. At PA levels where every watt counts, this is a meaningful disadvantage.
- Cost and complexity. TL enclosures are more complex to build than ported boxes (precise pipe dimensions, careful damping placement) and less predictable to model (damping behavior is difficult to simulate accurately).
- Power handling. The damping material in a TL can compress at very high SPL, changing the enclosure's acoustic behavior under exactly the conditions where predictable performance matters most.
Some high-end PA manufacturers (notably PMC -- Professional Monitor Company 17) use transmission line designs in their studio monitors and some larger systems. Their designs are closer to the lightly-damped (quarter-wave resonator) end of the spectrum, optimized for monitoring accuracy rather than maximum SPL. However, these are the exception -- the vast majority of PA subwoofers are ported or horn-loaded.
For our system, TL subwoofers are unlikely to be used at psytrance events (where high SPL, portability, and efficiency are priorities). They might be encountered if the owner uses hi-fi subwoofers for small venue or rehearsal setups. The pipeline handles them transparently -- the speaker profile just needs a "transmission line" option.
The preceding sections examined each topology independently. This section consolidates the evidence to answer the primary question directly: which topology gives the crispest transients, and by how much?
| Property | Sealed | Ported | Horn-loaded | True TL | QW Resonator / MLTL |
|---|---|---|---|---|---|
| Rolloff slope | 12 dB/oct | 24 dB/oct | Design-dependent | ~12 dB/oct | ~24 dB/oct |
| Filter order | 2nd | 4th | Complex | ~2nd (damped) | ~4th |
| Group delay at 50 Hz | ~15-20 ms | ~25-35 ms | ~20-40 ms | ~10-20 ms | ~25-40 ms |
| Cone unloading below tuning | No | Yes | Moderate | Minimal | Moderate |
| Subsonic protection (D-010) | Not needed | Mandatory | Advisable | Advisable (less aggressive) | Advisable |
| Efficiency | Low | Medium | High | Low | Medium |
| Size for equivalent output | Medium | Medium | Large | Large | Medium-large |
| "Tighter bass" claim | Reference | No | No | Supported (physical basis) | No |
| PA suitability | Rare | Standard | Common (high SPL) | Rare (low efficiency) | Uncommon |
| FIR correction compatibility | Full | Full above Fb | Full | Full | Full above tuning |
| Pipeline handles it | Yes | Yes | Yes | Yes | Yes |
| Significance in uncorrected rooms | Minimal (room modes dominate) | Minimal (room modes dominate) | Minimal (room modes dominate) | Minimal (room modes dominate) | Minimal (room modes dominate) |
| Significance in corrected system | Lowest group delay; modest advantage | Higher group delay; modest disadvantage | Variable | Comparable to sealed; modest advantage | Comparable to ported; modest disadvantage |
| Topology | Group delay mechanism | Approximate magnitude at 50 Hz |
|---|---|---|
| Sealed (Qtc = 0.707) | System resonance (2nd order) | ~15-20 ms |
| Ported (B4, Fb = 35 Hz) | Port resonance (4th order) | ~25-35 ms |
| Horn-loaded (Fb = 40 Hz) | Path length + resonance | ~20-40 ms (design-dependent) |
| True TL (heavily damped) | Propagation + attenuated resonance | ~10-20 ms |
| Quarter-wave resonator / MLTL | Propagation + strong resonance | ~25-40 ms |
The Thiele-Small parameters are the foundational relationships of loudspeaker/enclosure design 6 7.
Qts is the most important parameter for determining whether a driver suits a sealed or ported enclosure:
| Qts range | Recommendation |
|---|---|
| < 0.4 | Well-suited to ported. Strong electromagnetic damping compensates for the port's reduced damping below Fb. |
| 0.4 - 0.7 | Either topology. Sealed gives Qtc ~ 0.7 in a moderate box; ported extends bass response. |
| > 0.7 | Best in sealed. High-Q drivers in ported boxes tend to produce boomy, underdamped bass with excessive group delay. |
In a sealed box, the system Q (Qtc) relates to the driver Qts by:
Qtc = Qts * sqrt(Vas / Vb + 1)
where Vb is the box volume and Vas is the driver's equivalent compliance volume. Smaller boxes yield higher Qtc (less damped, more ringing); larger boxes let Qtc approach Qts. For optimal transient response (Bessel alignment, Qtc = 0.577), the box volume must be chosen to satisfy the equation -- this gives the flattest group delay at the cost of less bass extension compared to a Butterworth alignment.
Vas (equivalent compliance volume): Determines box size. Ported boxes are typically 1.5-3x the volume of sealed boxes for the same driver.
Fs (free-air resonance): Sets the lower frequency limit. In a sealed box Fc > Fs (the air spring raises it). In a ported box, the -3 dB point can be below Fs if the port is tuned low enough.
This section addresses the secondary question: does our pipeline handle all four topologies? The short answer is yes. The longer answer explains what the pipeline can and cannot do for each, and why room correction changes the decades-old consensus about enclosure topology's importance.
Above the enclosure's tuning or rolloff frequency, all topologies have essentially flat response that the room correction pipeline addresses effectively. The enclosure type does not meaningfully affect the pipeline's ability to tame room modes in this region.
A minimum-phase FIR correction filter can also reduce group delay near the enclosure's rolloff frequency by flattening the magnitude response. Since group delay is linked to magnitude response slope via the minimum-phase relationship, smoothing out a response peak near the tuning frequency also reduces the group delay peak. This is particularly relevant for ported and quarter-wave resonator designs, where the magnitude peak near the tuning frequency contributes significantly to group delay. Flattening this peak through correction can bring the group delay in the tuning region closer to -- though not equal to -- sealed-box levels.
Room modal peaks are the pipeline's primary target in the sub-bass range. A room mode produces both a magnitude peak and associated time-domain ringing (the two are linked through the minimum-phase relationship). When the correction filter attenuates a 15 dB room mode peak to flat, the ringing energy decreases proportionally -- roughly 97% energy reduction (magnitude squared). This is the pipeline's core value proposition for sub-bass quality.
However, the correction does not eliminate room effects entirely:
- Modal ringing is reduced, not zeroed. The correction targets the magnitude peak, which reduces ringing, but residual energy remains. A 15 dB mode cut leaves ~3% of the original ringing energy -- small but nonzero. Multiple overlapping modes leave a residual floor of time-domain energy.
- Early reflections are deliberately uncorrected. Floor, wall, and ceiling reflections arriving within 5-30 ms of the direct sound produce comb filtering that varies with listener position. The pipeline's frequency-dependent windowing ignores these -- correcting them at the measurement position would worsen the response everywhere else.
- Diffuse reverberation is untouched. The late-arriving, broadband decay of the room is not targeted by the correction.
The consequence: after room correction, the sub-bass time domain still contains early reflections and residual modal energy. But the dominant modal artifacts (50-200 ms of ringing) are dramatically reduced, which elevates the relative importance of the enclosure's intrinsic group delay.
The FIR correction cannot reduce group delay below the minimum-phase limit for the corrected magnitude response. It cannot eliminate the excess phase from the port's phase inversion below Fb -- that is a physical property of the enclosure. Similarly, it cannot remove the propagation delay through a horn's or transmission line's acoustic path.
At frequencies well below the tuning frequency, the enclosure's output is too low to correct anyway: D-009 prohibits boost, so the correction rolls off naturally. The group delay in that region is irrelevant because no signal passes through it.
Net effect: FIR correction can bring group delay near and above the tuning frequency closer to sealed-box levels, but cannot eliminate the enclosure's intrinsic minimum-phase delay. In a room-corrected system, this residual enclosure delay is one of several comparable time-domain artifacts rather than a negligible one.
In an uncorrected system: Enclosure topology differences probably do not matter. Toole's work 1 shows that room reflections and modal behavior dominate the sub-bass time response in untreated rooms. Modal ringing at 50-200 ms swamps the 15-40 ms range of enclosure group delay variation. The difference between a sealed and a ported sub is buried under room artifacts. This is the context in which Toole, Thiele, and Linkwitz made their assessments -- none of them assumed a room-corrected system, because real-time FIR correction was either nonexistent (Blauert 1978 4, Lipshitz 1982 5) or exotic and expensive (Toole 2008 1) at the time of their work.
In a room-corrected system (our system): The balance shifts. Our pipeline applies 16,384-tap minimum-phase FIR correction that specifically targets room modes. When the correction attenuates a 15 dB room mode peak to flat, the associated ringing energy decreases by roughly 97% (power scales with magnitude squared). This is highly effective -- but the correction does not eliminate room effects entirely. Early reflections from nearby surfaces (5-30 ms) are deliberately uncorrected because they vary with listener position. Diffuse reverberation is untouched. And the enclosure's intrinsic group delay -- a minimum-phase property of the enclosure itself -- cannot be reduced below the minimum-phase limit by magnitude correction.
After room correction, the dominant time-domain artifacts in the sub-bass range are, in approximate order: (1) early reflections from nearby boundaries (5-30 ms, uncorrected by design), (2) enclosure group delay (10-40 ms depending on topology, partially reducible but not below minimum-phase limit), (3) residual modal ringing (reduced but not zero), (4) diffuse reverberation (broadband, not targeted by correction). Enclosure group delay moves from a distant also-ran (in uncorrected rooms) to a peer of early reflections as a time-domain factor.
This does not mean enclosure topology becomes the dominant effect. Early reflections from the floor, nearest wall, and ceiling are still present and produce comb filtering that overlaps the enclosure group delay range. But enclosure topology moves from "negligible compared to room effects" to "one of several comparable contributors to time-domain behavior." The 10-20 ms difference between a sealed sub and a ported sub is no longer buried under 100+ ms of modal ringing -- it is now in the same order of magnitude as the early reflection pattern.
With two independent subs, each sub can use a different enclosure type. The per-sub FIR correction handles the differences naturally:
- Both ported with the same Fb: identical subsonic rolloff in both filters, delays differ only due to room placement.
- One sealed, one ported: the FIR filters have different shapes, but the pipeline handles this -- each sub is measured and corrected independently.
- Both ported with different Fb: each filter includes the appropriate subsonic protection rolloff for its Fb value (D-010 speaker profile parameter).
- One horn-loaded or TL, one direct-radiating: the measurement pipeline captures the acoustic path delay and the correction adjusts accordingly.
This is an advantage of the combined FIR approach. An IIR crossover would need different crossover parameters for each sub, which adds configuration complexity that the combined FIR approach absorbs automatically.
For our system's purposes, the measurement pipeline (D-008) will measure the actual impulse response of whatever subwoofer is connected, including any internal acoustic path delay. The per-sub FIR correction (D-004) compensates for the measured response. The pipeline does not need to know the enclosure topology -- it measures the result and corrects accordingly. The speaker profile (D-010) should include enclosure type options (sealed, ported, horn, transmission line) to set the appropriate subsonic protection behavior. For transmission lines, subsonic protection should default to enabled (conservative) with the option to disable for verified heavily-damped true TLs. The correction approach itself is the same regardless of topology: measure, compute inverse, apply as minimum-phase FIR.
Sealed and true transmission line designs, with group delay of 10-20 ms at 50 Hz. Ported and quarter-wave resonator designs are measurably worse at 25-40 ms.
Whether that difference matters depends on room correction. In an uncorrected room, it does not -- modal ringing at 50-200 ms buries the 15 ms gap between sealed and ported. This is why Toole 1, Thiele 2, and Linkwitz 3 all concluded that enclosure topology is a second-order effect. They were right, for uncorrected rooms.
In our room-corrected system, it does matter -- modestly. The FIR pipeline reduces modal ringing by roughly 97% (for a 15 dB mode cut), removing the dominant masking. Enclosure group delay then sits alongside early reflections (5-30 ms) and residual modal energy as one of several comparable time-domain artifacts. The 15 ms difference between a sealed sub and a ported sub is no longer buried under 100+ ms of room ringing; it occupies the same order of magnitude as the early reflection pattern.
This should not be overstated. The difference is subtle, operating near the edge of audibility even by the generous extrapolated Blauert thresholds 4. A well-aligned ported sub in a room-corrected system will sound very good. But if the owner is choosing between enclosure types and other factors (size, cost, efficiency) are equal, lower group delay is a legitimate tiebreaker in a corrected system -- whereas in an uncorrected system, it would not be.
The key insight: cheap, real-time room correction -- running on a $75 Raspberry Pi -- shifts a decades-old consensus. Enclosure topology moves from "negligible" to "one of several comparable contributors," not because the physics changed, but because we removed the thing that was hiding it.
Within sealed designs, the D-009 cut-only constraint favors Butterworth (Qtc = 0.707) over Bessel (Qtc = 0.577). Bessel has lower group delay but sacrifices 6+ dB of output at 30-50 Hz with no recovery path -- a permanent loss in the psytrance kick fundamental range. The FIR correction can slide a Butterworth system toward Bessel-like group delay in post-processing if needed, but cannot boost a Bessel system's lost extension (see 3.1.6) 8.
Yes. The pipeline measures, corrects, and deploys filters for any enclosure type. It does not need to know the topology -- it measures the actual impulse response and computes the inverse. The practical differences are limited to:
- Subsonic protection (D-010): Mandatory for ported, advisable for horn-loaded and transmission line, not needed for sealed.
- Time alignment: Designs with internal acoustic paths (horns, transmission lines) have acoustic centers that differ from physical driver position. The measurement pipeline captures this automatically (D-008).
- Per-sub independence (D-004): Mixed topologies (one sealed, one ported) are handled transparently -- each sub gets its own FIR correction.
Each entry includes an assessed confidence level for the claims attributed to that source in this document.
Footnotes
-
Toole, F.E. (2008). Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms. Focal Press. Confidence: MEDIUM-HIGH -- widely referenced, consistent with field experience. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7
-
Thiele, A.N. Transient behavior analysis of sealed and ported enclosures. Cited by Small; specific paper unverified. Confidence: MEDIUM -- secondary citation; specific paper should be identified and verified. ↩ ↩2 ↩3
-
Linkwitz, S. Writings on LR4 crossover phase distortion audibility. Specific publication unverified. Confidence: MEDIUM -- consistent with known positions; specific publication should be identified for proper citation. ↩ ↩2 ↩3
-
Blauert, J. and Laws, P. (1978). "Group delay distortions in electroacoustical systems." Journal of the Acoustical Society of America, 63(5), pp. 1478-1483. Confidence: HIGH for the measured range (500 Hz - 8 kHz); sub-bass values commonly attributed to this study are extrapolations, not data. ↩ ↩2 ↩3
-
Lipshitz, S.P., Pocock, M., and Vanderkooy, J. (1982). "On the audibility of midrange phase distortion in audio systems." Journal of the Audio Engineering Society, 30(9), pp. 580-595. Confidence: MEDIUM -- cited from memory; claims consistent with known findings. ↩ ↩2
-
Small, R.H. (1972). "Direct-radiator loudspeaker system analysis." Journal of the Audio Engineering Society, 20(5), pp. 383-395. Small, R.H. (1973). "Closed-box loudspeaker systems." Journal of the Audio Engineering Society, 21(1-2). Confidence: HIGH -- foundational, widely reproduced. ↩ ↩2 ↩3 ↩4
-
Thiele, A.N. (1971). "Loudspeakers in vented boxes." Journal of the Audio Engineering Society, 19(5-6). (Originally published 1961 in Proceedings of the IRE Australia.) Confidence: HIGH -- foundational, widely reproduced. ↩ ↩2 ↩3 ↩4
-
Bessel vs Butterworth alignment analysis for cut-only corrected systems (this document). Application of second-order transfer function relationships 6 7 to the D-009 cut-only constraint. Confidence: HIGH for transfer function relationships; MEDIUM-HIGH for the practical recommendation (engineering judgment in an unstudied perceptual regime). ↩ ↩2
-
Olson, H.F. (1957). Acoustical Engineering. Van Nostrand. Confidence: HIGH -- foundational. ↩
-
Beranek, L.L. (1954). Acoustics. McGraw-Hill. Confidence: HIGH -- foundational. ↩
-
Keele, D.B. (1975). "Low-frequency horn design using Thiele/Small driver parameters." AES Preprint No. 1032, presented at the 51st AES Convention. Confidence: HIGH -- foundational. ↩
-
Bailey, A.R. (1965). "A non-resonant loudspeaker enclosure design." Wireless World, 71(10), pp. 483-486. Confidence: MEDIUM-HIGH -- foundational for TL design, widely referenced, but less mainstream than Thiele-Small. ↩
-
Bradbury, L.J.S. (1976). "The use of fibrous materials in loudspeaker enclosures." Journal of the Audio Engineering Society, 24(3), pp. 162-170. Confidence: MEDIUM-HIGH -- foundational for TL design, widely referenced. ↩
-
King, M.J. (1996). Chapter in Loudspeaker and Headphone Handbook, ed. J. Borwick. Focal Press. Confidence: MEDIUM -- referenced in TL design literature, not independently verified against original. ↩
-
King, M.J. (2001-2010). Transmission line loudspeaker analysis and simulation. Published on quarter-wave.com and in peer-reviewed journals. Confidence: HIGH -- peer-reviewed and independently verifiable. ↩
-
HOQS (High Order Quarterwave Society). Paraflex subwoofer designs. Community-developed, shared via diyAudio forums. Confidence: LOW -- no peer-reviewed measurements published. ↩
-
PMC (Professional Monitor Company). Transmission line designs in professional monitoring. Confidence: HIGH -- commercially verifiable, well-documented product line. ↩