Calculate a fourth order speaker crossover by specifying the filter alignment and implementation before evaluating its response. For an ideal fourth-order Linkwitz-Riley alignment, each branch is −6.02 dB at the shared corner, with complementary low-pass and high-pass magnitudes that sum flat only under the intended phase and acoustic conditions.
Open the Fourth Order Crossover Calculator.
How to Calculate a Fourth Order Speaker Crossover
Fourth order describes the filter order, but it does not uniquely identify the crossover. Butterworth, Linkwitz-Riley, and other alignments can share a fourth-order stopband slope while having different corner attenuation and summation behavior.
The embedded calculator evaluates an ideal fourth-order Linkwitz-Riley response, often abbreviated LR4. Enter a corner frequency and a response frequency, then select the low-pass or high-pass amplitude ratio or its decibel equivalent.
It does not synthesize a passive crossover network or provide capacitor and inductor values. A passive implementation requires a defined circuit and load model; a DSP implementation requires the appropriate filter coefficients or verified processor settings. Those are different design tasks from evaluating the ideal response.
For an actual loudspeaker, the acoustic target also includes each driver’s natural response, phase, spacing, and enclosure. A filter setting labeled LR4 does not automatically make the complete driver output an ideal fourth-order acoustic response.
Distinguish LR4 from Fourth-Order Butterworth
An ideal LR4 branch can be formed by cascading two identical second-order Butterworth sections with the same corner frequency. Each section has Q approximately 0.7071 and contributes about −3.01 dB at that corner.
Two such sections in cascade therefore give −6.02 dB at the corner and an asymptotic fourth-order slope. This is different from designing a single fourth-order Butterworth response, whose section requirements and corner magnitude differ.
Do not identify an alignment solely from two familiar-looking coefficients or a parts count. State the intended transfer function, then choose an implementation that realizes it. Incorrectly mixing section Q values can produce a different response even though the system still contains four poles.
Rane’s Linkwitz-Riley primer explains the cascaded-Butterworth construction and the importance of time alignment. The intended summation depends on phase relationships as well as the individual amplitude curves.
Calculate Ideal Magnitude from Frequency Ratio
Let x = evaluation frequency ÷ corner frequency. For the ideal LR4 response, low-pass amplitude magnitude is 1/(1 + x^4), and high-pass amplitude magnitude is x^4/(1 + x^4).
These are voltage or pressure amplitude ratios in the ideal transfer-function model. Convert either ratio A to decibels with 20 log10(A). Do not use the power-ratio factor of 10 directly on these amplitude values.
At x equal to one, both ratios equal 0.5. Their decibel values are 20 log10(0.5) = −6.0206 dB. An amplitude ratio of one corresponds to 0 dB, while ratios below one produce negative decibel values.
The formulas require a positive corner and a positive evaluation frequency. Enter hertz consistently. A corner of 2 kHz should be entered as 2,000 Hz, not 2, unless the field explicitly offers kilohertz.
Worked Example: A 2,000 Hz LR4 Corner
Assume an illustrative ideal crossover has a 2,000 Hz corner. To evaluate its response at 1,000 Hz, calculate x = 1000/2000 = 0.5, then x^4 = 0.0625.
The low-pass magnitude is 1/1.0625, approximately 0.94118. Its decibel value is approximately −0.5266 dB. The high-pass magnitude is 0.0625/1.0625, approximately 0.05882, or −24.6090 dB.
At 2,000 Hz, both branches have magnitude 0.5, or −6.0206 dB. At 4,000 Hz, x is two and x^4 is 16. The low-pass becomes 1/17, while the high-pass becomes 16/17; their results exchange the 1,000 Hz magnitudes.
Enter 2,000 Hz as the corner and each evaluation frequency in turn. Select low-pass amplitude, high-pass amplitude, and their decibel choices to verify that the units remain clear. An output of 0.5 is an amplitude ratio, not −0.5 dB or half the corner frequency.
Read the Response at Several Frequencies
The following values use the same illustrative 2,000 Hz corner:
| Evaluation frequency | Low-pass magnitude | Low-pass dB | High-pass magnitude | High-pass dB |
|---|---|---|---|---|
| 500 Hz | 0.99611 | −0.0339 | 0.00389 | −48.1987 |
| 1,000 Hz | 0.94118 | −0.5266 | 0.05882 | −24.6090 |
| 2,000 Hz | 0.50000 | −6.0206 | 0.50000 | −6.0206 |
| 4,000 Hz | 0.05882 | −24.6090 | 0.94118 | −0.5266 |
| 8,000 Hz | 0.00389 | −48.1987 | 0.99611 | −0.0339 |
Fourth order approaches a 24 dB-per-octave stopband slope far enough from the corner. It does not mean the level changes by exactly 24 dB between every adjacent pair of octave-spaced frequencies, particularly near the transition.
For example, the low-pass changes by about 18.59 dB from 2,000 to 4,000 Hz, then about 23.59 dB from 4,000 to 8,000 Hz. The response approaches its asymptotic slope gradually.
Sum the Branches with Their Phase Relationship
The ideal LR4 low-pass and high-pass branches have the intended relative phase relationship for non-inverted summation. Their magnitude expressions add to one at every positive frequency, but that simple addition represents acoustic summation only when the required relative phase and level conditions hold.
At the corner, two equal 0.5-amplitude contributions in phase add to a magnitude of one. If one contribution has an additional 180-degree relative phase shift, they instead cancel at that point. The same individual magnitude values can therefore produce very different sums.
Flat summed magnitude also does not mean zero phase shift or zero group delay. The ideal summed response has all-pass behavior: its magnitude remains flat while its phase varies with frequency. Do not describe LR4 as automatically preserving the original waveform’s phase at every frequency.
Linkwitz’s active-filter explanation distinguishes the intended electrical response from the completed loudspeaker behavior. The driver outputs must satisfy the acoustic conditions, including their timing and natural response, for the ideal sum to apply.
Choose an Implementation That Matches the Model
In DSP, verify that the processor’s selected LR4 setting uses the intended corner definition and filter structure. Some interfaces offer separate slope, alignment, and frequency controls; selecting 24 dB per octave alone may not identify LR4.
For an analog active circuit, define the individual sections and account for loading, component tolerances, supply conditions, and gain. Cascading properly isolated active sections is not equivalent to connecting arbitrary passive sections without considering their interaction.
A passive network connected to a loudspeaker has a frequency-dependent load. You cannot obtain a reliable fourth-order parts list merely by applying isolated first-order formulas repeatedly. Choose a complete topology, model its interaction with driver impedance, and verify the resulting transfer function.
Keep each capacitor, inductor, and resistor value separate with its unit and circuit position. A list of section coefficients or a sum of unlike component units does not constitute a wiring diagram.
Check Driver Output and Acoustic Crossover Targets
Measure or obtain the driver’s frequency response and impedance on the intended baffle. Its natural low-frequency or high-frequency roll-off contributes to the acoustic result, as do resonances, breakup, and baffle effects.
An electrical LR4 setting might be appropriate in one system but produce an acoustic response steeper or differently shaped in another. Conversely, a different electrical filter may be chosen to achieve a particular acoustic LR4 target when combined with the driver.
Also check driver operating limits and coverage. A steep filter does not permit any arbitrary crossover frequency, and it does not eliminate all out-of-band excursion. Manufacturer requirements and distortion behavior remain part of selecting the transition region.
Physical separation changes relative path length away from the reference axis. Even a good main-axis sum can have different off-axis behavior. Review the intended listening window rather than evaluating the design from one scalar corner result.
Verify the Completed System with Consistent Timing
Measure each filtered branch individually and the combined output using a consistent microphone position and timing reference. Preserve comparable level and delay conditions so the individual traces explain the combined response.
At the transition, compare amplitude and phase rather than assuming that two traces near −6 dB must sum flat. Investigate any mismatch through driver response, polarity, time alignment, filter settings, and the measurement method.
Save the intended transfer functions, implementation settings or schematic, actual component values, and measured responses together. Label the calculator results as ideal reference magnitudes so later tuning changes can be compared with the original target without mistaking the reference curve for a verified loudspeaker measurement.