Biquad Filter Calculator for DSP Coefficients

Biquad Filter Calculator

Calculate normalized IIR biquad coefficients for musical EQ, crossovers, tone controls, notch filters, and embedded DSP with response and stability checks.

🎚 Quick DSP Presets

🎛 Filter Inputs

Uses the RBJ audio EQ cookbook coefficient forms.
Nyquist is half the sample rate; keep Fc below Nyquist.
For peaking, notch, band-pass, and all-pass this is the center frequency.
Q controls resonance and bandwidth; shelf filters use this as slope S.
Used by peaking EQ and shelf types; ignored by flat filters.
Estimates headroom after boost and cascade count.
Repeats the same biquad for steeper slopes or stronger EQ.
Quantized output lets you inspect embedded DSP coefficient risk.
Coefficient signs are shown for common Direct Form code.
Some libraries store feedback coefficients with opposite signs.
Feedforward Coefficients
b0 b1 b2
normalized by a0
Feedback Coefficients
a1 a2
Direct Form feedback terms
Pole Radius
0.000
stable when less than 1.0
Response At Fc
0.00 dB
estimated group delay

Coefficient Export

NameDecimalFormatted

📈 Biquad Spec Grid

2
Poles per biquad
5
Normalized coefficients
Fs/2
Nyquist limit
<1
Stable pole radius

📝 Filter Type Reference

TypePrimary UseMain Formula ControlTypical Range
Low-passRemove treble, anti-alias smoothing, crossover low branchFc and Q; Q 0.707 gives Butterworth corner20 Hz to 0.45 Fs
High-passRemove rumble, DC, stage vibration, crossover high branchFc and Q; Q below 1 avoids strong resonance10 Hz to 0.45 Fs
Peaking EQBoost or cut a focused music bandFc, Q, and gain where A = 10^(gain/40)Q 0.3 to 12
ShelfTilt bass or treble while leaving the opposite band mostly flatFc, shelf slope S, and gainS 0.3 to 1.5
NotchReject hum, whistle, resonance, or a narrow feedback toneFc and high Q; alpha = sin(w0)/(2Q)Q 5 to 50
All-passRotate phase without changing ideal magnitudeFc and Q set the phase transition shapeQ 0.3 to 5

🧮 Coefficient Formula Table

QuantityEquationMeaningCalculator Use
w02 pi Fc / FsDigital radian frequencyMaps Hz to the unit circle
alphasin(w0) / (2Q)Bandwidth control for most filtersSets damping and pole radius
A10^(gain / 40)Amplitude factor for EQ and shelvesConverts dB gain into coefficient scale
Normalizeb0/a0, b1/a0, b2/a0, a1/a0, a2/a0Standard one-section biquad formReturns coefficients ready for code

📊 Practical DSP Ranges

ScenarioSample RateFilter TargetPractical Note
Music plug-in EQ44.1 or 48 kHzPeaking and shelvesUse float coefficients and automate gain smoothly
Embedded speaker DSP48 or 96 kHzCrossovers and correction EQCheck fixed-point rounding and pole radius
Measurement cleanup48 kHzNotch or high-passHigh Q notches can ring; verify time response
Synth or pedal code48 kHzLow-pass and all-passClamp Fc during modulation to avoid instability

🎛 Common Preset Targets

PresetTypeStarting ValuesWhat To Watch
Rumble HPFHigh-pass35 Hz, Q 0.707, 48 kHzToo high can thin kick and bass fundamentals
Hum NotchNotch60 Hz, Q 18, 48 kHzVery high Q can expose coefficient precision limits
Bass ShelfLow shelf110 Hz, S 0.8, +4 dBBoost may need input trim before the filter
Two Stage LPFLow-pass2400 Hz, Q 0.5, cascade 2Cascading doubles slope and passband effect

🔍 Implementation Notes

Direct Form IGood when separate input and output delay lines match older fixed-point examples.
DF2TCommon for floating point audio because it uses two state variables per section.
Fixed PointQuantize coefficients, then re-check pole radius and response before deploying.
Cascade OrderPlace sensitive high-Q or high-boost sections carefully to manage internal headroom.
Tip: For musical tone controls, start with Q around 0.707 and adjust by ear or measurement after the coefficients are stable.
Tip: If a fixed-point export moves pole radius close to 1.0, lower Q, raise precision, or split the shape across cascaded stages.

A biquad is an second-order recursive filter. A biquad is a tool used to create digital filters for audios. A digital computer dont understand curves.

A computer must use a stream of numbers to create a digital filter. These coefficient create the filter. The five coefficients will determine how the filter function.

What is a biquad filter

Using an incorrect coefficients will create a digital scream that can damage you’re speakers. Coefficients is required to convert a stream of numbers into a digital filter. A biquad is efficient and versatile.

A biquad is a second order recursive filter. This means that a second-order recursive filter use the current input. However, a second-order recursive filter also use the previous inputs and the previous outputs to calculate the next sample of an audio.

This allow the biquad to create very steep slopes and very sharp peaks in the digital filter. Additionally, the way that a biquad uses previous samples of the audio to calculate the next sample means that the biquad does not use much processing power to perform its duty. Furthemore, biquads utilize a parameter call the Q factor.

The Q factor is a measurement used in a biquad. The Q factor allow the user to control the width of the filter. For example, using a peaking EQ with a high Q factor will create a narrow spike in the audio.

Using a low Q factor will create a broad swell in the audio. Additionally, using a low-pass filter with a Q factor of 0.707 will allow for a Butterworth response. This Butterworth response will allow the audio to have a flat passband before the cutoff frequency.

Using a Q factor more higher than 0.707 will create a resonant peak at the corner frequency of the filter. This resonant peak will only be useful for use in synthesizers, not in creating transparent audio crossovers. Another requirement for a biquad is stability.

A biquad can become unstable due to its use of feedback. A biquad will become unstable if the pole of the filter go outside of the unit circle. If the poles of the filter go outside of the unit circle, the audio output will continue to grow in magnitude until it either clip or the system that is running the program crash.

This can happen if the corner frequency is too close to the Nyquist limit. The Nyquist limit is half of the sample rate. Therefore, both checking the pole radius and ensuring that the filter remains within its stable range are required to ensure that it remains stable and performs as it should of.

When implementing a biquad filter, one must choose between floating point math and fixed point math. A programmer can use fixed point math in instances where the program is compile for a microcontroller. This is because fixed point math is faster on cheaper hardware.

However, fixed point math introduce rounding errors into the program. If the precision of the coefficients isnt appropriate for the hardware that will compile the program, the audio signal can either shift in frequency or become unstable. The precision of the coefficients can be seen in the coefficients in Q1.31 or Q1.15 formats.

Another important structure to understand for the biquad is the structure of the algorithm. The most common is known as the Direct Form I. However, the Direct Form II Transposed is the preferred structure for floating-point implementations. The Direct Form II Transposed is preferred than the Direct Form I structure because the Direct Form II Transposed is more computationally efficient and handle noise better.

Additionally, the user should not increase the Q factor to create a steeper slope. This is because increasing the Q factor will create a massive peak in the output of the audio. An alternative is to cascade the biquad filters.

Cascading biquad filters mean stacking multiple biquad filters on top of each other. Stacking multiple biquad filters will increase the slope of the digital filter. Additionally, cascading the filters is the preferred way to create a steep cut filter since cascading the filters will not ruin the phase response of the signal.

Biquad Filter Calculator for DSP Coefficients

Leave a Comment