A high pass filter removes low frequencies from a signal while letting higher frequencies through untouched. You set a cutoff frequency, and everything below that point gets progressively reduced. Whether you’re mixing music, cleaning up a podcast recording, processing medical signals, or sharpening an image, the core principle is the same: cut the low-end content you don’t want so the information you care about comes through clearly.
How a High Pass Filter Works
Every high pass filter has two key settings: a cutoff frequency and a slope. The cutoff frequency is the point where the filter begins attenuating the signal. At this exact frequency, the signal is reduced by about 3 decibels, which is roughly half the original power. Frequencies well above the cutoff pass through unchanged, while frequencies below it get progressively quieter the further they fall from the cutoff point.
The slope determines how aggressively those lower frequencies are reduced. A gentle 6 dB per octave slope lets frequencies below the cutoff fade out gradually. A 12 dB per octave slope is steeper and more noticeable. A 24 dB per octave slope acts almost like a cliff edge, cutting off low frequencies sharply. Each octave below the cutoff, the signal drops by that amount. So with a 12 dB slope and a 200 Hz cutoff, a tone at 100 Hz (one octave below) would be 12 dB quieter, and a tone at 50 Hz (two octaves below) would be 24 dB quieter.
Choosing the Right Slope
A 6 dB per octave slope is the simplest filter design and works well for gentle shaping. If you just want to slightly thin out a sound without an obvious filtering effect, this is where to start. A 12 dB per octave slope is more useful in creative musical contexts, giving you enough control to meaningfully clean up a track without sounding harsh.
At 24 dB per octave, the filtering becomes aggressive. You can hear the “edge” where the response suddenly drops off, which makes it less ideal for subtle work on recorded instruments or vocals. However, 24 dB slopes are standard in synthesizer design, particularly subtractive (analog-style) synthesis, where the sharp cutoff helps sculpt sounds while keeping enough bite in the remaining frequencies. For removing subsonic rumble from kick drums and bass instruments, 24 to 48 dB slopes work well because you’re cutting content that shouldn’t be audible anyway.
Using High Pass Filters in Music Production
The most common use of a high pass filter in mixing is cleaning up low-end buildup across your tracks. Most instruments produce some amount of low-frequency content that isn’t musically useful, and when several tracks stack up, that unwanted bass energy muddies the mix and eats up headroom. A high pass filter on individual tracks solves this.
Start by cutting everything below 25 to 45 Hz on the master bus or individual tracks as standard practice. This removes subsonic rumble from microphone handling, air conditioning, and floor vibrations that you can’t hear but that still consume headroom in your mix. For kick drums and bass guitars, the fundamental frequencies sit between 60 and 250 Hz, with the “chest punch” concentrated in the 60 to 100 Hz range. You generally don’t want to high pass these instruments above their fundamental, so filtering at 20 to 40 Hz is typical.
For vocals, guitars, keyboards, and other mid-to-high-range instruments, you can set the cutoff higher. Many engineers start around 80 to 100 Hz for vocals and sweep upward until they hear the body of the voice start to thin out, then back off slightly. The goal is to remove rumble and proximity effect without making the sound feel hollow. Acoustic guitars often benefit from a cutoff somewhere between 80 and 120 Hz, while bright instruments like cymbals and hi-hats can be filtered even higher since their useful content sits well above the bass range.
The Sweep-and-Listen Method
The most reliable technique is to solo the track, set your high pass filter to a moderate slope (12 dB per octave is a good default), and slowly sweep the cutoff frequency upward from 20 Hz. Listen for the point where the sound starts to change in a way you don’t want, then pull the cutoff back down just below that point. You’ve now removed everything that wasn’t contributing to the instrument’s tone. Once you’ve set filters on individual tracks, listen to the full mix to confirm the low end still feels full and balanced.
One Side Effect: Phase Shift
High pass filters don’t just reduce volume below the cutoff. They also shift the timing of the signal slightly, particularly near the cutoff frequency. At the cutoff point of a simple single-stage filter, the output leads the input by about 45 degrees of phase. This means the filtered signal arrives slightly ahead of where it would be without the filter.
In most mixing situations, this phase shift is inaudible and irrelevant. Where it matters is when you’re combining a filtered signal with an unfiltered version of the same source, such as a close mic and a room mic on the same drum kit, or parallel processing chains. The phase difference can cause certain frequencies to partially cancel out, thinning the sound near the crossover point. If you notice this, try matching your filter settings across related tracks, or use a linear phase EQ, which avoids phase shift at the cost of higher processing demands and a small amount of latency.
Cleaning Up Medical and Scientific Signals
High pass filters play a critical role in medical diagnostics. Heart monitoring equipment (ECG/EKG) relies on them to remove a problem called baseline wander: a slow, rolling drift in the signal caused by the patient’s breathing, body movement, or electrical charge building up on the electrodes. This drift can obscure the subtle voltage changes that doctors use to diagnose heart conditions like ischemia, where part of the heart isn’t getting enough blood.
The challenge is that the useful heart signal also contains low-frequency components, so the cutoff has to be set very low. Published recommendations suggest using a cutoff no higher than 0.05 Hz for cardiac monitoring to avoid distorting the diagnostically important ST segment of the heartbeat. A Butterworth-type filter is the standard choice for these applications because it processes quickly and introduces minimal distortion, both of which matter when decisions need to happen in real time. Brain wave recordings (EEG) use similar filtering strategies to separate meaningful neural activity from slow electrode drift and movement artifacts.
Sharpening Images With High Pass Filtering
In image processing, a high pass filter works on spatial frequencies rather than time-based frequencies. Low spatial frequency means gradual changes in brightness across the image (smooth gradients, solid backgrounds). High spatial frequency means rapid changes (edges, fine textures, details). A high pass filter suppresses the smooth, gradually changing areas while preserving and enhancing the sharp transitions.
The filter works by sliding a small grid of numbers (called a kernel) across every pixel in the image. A typical 3-by-3 high pass kernel has a positive value in the center surrounded by negative values. For example, the center might be 8 while all eight surrounding cells are negative 1. As this kernel moves across the image, it increases the brightness of pixels that differ from their neighbors (edges) and suppresses pixels that are similar to their surroundings (flat areas).
This technique is widely used in satellite imagery, microscopy, and photography for edge detection and sharpening. In photo editing software like Photoshop, the “High Pass” filter combined with an overlay blend mode is a popular sharpening method. You duplicate the image layer, apply the high pass filter at a small radius (typically 1 to 5 pixels), and blend the result back over the original. The edges get enhanced while smooth areas remain untouched.
Practical Tips Across All Applications
- Start conservative. Set the cutoff frequency low and increase it gradually. It’s easier to hear or see when you’ve gone too far than to guess the right value upfront.
- Match slope to purpose. Use gentle slopes (6 to 12 dB per octave) when you want natural, transparent results. Use steep slopes (24 dB or higher) when you need to hard-block unwanted content.
- Compare filtered and unfiltered. Toggle the filter on and off frequently. Your ears and eyes adjust quickly, and A/B comparison keeps you honest about whether the filter is actually improving things.
- Don’t filter everything the same way. Each signal source has different frequency content. A cutoff that works perfectly for a vocal track will be wrong for a cello or a kick drum. Set filters individually based on what each source actually needs.