It is the harsh distortion you get when a digital signal is pushed past the maximum it can store.
Clean peaks stay rounded below 0 dBFS; pushed too hot, they slam the ceiling, go flat, and turn to permanent fizz.
What it is
Harsh, unfixable distortion that happens when a digital signal goes louder than the maximum number the system can store.
Key facts
Digital has a HARD ceiling called 0 dBFS (decibels Full Scale) - the loudest sample possible. You cannot go above it.
All digital levels are NEGATIVE dBFS: 0 = max, -20 dBFS = quiet, -infinity = silence.
Clipping = the wave's peaks get chopped flat because the number maxes out and stops counting; flat tops add harsh harmonics = the fizz.
16-bit (CD) stores 65,536 levels; 24-bit stores 16,777,216 levels - both top out at the same 0 dBFS.
Dynamic range: 16-bit = approx 96 dB, 24-bit = approx 144 dB. Formula: range (dB) = 6.02 x number of bits.
+6 dB = double the amplitude (voltage). -6 dB = half. -3 dB = half the POWER (watts), a different rule.
Analog clipping is soft-ish; digital clipping is brittle, ugly and almost always UNUSABLE.
Once a sample hits 0 dBFS, the data above the ceiling is GONE - turning it down later cannot restore it.
Safe target: keep peaks around -6 to -12 dBFS. Use a True Peak meter and aim for -1 dBTP to catch inter-sample peaks.
Bit depth controls headroom/noise floor; sample rate (44.1k, 48k Hz) controls frequency range, NOT clipping.
How it works
Signal gets converted to numbers - each sample is a value on a fixed scale.
The scale has a top value = 0 dBFS. That is the biggest number available.
Push too hot and a peak needs a number bigger than the max.
The system can't store it, so it writes the max value instead.
Every peak gets pinned to the same flat top = a squared-off wave.
Flat tops generate harsh harmonics = the crackle and fizz you hear.
Real examples
Gain on a digital mixer cranked too high - vocal peaks hit 0 dBFS and the channel meter sticks red and crackles.
Recording a loud kick drum at -1 dBFS - the transient clips and the punch turns to fizz.
A mastered track exported too loud - the limiter pins waveforms flat and streaming sounds harsh.
DJ pushing the master into the red on a controller - the whole mix gets brittle and edgy.
Input gain too hot for a hot guitar DI - the recording is permanently chopped.
How it helps in live sound
Set gain structure first: ring out each channel so peaks sit around -12 dBFS, never the red.
Watch the desk's PEAK/CLIP LEDs - if a channel light flashes red, pull its gain/trim down.
Leave master bus headroom: aim master peaks at -6 dBFS so transients have room.
Use a True Peak / -1 dBTP target on any limiter feeding speakers or a recorder.
On digital stage boxes, clipping at the PREAMP is unfixable downstream - fix it at the input, not the fader.
If it sounds fizzy/brittle not just loud, suspect clipping before you reach for EQ.
Everyday analogy
Like pouring water into a glass that is already full - the extra has nowhere to go, so it spills flat across the top instead of rising higher.
Watch out
Myth: 'I'll just turn it down later to fix the clip.' Wrong - the peaks above 0 dBFS were never stored, so lowering the fader only gives you quiet, distorted audio. Clipping is captured permanently.
Fun fact
A signal can momentarily exceed 0 dBFS by an 'inter-sample peak' your normal meter never shows - which is why mastering engineers target -1 dBTP, not 0.
Key takeaways
0 dBFS is a brick wall - you physically cannot store anything louder.
Digital clipping is ugly AND permanent, unlike soft analog overdrive.
Keep peaks at -6 to -12 dBFS and never let meters hit the red.
More bits = more headroom (about 6 dB per bit), not a louder max.
Fix levels at the input/preamp - you can't repair a clip after the fact.