What is clipping ?
What is clipping?
Clipping is the name used to describe the result of a digital recording or mixing system being over loaded by an input or output signal level. Clipping means the the mathematics that represent an audio signal in the digital domain have exceeded. The finite numerical values end at 0dBFS and beyond this point invalid values are recorded or reproduced which produce distortion.
Digital systems are closed meaning that there is a very definite line between accurate representation of the signal and gross distortion. With analogue for example this is quite different because the distortion increases much more gradually and progressively as the level increases. Electronic components and/or recording medium (such as tape) slowly saturate as they exceed their signal level capabilities or that of the power supply providing the power to the circuit.
Digital distortion for the first 1 - 2dB above 0dBFS (zero decibels full scale) is often not audible on inadequate monitoring systems. In the best reference systems this can be audible as the high frequencies and upper mid range starts to become edgy, strident and harsh. As digital distortion increases the bass in a track will start to sound crunchy and the clipping will cause harmonics which sound unpleasant to the ear.
Often a low quality digital to analogue converter coupled with near field speakers will not have the clarity and detail to be able to reveal this clipping distortion. A novice producer or engineer can mistake this for not being distortion at all. It can be difficult to hear the first 1dB of peak distortion above 0dBFS, even on a good system.
Analogue distortion can have a very different harmonic character than digital distortion and generally speaking analogue distortion is regarded as being slightly less obtrusive or more musical than digital distortion. In fact this is often used to good effect in guitar amplifiers and tape machines. Analogue distortion is progressive and the harmonic relationships relate to the fundamental frequencies more than digital distortion.
Of course clipping has it’s own character and it is subjective as to whether any type of distortion be it digital or analogue is a subjective enhancement or not.
There are currently only 2 reasons to clip audio, one being to increase perceived loudness and the other as a subjective enhancement. Few people agree that clipping produces beneficial changes in sound however for some aggressive musical styles it may suit the material if very high perceived volumes are required. For either reason one has to very careful consider the negative impact of clipping. As a producer or engineer you must consider the possible effects of clipping further down the reproduction line. Audio can be clipped digitally “in the box” or during recording a signal via the analogue to digital converters.
What is the problem with clipping audio ?
Cheap digital to analogue converters suffer from bad audio fidelity when audio is clipped. In addition to the inharmonic artifacts caused, clipping tends to suggest very high signal levels and a cheap DA converter will have electronics that are ill equipped to deal with very high signal levels so these factors are compounded. This causes a loss of fidelity and detail and a much more distorted reproduction of the sound.
Lossy file formats such as MP3 and AAC also react badly to clipping artifacts, again producing even more distortion and loss of overall fidelity in files which are already compromised because they extrapolate data from the full resolution file and you have a lower quality audio file to begin with. I do not recommend clipping recordings or sound sources in mixing or the master output of a digital audio workstation. I feel the compromises are significant and that music is not usually best served. However from time to time with specific tracks it can be use with care and the understanding of the possible side effects further down the listening chain.
To avoid clipping, the gain structure of your mix is very important. I suggest working in a way that avoids clipping altogether. I suggest operating your DAW at 24 bit resolution and keep the recorded peak signals at around -12dBFS. This produces adequate headroom for mix down. When mixing try peaking your drums sources at -12dBFS and balance all other song elements against these tracks as a reference. You will hear better mix details and have plenty of headroom to avoid clipping.
How do I identify clipping?
Clipping can be identified in 2 ways.
1) During recording or mixing by looking at input and output metering. If your meter hits the end of the meter you have probably clipped the audio if you do not have a limiter in place. (when mixing) Peak signals lower to avoid this occurring.
2) In an audio file clipping can be identified by zooming in closely on one of the loudest sections of the track and looking at one of the biggest peaks at that point in time. Zoom in deeply and if you see a flat topped or “squared off” top of the waveform it is likely the file has exceeded 0dBFS (also known as clip point, end stops, or “overing”/ “overed”)
Summary
Clipping audio is essentially a practice which has negative side effects and it is generally best practice to avoid it. It is after all distortion, however in some unique circumstances clipping can be performed with a minimal of audible effects as long as all factors are taken into consideration. It is never an easy option or choice as it is always a compromise. It is best to avoid clipping music because of the unpredictable response of low quality reproduction systems that an end user maybe be listening on.
This is a post from Barry Gardner who operates SafeandSound Mastering. An online mastering studio which is highly experienced at low cost analog dance music mastering
Reader Comments