Understanding phase

The phase of 2 waveforms is a very important aspect within the realms of audio engineering. I am writing this because I occasionally hear out of phase (or nearly out of phase) signals in the 2 track mix downs that I receive for mastering. It can be quite a serious problem as it has the potential to render the out of phase source silent if the mix is ever summed to mono. (and this might happen more often that one might think) The phase relationship of 2 signals can directly dictate the sonic quality of a 2 microphone recording or synthetic waveform. Initially I will explain phase by explaining the definition of mono and stereo recording in the traditional sense using microphones.

If you have a band playing and you place a single microphone in front of the band and record it to a single track in a workstation you are creating a mono recording. Mono can be defined as a single microphone recording. If you were to place 2 microphones in front of the band spaced apart and recorded their outputs to 2 discrete tracks you have sampled the sound field from 2 points. Upon play back if you pan the signals hard left and hard right you will have captured a stereo field. You will hear a sense of stereophonic space in the recording as the time arrival differences between the 2 mics in the recording will reach your ears and produce the sensation of being in front of the band with distinct L-R separation.

The easiest way to show the phase relationship between 2 waves is to use a simple wave form. We will use the simplest waveform (albeit a mathematical construct) which is a sine wave. A sine wave is a curvy wave that has equal positive and negative energy and equal displacement from the zero crossing at any given frequency. In these diagrams you have 2 sine waves. When the positive and negative cycle of the wave are in sync with each other as in diagram A the waves can be said to be in phase with each other.

Diagram A

When in phase waves produce an increased signal level. When the positive and negative cycles are opposed to each other (Diagram B) the waves can be said to be out of phase with each other, this results in a lowered output level when summed. (in this example they will cancel completely)

Diagram B

In most stereophonic recordings there is a mixture of phase relationships at different frequencies. Music produces a waveform which is complex in terms of frequency content. As each frequency has it’s own wavelength it will mean that the time arrival differences in a 2 microphone recording will be different dependent on frequency and thus a highly complex phaseand frequency relationship is created when recording music.

In multi track recordings and mix downs we usually have either mono or stereo source signals. These may well be true stereo as recorded in a studio such as a pair of drum overhead microphones or pseudo stereo signals created inside a synthesizer with multiple oscillators panned left and right. In addition whenever one uses a stereo delay or reverb you will be creating a pseudo stereo images even if the source is mono. Recording engineers pay great attention to phase when placing microphones to record a band. This is because they understand that signals that are close to being out of phase will cancel and the recording will not have good mono compatibility. A general guideline is known as the “3 to 1 rule”. This means that given 2 mics on a source the second microphone must be 3 times the distance from the other mic than it is from the source. This largely generates mainly in phase stereo recordings and microphone relationships. The goal for a stereo recording is to capture space and depth so there will always be frequency dependent phase relationships that are neither exactly in phase or exactly out of phase. With care and attention this is a good thing as stereo adds realism and interest to a recording over a sole mono microphone recording.

As part of mixing practice it is a very good idea to be able to audibly recognize an out of phase signal. If you cannot audibly identify this yet you will in time and with practice but you can also use a phase meter to help you learn. I suggest importing a stereo mix into your workstation but use the split into L and R mono files capability of your DAW upon import. Align the L and R parts of the track to start at exactly same point in your timeline, pan them hard left and hard right. Then you can try the “phase flip” button on 1 channel. Listen carefully to the result. You may now be hearing less bass or the bass line may nearly vanish, you may not be able to easily detect the position of specific instruments in the stereo image anymore. Try moving your head from left to right slowly and listen to the peculiar shifting of sound as your head travels. This can make you feel very slightly queasy or uneasy as this is very unnatural sound. As an extension of this experiment you can then sum the track to mono using either your monitor controller or DAW master output mono button. You will then hear the bass output may drop significantly and some instrumental balances have changed. This is an extreme example purely for learning how out of phase images sound, if you hear overly wide images in certain recordings, tracks or mix elements there could be a phase problem.

When mixing it is important not to use overly wide stereo images. If the sources approach being out of phase the instrument can almost vanish from the mix when summed to mono. This is obviously not a good situation. There are numerous ways this can happen in the real world. Radio transmissions are often summed to mono in weak signals areas, many PA systems may be wired in mono and even some notebook computers have a single speaker output. Almost out of phase images can be created by over use of “stereo imagizers” or “stereo width enhancers” it might sound impressive and spacious at first but it is important not to push it too far as you can end up with the problems outlined above. There is a freeware tool called “Flux - Stereo Tool” You can use the phase meter to help with identifying overly wide stereo images if you are unable to hear them as yet. Ideally you want to see the meter hovering largely between 0 and +1. If the signal hovers closer to the -1 then there is likely to be a phase problem. You can insert this meter and solo any suspect track.

So when you record and mix consider the phase of the signals between the left and right channels (and between mono mics on a drum kit for example). It will take some time to identify and hear signals not in phase but it is a very important concept to grasp and to aurally recognize.

Barry Gardner operates SafeandSound online mastering services a low cost music mastering studio based online.

Comments

[Darth_Voda]

Very useful and well written article on phase, Mr. Gardner. Thanks for sharing, and for referencing the Flux - Stereo Tool!