## How to read and use a Phase / Vectorscope

6

1

I been using Izotope Ozone 4 for some time, but I'm still a bit intimidated by the Vectorscope in the stereo-imaging section.

When Mastering / Mixing a signal...

• What "lissajous shape" are you trying to achieve?
• And which ones should you absolutely avoid?
• Should some phasing switches be used while verifying your signal (toggle L/R channels, Mono, etc)?

I understand it is used to eliminate Phasing Issues, but how do you actually achieve this? Is this something that has to be typically verified and fixed per instruments or on the overall master mix?

Any free online resources on this subject would be very appreciated.

Thanks!

I guess this is also called a "Goniometer". My bad! – None – 2011-01-13T20:17:05.330

## Answers

8

The lissajous shape is generated by graphing the left signal on one axis against the right signal on the other axis. So consider a few scenarios for a sine wave tone.

• Mono signal: L is always equal to R, so the point (L, R) is always on the x=y line, so the plot is a straight line. Vectorscopes usually rotate the axes 45 degrees, so x=y goes straight up and down instead of 45 degrees.

• 180 degrees out of phase: L is always equal to -R, so again a straight line but on the rotated axes the line x=-y goes straight left and right

• 90 degrees out of phase: L is following R (or vice versa) by 90 degrees, so the plot is a circle.

So "vertically-oriented" patterns indicate in-phaseness, "rounder" patterns are out of phase to some extent, "horizontally-oriented" patterns are out of phase to the point of cancelling each other out.

More complex frequency content will make this look like a mess... phase relationships are wavelength-dependent, so when some are out of phase, others are in phase. You want to look at single frequencies or narrow bands.

Nice, thank you very much for the detailed answer. So is reading a vectorscope in soloed "multiband" effect processor (say 4 seperate ranges) sufficiently narrow to measure for phase-cancellations? Or would it be better to pass it through a thin Q range frequency sweeped via an EQ? – None – 2011-01-14T00:27:22.213