Understanding Phase and the Phase Correlation Meter

Hello All.

I have some basic questions to do with phase in samples, and thePhase Correlation Meter.

  1. To obtain a great sounding mix I must avoid sounds that are out of phase ?

  2. Renoise Phase Correlation Meter shows audio out of phase when the meter goes to+1 ?

  3. Only stereo audio can be out of phase ?

I bought sample pack from different companies I consider being professional, but now I discover that 20% of all my samples meters +1, it just made wonder how you guys deal with this, and if its important at all ?

Also I noticed that some VST effects puts audio out of phase, is there a reason for doing this ?

Hope someone help and explain the relationship between these phase questions, thanks in advance.

Hello All.

I have some basic questions to do with phase in samples, and thePhase Correlation Meter.

  1. To obtain a great sounding mix I must avoid sounds that are out of phase ?

  2. Renoise Phase Correlation Meter shows audio out of phase when the meter goes to+1 ?

  3. Only stereo audio can be out of phase ?

I bought sample pack from different companies I consider being professional, but now I discover that 20% of all my samples meters +1, it just made wonder how you guys deal with this, and if its important at all ?

Also I noticed that some VST effects puts audio out of phase, is there a reason for doing this ?

Hope someone help and explain the relationship between these phase questions, thanks in advance.

Hello !
Which phase do you mean ? There is a vertical phase that can be seen on stereo spectrometer . And a horizontal phase that can make your sounds more left or more right .
Vertical phase = up / down

Horizontal phase= left right

Negative correlation means that your horizontal phase is getting out of balance
The balance of the main phase depends on the bpm and the song dynamic range . It is not bad to have -1 correlation in the break part of the song " There is no rhythm "
When rhythm is acting, it is preferred to have ~ +1 correlation " When your Bass sound get stereoed = The +1 correlation will swing to 0 then -1, depends on the stereo specter and the amount of wide sounds in the low band area “0 - 300 hz”

Renoise correlation meter shows when this dynamic/stereo relationship is in harmony. When your correlation range starts to swing between -1 and +1 two often = Make your low band "0 - 200 hz " range more MONO . Dont use wide kicks or something like that if you want to save the dynamic of your track.

The formula of phases is simple :
1- Stereo sounds = Your correlation range will swing between -1 and +1
2- Mono sounds = Your correlation range will STAY in +1 area .

Wide sounds in low pass area " o - 300 hz " = Correlation will Swing between 0 and +1
Wide sounds in SUB area " 0 - 100 hz " = Correlation will swing like a crazy b*** :stuck_out_tongue:
Try to balance between horizontal and vertical phases.

Hey

The horizontal one was the one that made me wonder. The one in the bottom, where the white line moves from -1 to +1.

I see I made a mistake in my first post, because it seams to me now that -1 is out of phase, and not +1, if I got it right, still trying to understand it all, its a bit confusing.

looking at random stereo samples, none goes below 0, no swing between -1 and+1 here!

Try to balance between horizontal and vertical phases.

I have no idea how/where to do that :unsure: :huh:

Thanks for the help.

I see I made a mistake in my first post, because it seams to me now that -1 is out of phase, and not +1, if I got it right, still trying to understand it all, its a bit confusing.

looking at random stereo samples, none goes below 0, no swing between -1 and+1 here!

-1 is called a Negative Correlation. It is not bad when pointing in ambient music or in some parts of your track, where the rhythm stops / BUT / when rhythm goes on = Try to make the correlation near to +1 as much as possible. Unless you will have tons of problems in mono systems and you will lose a lot of dynamic.

Having a stereo samples that swing around 0 means that they are not 100 percent wide . You can make them absolutely wide by :
1- Putting a delay on the track
2- Mute the source in it " Mute Src."
3- Set the L and R feedback to 0
4- Set the L Delay to 1 \ R Delay to 10 :>>> it will start to swing. The more difference between L and R delay ms, the wider goes your sound." Till it starts to Stutter" or act like a fast ms delay between L and R.

About the Horizontal / Vertical phase balance : Try to select more mono sounds when getting near to the 500 hz zone / more stereo when getting to the 8 khz zone.
The kick should be absolutely mono / unless you will lose the dynamic as well
The clap can be 80 % mono / 20 % Stereo [Inside of each sound try to make the balance between the mono lowpass area and the stereo highpass area]

Do not use stereowiders or any kind of stereo imagers on your master channel [This will kill the vertical phase because it manipulate the whole specter of the track]

the phase correlation meter is a means to analyse how similar (or opposed) frequencies of the left channel are in respect to the right, and at the same time the other way round. +1 means both sides identical, resulting in twice the amplitude when summed together. -1 means they are completely inverted from each other, resulting in some strange very stereo sound but also in silence when both channels are summed. zero point is neutral, so when summed the amplitude will be the same as the mean of both stereo channels.

A pure mono sound would always be +1. Inverting one stereo channel of a mono sound -1 (the gainer device can invert either left, right, or both). The stereo expander can reduce width to mono mix, if you use l+r mix mode you will see the effect of the correlation in the (mono) result, i.e. a bass totally out of phase would result in silence. You can also put filters on the master channel to analyse specific frequency ranges for their correllation.

Having out of phase sounds will mean they will sound weird on headphones (though it can be an interesting experience, only parts of sound out of phase like with reverbs can sound very…wiiiiiiide), or when playback via speakers the frequencies can sound with a different balance depending on the position you have relative to the speakers. Also you should consider your music being converted to mono at some occasions, i.e. a big mono soundsystem in a club, a mono playback smartphone, a subwoofer that will sum left/right for bass frequencies (some sub setups also seem to choose either left or right for the sub).

Theres lots of things to consider when it comes to phase correlation/stereo width, its black magic. Every stereo effect and stereo layerings etc. will affect it in different, not always perfectly controllable ways, so if you want to be anal about the few dbs to be gained by correlation vs width games, you have to watch your sound design and mix business closely for possible phase problems. It is common to keep at least the bass in mono (below 100-200 hz - correlation near +1), or at least almost completely mono, to make sure it will always have full power, and also as stereo bass isn’t useful anyways, only a bit weird on headphones. Some people mix/eq/balance their sounds with a mono mix on master completely in mono (just put stereo expander on master channel), so their balance of stereo sounds will work better regardless of the angle the listener has to the speakers.

1 Like

It is common to keep at least the bass in mono (below 100-200 hz - correlation near +1), or at least almost completely mono, to make sure it will always have full power, and also as stereo bass isn’t useful anyways, only a bit weird on headphones. Some people mix/eq/balance their sounds with a mono mix on master completely in mono (just put stereo expander on master channel), so their balance of stereo sounds will work better regardless of the angle the listener has to the speakers.

Low pass sounds can swing the correlation more effectively than the high pass sounds.
For example : Sub bass with a stereowider will cause a dis-balance in the correlation specter 90 % stronger than a stereo whitenoise …etc
The correlation swings stronger between +1 and -1 when Low band sounds get wider / weaker when other sounds are stereo " from 300 hz to 20k "
The area between 0 and 150 hz cant be Absolutely stereo [Stereo Sub can kill the dynamic if it is not sidechained by a strong mono kick …etc]
A lot of nuances can flip the whole phase of your track , especially in low band area,

We should always remember that there is a parallel VERTICAL phase that gives the sound a 3D feeling , it can be controlled by a stereowider on each channel.

Thanks for helping me out guys, think I slowly get it… :mellow:

Guess its a balance in general. I had Levels from29palms as a track inspector and it just shows stereo image problems in everything I ever made, even when it sounds good… I will go by the rules and advises from here and keep the balance in mind.

back to making beats :drummer:

Thank you people for developing this topic. I find it interesting!

Renoise’s Phase displays would be much, much more useful if it was following the track selection. For master select master track.

And maybe even following the slope factor as weighting, so at a +4dB slope the phase of high frequencies was stronger weighted…

1 Like

Yes such an update would be very fortunate. Also for the scopes. I’ve been dreaming of a new native “meta device” that you place somewhere in the dsp chains and that will let you select an analyser slot and a gain, so you can view action in your dsp chains also from within effect chains, doofers and instruments.

1 Like

horizontal phase is left/right, stereo-wise - the difference of overlaying the left and right portions of a sample

vertical phase is the forward/backward motion of the speaker (a low sine pushes the speaker forward and then backward, in an observable manner)

you could look at the phase meter “from above” as if both left and right speaker were positioned at the exact same point in space. this point-source is the eruption of the sound, the phase meters “vertical” movement is what the speakers motion looks like (forward/backward), the phase meters horizontal movement shows in which direction the sound is pushed as a result of two stereo speakers which actually are in different space-wise positions.

(there is a tilted square in the background of the phase meter, in that manner of looking at the things, the corner at the top would be the front, pointed at the audience)

that’s not the scientific explanation but is it at least a logical metaphor when not overinterpreted?

And now we are coming up on 8 years later. Since no one seems to have shared how to ACTUALLY BALANCE a stereo signal that has severe correlation issues in a mix, or mentioned the fact an unbalanced stereo field will lose mono compatibility, as well the compatibility with 2 channel surround systems (Dolby) and pro-audio gear. If you disagree with that statement for what ever reason, check out the Creative Lab Pebbles, which have a

“Far-field drivers and passive radiators, custom-tuned far-field driver solution with rear facing passive radiators for excellent audio and enhanced bass reproduction, and a 45° elevated drivers for enhanced audio projection and an immersive personal listening experience.”

It’s surround sound that doesn’t violate patent laws, or I would assume, either way, if your mix isn’t mono compatible, it won’t be “Far Field” compatible either.

IF YOU DON’T CARE ABOUT THE TECHNICAL BS OR THE CONCEPT… SKIP TO “THE METHOD”
(You could just buy the software that does it, and hope one of the presets works, but I am broke and OCD)

It is important to shift your paradigm a bit for this to make sense, because there are in fact 5 different stereo modes, which I only figured out because WinAmp will display the stereo mode, and I noticed that it sometimes said Pseudo Stereo, and other times it said True Stereo. And when programming audio software, the wave writer does contain 5 different flags, granted there in no explanation anywhere that I found other than 0 is Dual Mono, and 1 is Mono or Pseudo Stereo. However there are 5, according to WinAmp there is MS Stereo and True Stereo, and I never learned what the last mode was, but I am pretty sure it is used for stereo signals that contain an LFE channel.

I have concluded that True Stereo is the same thing as having a Balanced Left and Right channel in the professional audio world, which means you have 2 parity signals channels that accompany your left and right channels… I am going to start calling them voice channels since that is how the computer sees them, the change in paradigm is this.

Balanced Mono and L/R Stereo are the same, because they both contain 2 voice channels, Mid Side signals have 3 voice channels,1 for your mid, and 2 for your side. The Mid Side signal still uses the same 2 voice channels, but it processes that signal differently. The Left channel is processed as a Mid signal, and the Right is processed as a side signal. True Stereo contains 4 voice channels, and CANNOT be achieved by summing to 2 channels in a digital environment, because the inverted signals will cancel out the L/R signals. I feel another paradigm that has to shift is what the 3rd prong on an XLR is for, as a consumer, the only benefit to using an XLR is that it cancels out line noise, or so we are told, however professionals use XLRs for multiple reasons, none of which anyone seems to want to share. So lets us deduce how balanced stereo is achieved with 2 XLR cables, but not with a stereo 1/4 inch, even though a 1/8 inch aux cable can carry a balanced stereo signal from your computer to your “Pebbles.” Which means the Pebble’s are using the parity signals to generate the Mid, SLeft, and SRight channels using only the left and right channel. This is why correlation is important, I use the Correlmeter from Voxengo, which allows you to view any problems with correlation in your stereo field as well. the difference between your Left and Stereo Left should either be at +1 across the board, indicating a mono sound, or be dancing around 0 evenly, which seems to mean it is balanced. Long story short, you need to account for all 4 voice channels in your mix if you want a balanced stereo mix and not a pseudo stereo mix.

THE METHOD!!!

Ok, so first you are going to want to set up some busses. I use : Drums, Bass, Synth, FX, and Pre-Master. Then throw a correlmeter on all the busses. Lets assume you see a problem with your Drum Buss, first, you want to find which channel is having the issue, the correlmeter will have a rhythm to it so it should be easy to spot the issue right away. Let’s say you notice a glitchy drum loop is culprit, you could toss it, but it happens to be the glue holding the drums together. Your mind set at this point should be you are going to use the Left channel as your Mid, and your Right channel as your Side in order to eliminate the the phase issue between L and R. You are going to need MSED.

This will require you duplicate the channel with the problem and then send both output to the same bus.
METHOD #1 - Hard-pan one channel all the way left. On the other channel, you want to invert the right signal and then add on an MSED set to encode.If for some reason this doesn’t leave you with just a right signal, you will want to hard-pan right after MSED (which happens in you don’t start with a balanced mono).
on the bus, drop another MSED on and set it to decode. Basically you are encoding the parity signal into the wave file so the DAC doesn’t combine the signals and cancel out all your sound.

METHOD #2 - Render or bounce the LR channels into separate files, drop those file onto separate channels, and again, send the output to the same bus. On the Left channel, just drop an MSED set to encode. On the Right, invert the right channel, and drop on an MSED set to encode. This should make it so the channels appear hard-panned, and then on the bus, you want to drop on an MSED set to decode. As a side note, you can add a stereo delay (haas effect) to your side mix before it is decoded back to LR to increase the width without destroying mono compatibility.

So now that channel is now longer having issue, and your Drum Buss Correlmeter is mostly even at 1, you move one, you probably had to do the same thing to your FX bus… But now you are having problems on your pre-master Bus in the low frequencies. This is probably do to line noise in one of the recording. This is why you want to adhere to the Rule of 300, which states, if it is not a Kick, a Snare, or a Bass, then you want to highpass the channel at 300 hrtz. I make Bass music, and I am part of a fairly large community of electronic music producers, and i have attempted to share the rule of 300, which is a term coined by Nyquist I believe, and every time I am told that that would destroy my mix, by people that are 100% “experts” on the subject. I personally will use the rule of 300 every time, and mixes are always perfect… even when my mixes are crap.

Anyway, I had to change the way I understood what Stereo was, which is to say stereo is 4 channels, pseudo stereo is 2 channels.