Giniometer Spectrums (Analysis)

[ChocolateChicken]

Here, we have a giniometer spectrum of "Party With Pinkie [VIP]" by Alex S.



And here, we have "Hot N Cold" by Katy Perry.



And this... is "The City" by Madeon.



Wow. A perfect solid diamond.

Who here knows about giniometers and what's going on and can provide some knowledge about this? Like, how did Madeon get that perfect spectrum shape?

Let's see if mylittleremix can actually provide some useful information.

[Mr. Bigglesworth]

Well it obviously has something to do with panning, likely dynamic range detection too.

[itroitnyah]

It probably measures the dynamic range and frequencies or something. Like, the higher you go, the higher the frequency, and then to the left and right are the panning, like Bigglesworth suggested. But can you provide an accurate name for this? I google it and it doesn't come up with those spectrums at all, it just suggests something else.

[SpiderPubes]

Easy, he's french. That's why.

[the4thImpulse]

It's a display of the phase correlation in a track, differences in both the left and right channels (the horizontal measurement) and the phase difference (if one channel is 180 degrees out of phase than it appears in the meter at -1).

There is no perfect shape, that diamond is likely because its been brickwalled limited (and poorly at that).

[the4thImpulse]

Example:

(Keep in mind these graphs are often rotated 45° to the right so you can find the center easier)

In a perfect mono signal the left and right channels are identical ( L=R ), the graph would read it as ( x=y ) which would result in a perfect vertical angle. (with the length being how strong the signal is)

In a complete 180° out of phase signal ( L=-R) the vectorscope will read a horizontal line, this usually results in complete phase cancellations (if the two sides are equal strength).

[GumsOfGabby]

that diamond is likely because its been brickwalled limited (and poorly at that).

Please tell me you're joking...

[the4thImpulse]

that diamond is likely because its been brickwalled limited (and poorly at that).

Please tell me you're joking...

Well if you know any better than why don't you share with us all..

I have analyzed a bunch of songs and the ones that usually produce those rough diamond shapes are the ones that are extreme brick walled.

Soft classical piano (Für Elise) music makes more swirling circle shapes, they have much simpler waveforms than 'dubstep' so you can more easily see the movements of the correlation.

[froggy]

Could be due to brick-walling, but Party With Pinkie [VIP] and The City both look pretty slammed

Party With Pinkie


The City

[GumsOfGabby]

Okay, turns out 4thimpulse was on the right track. My apologies.

The screen cap of the vectorscope analysing "The City" was most likely taken during the chorus, where the track is constantly pushing 0db. The screen cap of "Party with Pinkie" was most likely taken during a section (perhaps within the chorus) which had a lower volume.

[Forza SoundFire]

Yup, it's simply an X/Y graph of the amplitude of the L/R channels.

[ChocolateChicken]

The screen cap of the vectorscope analysing "The City" was most likely taken during the chorus, where the track is constantly pushing 0db. The screen cap of "Party with Pinkie" was most likely taken during a section (perhaps within the chorus) which had a lower volume.

Actually I had screencapped "The City" during one of the verses, where the singer says, "YEAAAAAAAH. I HAVE WHAT I WANT..." and I screencapped "Party With Pinkie" during the final section of the song, where it was brickwalled. "Hot N Cold" was screencapped at the last chorus, which isn't really brickwalled, but it's close. So "Hot N Cold" and "Party With Pinkie" were screencapped where the loudest parts were, and "The City" was screencaped at the not-as-loud part, but all of "The City" looked like a diamond pretty much.


I thought a giniometer was more of a tool to measure stereo width in regard to amplitude. For phase correlation, isn't that just with a horizontal correlation meter, with a "-1" on the left end and a "1" on the other end? Or perhaps those are just mono compatibility meter.
When I summed "The City" to mono, it showed up as a single vertical line in the center, which 4thImpulse said would happen.

[Forza SoundFire]

I thought a giniometer was more of a tool to measure stereo width in regard to amplitude. For phase correlation, isn't that just with a horizontal correlation meter, with a "-1" on the left end and a "1" on the other end? Or perhaps those are just mono compatibility meter.
When I summed "The City" to mono, it showed up as a single vertical line in the center, which 4thImpulse said would happen.

Correlation is easy to see on the giniometer, because a mono signal (or perfectly correlated signal) will appear as a straight, vertical line! ^^

I usually have it up in final mixing, as it helps me deal with setting the mix depth, localisation and such.

[the4thImpulse]

I thought a giniometer was more of a tool to measure stereo width in regard to amplitude. For phase correlation, isn't that just with a horizontal correlation meter, with a "-1" on the left end and a "1" on the other end? Or perhaps those are just mono compatibility meter.
When I summed "The City" to mono, it showed up as a single vertical line in the center, which 4thImpulse said would happen.

Like I said its a comparison between the left and right channels. A mono signal means the left and right channels are duplicates, if you gain the right side up while leaving the left the same the vector scope would show there's more signal on the right while still remaining a straight line.

Phase correlation is the difference between the right and left signals, those differences result in phasing. If the right signal peaks at 1 at the same time the left peaks at -1 the meter would read it as 180 degrees out of phase (should be a horizontal line). Anywhere in between would result in a circular motion (with a pure sine wave).

Mess around with sine waves, delay the left channel. Add some reverb to see how the added frequency context changes the scope.

[Sugarholik]

By the way, this thingie seems to be called either Vectorscope or Goniometer

I thought a giniometer was more of a tool to measure stereo width in regard to amplitude.

Just to reiterate what others said: When the pattern is wider horizonally, the more stereo width there is. Handy amirite?

And here's something to demonstrate the brickwalling and how it shows up on vectorscope (<-- that name sounds cooler!). I mildly distorted Party With Pinkie which is basically extreme brickwalling.




Also here's a bit closer zoom of Party With Pinkie and The City. Can you guess which is which?


I've never really paid that much attention to loudness war business but gosh how The City sounds obnoxious! And yes, this was the first time I heard the song. Say that I live under a rock if you want, it might be true.

[Lavender_Harmony]

Guys, the reason it's a triangle is because it replies on transients and the shape of a wave. Play a sine through a Goniometer and you get a circle:



it's esentially a linear waveform wrapping itself around at the M point. So if your track has little or no transients, like a commercial, high RMS track, it's going to push up against the top of the spectrum constantly rather than having lots of peaks and troughs, like Alex's early material does. Simply put, it's not a gain issue, it's a transient issue. I can guarantee the pre-master of The City had plenty stereo movement, and this mix itself is very clean. If you're worried about phasing when your mix is summed to mono, you want to keep an eye on the correlation meter, not the Goniometer.

[Sugarholik]

Guys, the reason it's a triangle is because it replies on transients and the shape of a wave. Play a sine through a Goniometer and you get a circle

Huh? wait...wut?


One of these goniometers is a hipster. They do not operate with the same principle.

I can guarantee the pre-master of The City had plenty stereo movement, and this mix itself is very clean

Clean: yes. Limited to oblivion: totally.
Also the vectorscope too suggests that there is plenty of stereo movement. (If it works like me, 4th and others think it works) No one said there wouldn't be.

[Alycs]

Looking at it, I'm guessing Lavender's is using Polar Graphing, and yours is using Parametric, as setting sin(t) to equal both x(t) and y(t) for a parametric graph will give you just straight 90 degree lines. However, I'd imagine a good one would be using Polar, as its much more "circle-based"

They work on the same principle, just using different graphs; see if there is an option for changing that and you should get the same result.

((Yup, and there is the "Param" setting on yours, see if you can change that))

[Sugarholik]

They work on the same principle, just using different graphs; see if there is an option for changing that and you should get the same result.

((Yup, and there is the "Param" setting on yours, see if you can change that))

I figured it'd be something like that. I'm more used to the one I'm using though. And that button isn't the param you are looking for. It's effect parameters. For automation and that stuff.

EDIT:
Or! OR! Lavender is just trying to mindfuck us.

[Lavender_Harmony]

Its a straight line because you're feeding it a monophonic signal. I fed it a duplicate with a slightly offset phase.

[the4thImpulse]

Or! OR! Lavender is just trying to mindfuck us.

Lavender is saying pretty much what I said in different words.

[Sugarholik]

Its a straight line because you're feeding it a monophonic signal. I fed it a duplicate with a slightly offset phase.

I noticed.

Or! OR! Lavender is just trying to mindfuck us.

Lavender is saying pretty much what I said in different words.

Yea. Still got me a bit mindfucked
Never mind, carry on...