panner patch out volume isn't linear - known issue ?

[La Tenaille]

Hi,
I'm using the "panner 1-2" patch and I've noticed a +6dB difference between in and out volume when panning a signal 100% left or right.
Is it a known issue ? Any way to fix that ?
Thanks

[nay-seven]

i suppose there's a mathematic reason , cause you came from a mono signal to a stereo one? ..can't answer precisely

alternate solution, use a cross fade module this way..?:

[gurulogic]

Hi, you can try my "MonsterPanner" for some ideas if you like.
http://www.sensomusic.com/forums/upload ... 2BGain.pat

In stereo mode, each LR channel is brought to the other side and gain compensated -6db during panning, in mono mode,
either or both inputs can be used and when panned is gain compensated -3db
Also there is a unified gain fader that works in conjunction with the pan gain compensation..

[gurulogic]

I have another configuration where pan dead centre bypasses the pan patch for cpu savings if interested...?

[La Tenaille]

Hi, Thanks for your help.
I've tried the crossfader but the trouble is nearly the same (the -6dB happens on centered position).
Gurulogic, I've tried your panner and same result : -6dB when centered if I tune the gain to have 0dB on 100% pans.

... so I've tried to analyse the "panner 1-2" patch and found a partial solution :



I've added the selected modules, where b=4 and fader ranger 0>2. The fader range truncate the numbers above 2 when the panner crosses the center position.
It works but the sound position isn't regular : it changes a few near 0 and a lot near 100...

I imagine a good curve would fix that but I don't have any idea of how to create the right curve...

[23fx23]

have a look to the mapper curve in /data, it will let you create your own panning compensation rule with a clean curve.

[La Tenaille]

yet another module I discover, and much more fun to come ! Thanks 23fx23

[La Tenaille]

Mapper curves are great ! Here is my new panner (fader still -100 to 100):


Gain is now linear and I can freely set the pan slope, great, great, great !

[nay-seven]

great !
well done la tenaille !
can i add this patch to the next Utils pack..?

[La Tenaille]

of course Nay-Seven, I'm pleased to share something
For real time panning the smoother avoids crackles, but I've noticed it costs CPU.

[gurulogic]

Gurulogic, I've tried your panner and same result : -6dB when centered if I tune the gain to have 0dB on 100% pans.

Hmm, yes. my patch seems to be losing -6db somewhere before the volume controller, but as I can tell it's is equal "perceived" gain through the entire pan width, which is what my goal was.
If I increase the volume module +6db and then pan slowly from one side to the other, the output gain is equal to the input gain and to my ears, I hear an even volume all throughough the entire pan. I f I pan quick, my ears play tricks on me and I hear quiet spots at half pan but I think that has something to do with the way I blend one side of a stereo channel into the other when panning.
But maybe I still need to do some tweaks, and perhaps even rebuild around your much simpler looking solution

Anyways, I am glad you have it working to your liking

[La Tenaille]

your explaination sound like psychoacoustics
I'm currently working on multiple delays with spatialization, and loosing gain when panning a delay distort the spatial image in that precise case.

[La Tenaille]

ok, so I was completely wrong !
I relied on stereo vu meters out and that's stupid because the signals add.
So -6dB when centered were ok (0dB when added) and your first solutions were good.
The only issue that seems to remain for panner 1-2 is it boosts the signal of 6dB ?

I've modified my curve panner to make it work too, here it is, but it's now just another approach...
Your explanation Gurulogic made me doubt about my patch and that was a really good thing !