Thanks Frank!

Pierre

2012/2/6 Frank Barknecht fbar@footils.org

Hi,

On Mon, Feb 06, 2012 at 02:22:34PM +0100, Pierre Massat wrote:

...

I need a simple equal-power crossfade between two signals. I asked the

same

...

question a few years ago, but i just can't remember how to do it...

This is used in rj's e_pan.pd:

left = cos(p) * signal right = sin(p) * signal

where p is in radians from 0 to PI/2 (i.e. multiply your 0...1 panning by 1.5708...)

Ciao

Frank Barknecht Do You RjDj.me? _ ______footils.org__

---

Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

There are some pan algorithms in the 'pan' library also, included in Pd-extended. They're all pd abstractions.

.hc

On Feb 6, 2012, at 9:29 AM, Jaime Oliver wrote:

Here's an argument for plain linear crossfade. I get power "boosts" with cosine crossfades...

best,

J

On Mon, Feb 6, 2012 at 8:53 AM, Pierre Massat pimassat@gmail.com wrote:

...

Thanks Frank!

Pierre

2012/2/6 Frank Barknecht fbar@footils.org

...

Hi,

On Mon, Feb 06, 2012 at 02:22:34PM +0100, Pierre Massat wrote:

...

I need a simple equal-power crossfade between two signals. I asked the same question a few years ago, but i just can't remember how to do it...

This is used in rj's e_pan.pd:

left = cos(p) * signal right = sin(p) * signal

where p is in radians from 0 to PI/2 (i.e. multiply your 0...1 panning by 1.5708...)

Ciao

Frank Barknecht Do You RjDj.me? _ ______footils.org__

---

Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

---

Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

-- Jaime E Oliver LR

jo2357@columbia.edu www.jaimeoliver.pe 858 750 0924 (cel) <linear-cross.pd>_______________________________________________ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

'You people have such restrictive dress for women,’ she said, hobbling away in three inch heels and panty hose to finish out another pink-collar temp pool day. - “Hijab Scene #2", by Mohja Kahf

On Mon, Feb 06, 2012 at 09:29:33AM -0500, Jaime Oliver wrote:

Here's an argument for plain linear crossfade. I get power "boosts" with cosine crossfades...

Your example is a bit broken (there is no panning going on at all), but I get the same result with a the fixed version or with [e_pan] in the rj library when adding left and right signal together and taking the RMS.

The "boost in the center" is to be expected: In the center with CP-panning, left and right channels both are multiplied by about 0.707 instead of 0.5 with a linear pan, so adding them together will make the center louder. Perception-wise this results in the signal appearing nearer to you than with linear panning, which is the reason this is used: With CP-panning the signal will appear to move around you in a constant circle whereas with linear panning it will move farther away from you as it gets to the center, it will move as if on an oval.

Avoiding this kind of movement is the motivation for CP-panning.

Ciao

Frank Barknecht Do You RjDj.me? _ ______footils.org__

Le 2012-02-06 à 14:42:00, Frank Barknecht a écrit :

On Mon, Feb 06, 2012 at 02:22:34PM +0100, Pierre Massat wrote:

...

I need a simple equal-power crossfade between two signals. I asked the same question a few years ago, but i just can't remember how to do it...

This is used in rj's e_pan.pd:

left = cos(p) * signal right = sin(p) * signal

where p is in radians from 0 to PI/2 (i.e. multiply your 0...1 panning by 1.5708...)

That's for panning, not for crossfading.

crossfading is similar, but is the other way around :

signal = left * cos(p) + right * sin(p)

But if left=-right (opposite phase) then the crossfade will quickly fade out to silence and back in ;

and if left=right, then your power gain will be (cos(p)+sin(p))² = cos²(p)+2sin(p)cos(p)+sin²(p) = 1+2sin(p)cos(p) = 1+sin(2p). This goes up to double power (amplitude is multiplied by 1,414).

Generally speaking, as long as L and R don't have frequencies in common, this works perfectly, but as soon as they do, you have to do proper phase addition : the cosines are added separately from the sines : if you know a sinusoïd by its amplitude and phase offset, convert to cartesian before adding... this will show you how such things add up.

| Mathieu BOUCHARD ----- téléphone : +1.514.383.3801 ----- Montréal, QC