Stereo imaging plugin

[bluebell]

What we really want is a delay time in ms or µs. That would be independent of the sample rate. But because it's so simple and works well I use the number of samples.

The plugin displays time too - in my case 0.625 ms.

Delay should be related to the panning percentage.
If we want to be very exact then we should calculate how far away the audio source should appear. If it's meant to be very far away (usually coupled with lots of reverb) then the angle is smaller than if it's right in front of your face and so the difference between your ears is smaller and the delay of one side should be smaller, too. But I don't consider this effect.

Interesting. Let's consider a simplified example: how would you calculate the delays for lets say a string quartet (every instrument would appear on the same radius from your head), plus a double bass which would go into the middle but a bit further away?

Isn't there a plugin available which would the this for us? I mean sound source positioning in a virtual space.

I estimate instead of calculating. The Head-related transfer function is frequency-dependent and complicated.

Something that's centered doesn't need pan/delay.

Something that's not too far away (distance ignored) and panned 100% left should be delayed on the right side as much as the right ear is farther from the source than the left ear, eg. 20 cm, which leads to 0.02/330 = 0,06 ms = 600 µs as a first approach. Try it out an play with the values until it sounds good to your ears.

For more than panning/delaying there is Ambisonics.

[sysrqer]

For stereo imaging you should use much lower delays than 30 ms. For my taste always under 10 ms.

You can get away with up to 30ms, I've done it with hi-hats before, depends on the sound.

[novalix]

Hi,

all those theoretical points are good and valid.

I sense one should keep in mind that with this technique one always alters the phase relationship of the signal and it is up to our ears if that is causing an issue or not. A/B-ing in the mix context should be mandatory.

Sometimes the better outcomes lie in those ranges where the two signals are distinct (but still connected) to our ears.
Recently i've delayed a hard panned electric guitar on the other side about 72ms which sounded the most exiting to my ears and minimized the phasing issues especially in the low end, which was crucial for that particulary performance.