Crosstalk and image.

[Mooly]

Stability of image is all important in a good system...

Anyone any thoughts on left and right crosstalk being equal... forgetting absolute figures, if the left channel say shows worsening crosstalk as frequency rises compared to the right then how much effect do you think that would have... would the image appear to move as say a solo instrument played higher up the scale.

I wonder if this was ever considered in the mastering and recording of performances... something we have no control over.

[Key]

Originally Posted by Mooly

would the image appear to move as say a solo instrument played higher up the scale.

No. Regardless of frequency the entire stereo image will shift to the side with the most crosstalk.

I wonder if this was ever considered in the mastering and recording of performances... something we have no control over.

Yes all the time. We use pan pots and stereo/matrix micing along with predelay/reverb, eq, dynamic range control to manipulate/optimize these things. This whole process is made to be as intuitive as possible through modern technology like mixing boards and GUIs for the mixing engineer so they don't exactly have to think about what is happening mathematically in terms of measurable crosstalk but it is still considered either way.

[Mooly]

I first came across this problem or issue on a Quad site and it's something I never really considered before but it seems to make a lot of sense...

I can understand why you say the whole image would shift (I think ) because the frequencies we use to locate sound probably start in the few hundred hertz range.
An audio component would have to be pretty poor to start having crosstalk that appeared at those frequencies.
I was thinking more at HF, say above 3 or 4 khz and a figure that worsened assymetrically with frequency.

I know very little about the way recordings are produced etc and the equipment used.

[Key]

I love quad.

Anyway I think for asymmetrical stuff to happen it would be a phase shift in the HF range on one side.

But the way pan pots work is by varying the crosstalk. When the signal is in both channels and in phase you perceive this as phantom center with ideally zero detectable crosstalk between the channels - the L and R speakers should seem like they are no even on sort of but the sound emanates in between these speakers.

Now to move these sounds in a semi circular fashion all you have to do is slowly fade out one side. The sound will shift slowly to the speaker with the most signal.

By messing with the math and how the two channels add together you can end up with different panners which either drop in percieved volume as you come closer to the center or stay somewhat consistent - 0dbcenter, -3dB center, -6dB center. You can actually come up with weird variations of those but those are the 3 most common. On an analog console you usually don't have any control over what kind are hard wired into the board but with digital you can usually switch, mix and match.

[Mooly]

Thanks for explaining it.
This is where I first came across this, about a third of the way down the page,
— QUAD 34 Modification and information —

the figures quoted are pretty poor these days it must be said.