I'm quite fully booked at the moment so cannot show visualization soon.
It should be well known that distance difference of 1/2 wavelength between two radiators with the same phase causes null, and 90 deg is the most significant in power response due to intensity conversion from dual plane to spherical surface so c-c=1.0 x wavelength would give smallest dip in power with assumptions that XO is phase matched i.e. acoustical L-R and directivities of individual radiators do not drop at XO band. Acoustic LR24 is quite typical XO so that's good starting point for simplified rules.
We also know that weight of power response is not 100% in sound balance. Therefore it's safest to include also vertical early reflections. CTA-2034 standard committee has decided that floor bounce of 20-40 deg down and ceiling bounce of 40-60 deg up are the most relevant assuming that listening distance is 2-3 m i.e. typical at home. Target is to aim nulls elsewhere than early reflection angles to get reflected spectrum smooth. Very near field has another issues, and very long distance changes the game because previous directions do not send 1st order early reflections.
Easiest way at the moment with VituixCAD is to optimize either preference rating (equation 9) or in-room response with three ideal omni radiators in MT or MTM (nothing loaded to Drivers tab). Both optimize blend of power and early reflections. XO LR24 linear phase. Variables to be optimized are Y coordinates of M drivers. Best result i.e. the most linear in-room response or highest PR should happen with c-c somewhere between 1.0 and 1.4 x wavelength. Higher weighting of early reflections increase c-c.
Phase mismatched crossovers such as Butterworth and Bessel will change this rule. In addition, weight of power response is smaller in practice with very directive radiators such as large cones and horns because they are able to block 60-90 deg quite much. Another c-c rules - or no rules at all is needed.
It should be well known that distance difference of 1/2 wavelength between two radiators with the same phase causes null, and 90 deg is the most significant in power response due to intensity conversion from dual plane to spherical surface so c-c=1.0 x wavelength would give smallest dip in power with assumptions that XO is phase matched i.e. acoustical L-R and directivities of individual radiators do not drop at XO band. Acoustic LR24 is quite typical XO so that's good starting point for simplified rules.
We also know that weight of power response is not 100% in sound balance. Therefore it's safest to include also vertical early reflections. CTA-2034 standard committee has decided that floor bounce of 20-40 deg down and ceiling bounce of 40-60 deg up are the most relevant assuming that listening distance is 2-3 m i.e. typical at home. Target is to aim nulls elsewhere than early reflection angles to get reflected spectrum smooth. Very near field has another issues, and very long distance changes the game because previous directions do not send 1st order early reflections.
Easiest way at the moment with VituixCAD is to optimize either preference rating (equation 9) or in-room response with three ideal omni radiators in MT or MTM (nothing loaded to Drivers tab). Both optimize blend of power and early reflections. XO LR24 linear phase. Variables to be optimized are Y coordinates of M drivers. Best result i.e. the most linear in-room response or highest PR should happen with c-c somewhere between 1.0 and 1.4 x wavelength. Higher weighting of early reflections increase c-c.
Phase mismatched crossovers such as Butterworth and Bessel will change this rule. In addition, weight of power response is smaller in practice with very directive radiators such as large cones and horns because they are able to block 60-90 deg quite much. Another c-c rules - or no rules at all is needed.
Comment