quad: smooth diagonal panning?

    Aug 26 2012 | 11:42 pm
    I am working on a project that uses a lot of quadraphonic panning. Easy enough to get smooth circular motion, but I am having difficulty creating smooth motion from say, FrontRight to BackLeft.
    Because this type of motion is key to the piece, I have long since abandoned working with any of the objects out there since they seem to only be capable of circular panning. Instead, I am just writing breakpoint-style curves for each speaker's gain~ slider through the "line" object.
    SO- any tips on how to create smooth diagonal motion? It seems that the other diagonal pair (in the case of my above example, FrontLeft and BackRight) needs to be involved somewhere along the line. I have come close, but it's just not quite there.
    Input is much appreciated.
    Thanks, Tomek

    • Aug 28 2012 | 9:48 am
      have you considered treating the two speakers diagonally opposed as a stereo pair, then just pan between them using equal power, or are you asking for something more sophisticated?
    • Aug 28 2012 | 3:04 pm
      Actually, that did the trick! I got so caught up in writing my lines that I overlooked the fact that I was not working with equal power. Thank you for bringing that to my attention, as I was going mad over here. Now, this works perfectly fine in the 5x5-foot square that I am currently sitting in, but do you imagine that the effect will translate well in a medium-sized recital hall? The effect being that of the sound traveling diagonally through a central point, rather than simply fading from one corner as it appears in the other. Or might it benefit from bringing the signal up in the opposite diagonal pair a bit around the midpoint of the transfer?
      Also- I achieved equal power using cppan~. One thing I noticed is that I get some popping if I pan between the two channels quickly. I have noticed this with other processes, such as switching dry/wet balance in reverb. Is this the nature of the beast, or is there a way to alleviate this?
    • Aug 28 2012 | 5:33 pm
      Panning between the diagonals is a good idea.
      Usually, I do an equal power pan left/right on that coordinate and then pan each channel of that result front to back. I think there's an example patch in the CNMAT Depot: http://cnmat.berkeley.edu/downloads
    • Aug 29 2012 | 1:43 am
      panning between the two corners should be all you need.
      think of the diagonal line as a segment in a circular setup of 4 speakers, where panning between speaker positions works likewise.
      a linear movement from other positions in your quad setup will also work using panning: when you want a source to describe a line from LF+RF (12 o clock) to LR(20%)+RR(80%) (4 o clock) just use regular panning between LF+RF and LR+RR.
      about "bigger rooms", well, all things positioning, weighting or phase only works as intended when the listener stands in the ideal spot. but i think even when you are standing outside the 4 speakers it will still be impressive to hear movements from one side to the other. the user will instead of the direction now recognise phaseshifts or a delay - but in the end that only reproduces what would happen with a moving _natural sound source, too, isnt it?
      if you think it sounds boring you might consider implementing a doppler effect, or put some slight first reflections on the other diagonal, whose dynamic delays times represent the actual (logical) position of the sound source.