Did not check the Bacch any closer, but face tracking and all processing could be made quite easily and cheaply with Arduino for example. Questimate person position and adjust delay, perhaps gain of L and R to keep sweetspot at listener. Dont see too much value here though, its still one listener system 
https://create.arduino.cc/projecthub/shubhamsantosh99/face-tracker-using-opencv-and-arduino-55412e
https://create.arduino.cc/projecthub/shubhamsantosh99/face-tracker-using-opencv-and-arduino-55412e
Giving another thought to it, if it was simple like this and worked fine then people would probably use it all the time
So, either its not simple or its just a gimmick. Made quick try with Audacity, just duplicating a stereo track, stereo flip, invert, highpass, tiny delay, tried low pass, vary the delay. It was kind of obvious it sounded poop so could not even concentrate what the effect was on the stereo image
well, perhaps it worked as should but sound is very much fine without. It probably is not so easy. Comb filtering on highs is quite obvious doing it like this.
Earlier in this thread I had linked to a paper by Prof. Choueiri... Sadly it has been taken down from his website.
I still have a copy, it's an older paper from 2015. You're right though, it isn't that easy. If you look at what he's doing you'll notice the (almost) side by side speaker setup to make the processing easier and more effective. Just like they advise over at ambiophonics.org. Prof. Choueiri used to be an active member of that research group, but split off and went commercial with his own stuff... It's available online and a review can be found at ASR.
I have played extensively with his ideas etc. Not without favorable results, except that at times I got listeners fatigue (no head tracking, so holding the head steady in one position gets old real fast). I still use parts of the ambiophonic or Bacch theories as inspiration in my DSP trickery. And yes, it partly involves what you were trying to do, but a little more groundwork would be needed to get an idea of what is needed. Go play with a mic and a dummy head to investigate, or even simpler, adding and subtracting time delayed impulses in REW, pretty sure I've shown quite a few of those in this thread. The time delay of cross talk for an average head is about ~0.27 ms by the way.(*) That should help.
(*) in a 60 degree stereo setup with ~3 m distance between both speakers and from speaker to listener. Adjust from there to get your personal delay.
My own choice of DSP has been sort of a combination of a few things, among which the shuffler, the subject of this thread. I like it better as it reacts more natural to head movement (less or no fatigue) and still has good tonal results outside the sweet spot. Tonal balance is No. 1 for me, but for the sweet spot I do try real hard to get it all. Been experimenting with it for years, I like what it does.
I still have a copy, it's an older paper from 2015. You're right though, it isn't that easy. If you look at what he's doing you'll notice the (almost) side by side speaker setup to make the processing easier and more effective. Just like they advise over at ambiophonics.org. Prof. Choueiri used to be an active member of that research group, but split off and went commercial with his own stuff... It's available online and a review can be found at ASR.
I have played extensively with his ideas etc. Not without favorable results, except that at times I got listeners fatigue (no head tracking, so holding the head steady in one position gets old real fast). I still use parts of the ambiophonic or Bacch theories as inspiration in my DSP trickery. And yes, it partly involves what you were trying to do, but a little more groundwork would be needed to get an idea of what is needed
. Go play with a mic and a dummy head to investigate, or even simpler, adding and subtracting time delayed impulses in REW, pretty sure I've shown quite a few of those in this thread. The time delay of cross talk for an average head is about ~0.27 ms by the way.(*) That should help.(*) in a 60 degree stereo setup with ~3 m distance between both speakers and from speaker to listener. Adjust from there to get your personal delay.
My own choice of DSP has been sort of a combination of a few things, among which the shuffler, the subject of this thread. I like it better as it reacts more natural to head movement (less or no fatigue) and still has good tonal results outside the sweet spot. Tonal balance is No. 1 for me, but for the sweet spot I do try real hard to get it all
. Been experimenting with it for years, I like what it does.
Last edited:
I'd love to, one day. I do have 6 channels of "OK" amplification and 8 channels of "OK" (48 kHz internal processing) miniDSP to drive them, for such experiments. I could also stack the two Zoudio amps I own for enough DSP channels to directly power two main, two SDA, two woofers, but that would take down one of my systems. Or I could spend the money for a 3rd Zoudio amp, or get passive crossover components for the Lii F15 / Goldwood 18 based OB system - with the advantage of being able to simply connect up a "normal" 2-channel amplifier to it, albeit costly for...50 lbs of copper and iron.
Until then, I'm stuck with two identical drivers in series. I suppose I could "cut" the highs to one driver with a parallel capacitor, then reduce the consequential "boost" to the other driver in the Zoudio amp's DSP. Likewise with an inductor for the lows; but that's everything I'd not want to have to juggle about, what with having DSP amps in the first place. There's just not enough output channels to go around...with attempting SDA and mono center.
I just started this thread about a way to set up 3 or 5 speakers for 2 channel stereo playback to solve the center phantom image problem. https://www.diyaudio.com/community/...-for-the-phantom-center-image-problem.393540/
It's just a L-R, L+R, R-L 3 speaker array spaced apart properly and listened to at an appropriate distance. Magic happens! You can add more side speakers to make a 5 or 7 or more channel array but 3 will suffice for excellent results.
It's just a L-R, L+R, R-L 3 speaker array spaced apart properly and listened to at an appropriate distance. Magic happens! You can add more side speakers to make a 5 or 7 or more channel array but 3 will suffice for excellent results.