Mono to Stereo non-linear phase splitting technique

[DSPaudio]

Hi All.

I'm curious if any of you wizards might have a go at reverse engineering the mono to stereo spatialization process used in the Bedini B.A.S.E...

Here is a synthmaker page (without source files) that shows examples (and they are stunning in how a mono signal can be 'stereoized'):

http://www.johnbedini.net/john34/test.html

Definitely take a listen to this (first half mono, second half processed with the above synthmaker patch):

http://www.johnbedini.net/john34/Clovers.mp3


Original patent for the hardware:

https://www.google.com/patents/US4555795


Abstract from the patent:

"An audio processor which converts monaural input to a realistic pleasant sounding binaural output in which optical coupling means, having a nonlinear transfer characteristic, creates an unbalanced output to a pair of audio outputs. The system provides for a high or low level input, which may be selected by a switch, to an emitter follower which drives an optical coupling circuit. The optical coupling circuit is comprised of a light emitting diode and a phototransistor connected in a phase splitting network. The phase splitting network provides an unbalanced output to terminals for connecting to preamplifiers, amplifiers, etc. An alternative embodiment employs two optical coupling circuits in two separate signal processing channels, including two phase splitting networks. "


And some quotes and marketing on the Bedini website:

http://www.bedini.com/base.htm


As I am involved in the mastering trade and have lots of familiarity with various techniques to fake stereo or produce holosonic fields (Qsound, Katz's plugins, Mid-Side processing, etc.) This processor's effect (on a bad day when you have only a mono source to work with) seems quite desirable and could be very useful for treating synth patches to.

/a

file under: (faux stereo, fake stereo, mono-to-stereo, mono-stereo, spatial abstraction, stereo field control, stereo abstraction)

[Steveg]

Hi DSPAudio,

I had a look at the patent and they have a circuit diagram of the "Audio Processor" in Figure 3. Q2 is a conventional split-load phase splitter of the type found in many Class AB valve amps just using a transistor rather than a triode. I can't even begin to reason why it is driven as a photo-transistor rather than using any conventional drive method. My guess is that the optocoupler will exhibit soft clipping when over-driven, but that is just a guess. What comes out the other side is just two copies of the input signal 180 degrees out of phase with whatever nonlinear effect the optocoupler has introduced. It does not look like it is going to be hifi.

Cheers,

Steve

[DSPaudio]

Hi Steve,

Thanks for your analysis. The non-linear material dependent process of abstracting tangentially related lightly distorted material and throwing that to the sides seems to be the crux of it.

It seems that non-linear phase distortion with the artifacts only being directed to the sides would allow a mono signal to take on a stereo 'presence'.

Unlike traditional frequency shufflers (200-400hz > Left 20%, 400-600hz > Right 20%, 600-800hz > Left 20%, etc etc.) the BASE is mono compatible and *some* regarded engineers seem to like it well enough (when used judiciously).

As far as it being nigh-fi or hi-fi, this thread explores mastering engineers impressions of the device:

https://www.gearslutz.com/board/mastering-forum/440659-brian-gardner-widening.html

Short summary is, a number of 'names' have used modified versions of the unit over the years on their releases. And like most processors it can also ruin a mix.

I think this Bedini approach is preferable to phase shufflers which can stereoize a mono signal but stress out our psycho-auditory recognition process (can even produce a headache under the wrong circumstances). And Mid-Side processing can't do anything for mono (obviously) and finally haas delays (like k-stereo) can add the illusion of space but typically leave the user feeling like they've traded the source clarity for stereo image.