[rec.audio] Stereo synthesis

trimble@laic.UUCP (Gary Trimble) (12/31/89)

I am interested in taking a "monophonic" source and synthesizing a
"stereo" like product from it. I can easily design a circut that
clips high versus low signals but that is really not adequate. Are
there any papers or articles (or thoughts) on this topic available?
Let me know via email and, of course, I'll summarize responses.
Thanks.

wkb@cbnews.ATT.COM (Wm. Keith Brummett) (01/02/90)

    In article <768@laic.UUCP>, trimble@laic.UUCP (Gary Trimble) writes:
    > I am interested in taking a "monophonic" source and synthesizing a
    > "stereo" like product from it. I can easily design a circut that
    > clips high versus low signals but that is really not adequate. Are
    > there any papers or articles (or thoughts) on this topic available?
    > Let me know via email and, of course, I'll summarize responses.

    I don't think you want all the highs coming out of one side and all the
    lows coming out of the other, if that's what you're thinking of.  I
    believe that most stereo synthesizers try to use two overlapping "comb"
    filters to separate the sound.  Consider the response of an equalizer
    that has all of its odd-numbered sliders set to max, and all its even
    sliders set to min -- that's one channel.  The other channel has the
    odds at min and the evens at max.  Of course there won't be as many
    "teeth" in the synthesizer as in an equalizer.  I've seen as few as
    three -- basically the bass and treble from one side, and the mid-range
    from the other.  I would think that you'd want a couple more than that.
    You might want both channels to pass voice frequencies so that vocals
    come from the center.  Also, you could play with the phase of one
    channel, but that wouldn't sit well with the "hi-fi" folks.  I can't
    recall any specific articles on this; it's off the top of my head.  Hope
    it helps.  Good luck.

-- 
    | W. Keith Brummett          (614) 860-3187          AT&T, Room 2B-252 |
    | att!cbnews!wkb               Cornet 353          6200 East Broad St. |
    | or, wkb@cbnews.ATT.COM                       Columbus, OH 43213-1550 |
    `----------------------------------------------------------------------'

bickford@topo.UUCP (Peter Bickford) (01/09/90)

In article <768@laic.UUCP> trimble@laic.UUCP (Gary Trimble) writes:
>I am interested in taking a "monophonic" source and synthesizing a
>"stereo" like product from it. I can easily design a circut that
>clips high versus low signals but that is really not adequate. Are
>there any papers or articles (or thoughts) on this topic available?

One approach involves filtering out the negative frequencies, thereby
producing a complex (stereo) signal from the original real (mono) signal.

(I think the complex signal is called "the analytic signal" associated
with the given real signal.  (see Papoulis, perhaps.))

An ideal filter would need the signal values for all time.  In practice,
an approximation would permit the use of a FIR filter implementation.

I don't know if this has ever been done in the audio world.
-- 
Peter Bickford
UUCP:		...!sun!topo!bickford