[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference napalm::commusic_v1

Title:* * Computer Music, MIDI, and Related Topics * *
Notice:Conference has been write-locked. Use new version.
Moderator:DYPSS1::SCHAFER
Created:Thu Feb 20 1986
Last Modified:Mon Aug 29 1994
Last Successful Update:Fri Jun 06 1997
Number of topics:2852
Total number of notes:33157

1929.0. "Soundstage in Recording Electronic Instruments" by FLOWER::JASNIEWSKI () Tue Mar 07 1989 18:29

                                                       
    	I'm interested in hearing what others have done to effect a
    good soundstage in their recordings of electronic instruments (signal
    cable only - no mikes) using a mixing board. We'll assume that the 
    mixing board has the so called "pan pots" which can vary signal 
    amplitude into each of the two channels, Left and Right.
    
    	I'd like to put forth some soundstage definitions, considering
    the L to R space available in the stereo format. While I can agree
    that there is a front to back dimension to soundstage, I have yet
    to formulate how you would set up discernable soundstage "depth" in 
    a studio when recording a monaural signal from an electronic instrument.
    
    	Of course, there is the monaural soundstage position, which
    is dead center in the field. It is assumed that the signal in the
    left channel is of the same relative phase as the signal in the
    right channel.
    
    	You can "pan" the signal halfway between dead center and one
    of the two stereo channels, giving two more positions on either
    side of dead center, for an instrument location. Again, relative phase
    is still the same.
    
    	You can "pan" the signal entirely to one channel or the other,
    giving two more positions for instrument placement. This makes a
    total of 5 "spots".
    
    	I believe you can get two more spots, which are located beyond
    the extremes of left and right, using the same technique as the
    "stereo wide" features on some EQs and boom boxes. My guess is to
    pan the signal all the way in to the desired side, then phase invert 
    the signal, and apply a proportion of it to the opposite speaker.
    This will give you two more spots, beyond the extremes of left and
    right.
    
    	The soundstage positions then available would be:
    
    L-"0.5"R     L     L+"0.5"R     L+R     R+"0.5"L     R     R-"0.5"L 
                                           
       1	 2 	   3	     4	       5	 6	  7
    
    	
    	Note most mixer boards are only setup to handle 5 of the 7 
    possibilities. To phase invert one channel and pan that to the opposite
    speaker, you'll have to use up a whole 'nother input! That is, if my guess
    is correct! Comments?
    
    	Joe Jas
    
    

T.RTitleUserPersonal
Name
DateLines
1929.1Delays and Fake StereoAQUA::ROSTShe's looking better every beerTue Mar 07 1989 18:5631
    
    All I would like to say is that the "extreme" positions , that is #1,
    2, 6, 7 in your drawing are not too useful unless you use stereoizing
    techniques like the ones Karl mentions in this "Low Tech Recording
    Primer" note.
    
    The way you do that of course is to pan a dry signal hard to one
    side, say left, and a reverbed or delayed version of that signal
    somewhere other than hard left.  Depending on where the reverberant
    signal appears and how loud it is affects the listener's perception
    of where the dry sound actually sits.   
    
    Delays in general are great for creating depth.  Try this, set up
    your delay so you get basically a "slap" echo, that is one repeat,
    then adjust the delay time until you can no longer perceive the
    delay as an echo.  You will eventually hit a delay time of about
    5-10 mS where the doubling fattens the original signal.  Now if
    you pan the dry and delayed signals just slightly apart from each
    other, you get a very wide, deep sound.  
    
    As a matter of fact, I just did a session last night where the engineer
    got a great guitar sound by running the guitar through two amps,
    a Fender Twin (very clean) and a Fender Deluxe (pretty dirty). 
    He set up the reverbs on the amp to be identical, then miked both
    amps, panned one to the left, one to the right.  The result was
    a huge guitar sound dead center which had a sweet overdrive with
    a very tight bottom and none of the nasality of a single overdriven
    amp.  To tweak the reverbs, he had the guitarist play clipped notes
    and adjusted the reverb controls until the reverb "slap" sounded
    like it was dead center, but because of the different tonal
    characteristics, it provided a nice stereo illusion.
1929.2WEFXEM::COTERead my lips, NO NEW AXES!!Tue Mar 07 1989 18:583
    Wouldn't your 2 outside positions be "L-R" and "R-L"?
    
    Edd
1929.3That came nextFLOWER::JASNIEWSKIWed Mar 08 1989 11:2748
    
    	Re .1 - that was my *next* step, to put in the echo's and
    reverberations placed optimally to enhance the 7 positions. The
    two outside positions are formed basically by a phase cancellation
    of the panned signal from the opposite speaker - what I assume the
    "stereo wide" controls do.
    
    	Re .2 - the "L-R" and R-L" that your thinking of represents
    a different thing than my letters do. Mine are panning positions,
    with some attenuation value and a phase sign. Your thinking of the
    final two signals coming from the board subtracted from one another.
    
    
    	I wished to set up a "minimal" discernable soundstage, leaving
    everything "dry", i.e. no reverbs, no echos, no phase shifts other
    than 180 deg. or simply "inverted". I would hope to be able to pick
    out the location of 7 different sounds, created by 7 monaural tracks 
    playing back simultaneously thru a console, panned as described in .0.
    
    	Adding delays will certainly enhance the sense of spaciousness,
    and there is no reason why the delayed sound cannot be panned to
    any one of the 7 positions. Adding some sense to the location choices
    - why would you want the dead center vocal's reverb to be coming
    from the hard left field only? - would make the soundstage a pleasant
    experience to listen to.
    
    	Personally, I think symmetry is important in a good soundstage
    presentation. Some drum mixes are particularly bad - snare over
    here, hi hat over there - at least the bass drum comes from the
    center - Gee, am I "playing the set myself" or listening to the whole
    band from the audience? 
    
    	I also feel that reverberations should be located symmetrically
    in the soundstage field. In fact, I'd consider each signal as
    potentially having it's own reverb, echo or delay of somekind, which
    is also located specifically in the soundstage. Of course, this
    means two channel per input, and lot of expensive delays, say, $100
    a channel.
    
    	However, this can be simplified to just three delays, I believe.
    One, for that which is panned dead center, another for that which
    is panned "left", and another for that which is panned "right".
    The outputs of these can be panned to any position in the soundstage.
    A judicious choice of such would of course enhance the hopefully
    already apparent sound locations.
                                     
    	Joe Jas
    
1929.4QuibblingDRUMS::FEHSKENSWed Mar 08 1989 11:475
    re .3 re .2 - they amount to the same thing.  Adding two out of
    phase signals i sthe same as subtracting one from the other.
    
    len.
    
1929.5Try telling the wife you need 10 reverbsANT::JANZENMr. MSI ECL TestWed Mar 08 1989 13:0210
    the only meaning for "depth" is the opposite of "presence" which
    is determined by 
    1. loudness
    2. the abscense of atmospheric filtering
    3. the presence of liveness, i.e., reverberation from the room.
    
    so depth comes from less loudness, more filtering, and more reverb
    on the signal.  Yes, different sources should get different amounts
    of reverb.
    Tom
1929.6How to do approx. first reflectionsCTHULU::YERAZUNISKwisach Haderach for HireWed Mar 08 1989 13:2867
    Adding a time-delayed version of the sound panned hard to one side
    to the opposite channel is an excellent way of generating a stereo
    illusion. 
                            
    It can help for the recording engineer to draw (on graph paper) a
    "virtual room" for the musicians to be playing in.   Each sound source
    location is marked on the graph paper, as well as the location of
    the listener, and several lines drawn in.
    
    	line 1:  Direct from the instrument to the listener
    
    	line 2:  From the instrument to the nearest wall
    
    	line 3:  From the wall to the listener
    
    Lines 2 and 3 should meet the wall at approximately the same angle.
    
    Measure the length of lines 1, 2 and 3 on your drawing.  Then, set
    the time delay _for_ _that_ _sound_ _source_ according to this formula: 
    
    	millisec delay time =     (line 2 length in scale feet) 
    				+ (line 3 length in scale feet) 
    				- (line 1 length in scale feet)
    
    This assumes the velocity of sound in air is 1000 ft/sec.  That's a
    little high but it's close enough for government work.  You can add
    reverb after the delay, but not before it. 
    
    -----
    
    Example: I've got a "horn" SGU that's off to the right.  Here's
    my graph paper drawing of the room:
    
    
    	-------------------------
    	|	   / \ 		|	
    	|	  /   \		|				  
    	|	 H     \	|				  
    	|		\	|				  
    	|		 \	|			      	  
    	|		  \	|				  
    	|		   L	|				  
    	|		    	|				  
    	|			|				  
    	|			|                                
    	|			|				  
    	|			|				  
    	|			|				
    	------------------------|
    			
    
             
    On my scale drawing (say, one inch = 3 feet of reality):
    
    		Line 1	= 6" --> 18 'feet'
    
    		Line 2	= 4" --> 12 'feet'
    		
    		Line 3 	= 8" --> 24 'feet'
                  
             
    The difference in (2) + (3) - (1) in feet is 18, so the formula says we
    should use 18 milliseconds delay before there is any reverb sound. 
    
    (this is sometimes called a "first reflection" calculation)
    
    	-Bill
1929.7Don't need four walls, either!CTHULU::YERAZUNISKwisach Haderach for HireWed Mar 08 1989 13:318
    I should add that you don't have to have a completely enclosed
    room to do the above calculation.  For example, try the calculation
    in an "open shell" bandstand with two or three sides and a high
    roof.  
    
    It makes the sound very "open-air outside" sounding.
    
    	-Bill
1929.8SALSA::MOELLERAudio/Video/MIDIophileWed Mar 08 1989 14:2916
    < Note 1929.3 by FLOWER::JASNIEWSKI >
>The two outside positions are formed basically by a phase cancellation
>    of the panned signal from the opposite speaker - what I assume the
>        "stereo wide" controls do.

    Disagree.  You can't get any 'wider' than hard left or right panning.
    "Stereo Wide" circuitry, at least in boomboxes, takes off-axis sounds
    from each channel, amplifies them slightly, and pans them hard left 
    or right, depending on which side of stereo they originated from.
    
    Good discussion so far.  In a sense, Joe's right, symmetrical
    reflections more closely recreate an acoustical soundstage.  But
    it sure is fun creating electronic listening environments that never
    occur in nature..
    
    karl
1929.9Yeah, But What's The Sound of Speed?DRUMS::FEHSKENSWed Mar 08 1989 16:279
    Last time I looked, the speed of sound in air at room temperature
    was closer to 1100 feet per second than 1000, so 1000 ft/sec is
    a little low rather than a little high.
    
    But then again, my memory's known to be faulty, and I don't have
    my American Institute of Physics Handbook handy.
    
    len.
     
1929.10tANT::JANZENMr. MSI ECL TestWed Mar 08 1989 17:414
    I think Len's right, although, I just sent my brother back his
    Chemical Rubber Company Handbook of Chemistry and Physics, just
    to help reduce my book load on the next move.
    om
1929.11Observation...WEFXEM::COTERead my lips, NO NEW AXES!!Wed Mar 08 1989 19:3637
    RE: Stereo-wide and "you can't get further than HL or HR"...
    
    Au contraire!!!
    
    I have a Sound Concepts IR2100 "sonic holograph" generator on my
    home stereo. There is no doubt in my mind that it can cause sounds
    to *seem* as if they are coming from someplace outside the stage
    defined by the speaker placement. 
    
    My favorite demo is "Birthday" by The Beatles. Without exception,
    everyone who listens thinks "nothing special" during the instrumental
    opening, but they invariably look to the same 2 empty spots on the
    wall when the vocals start. Why? Because that's where they hear
    the vocals coming from!! (These 2 spots are located about 70 degrees
    off center from the 'sweet spot' while my speakers are form an equi-
    lateral triangle with the listening chair/sweet spot.
    
    I agree with Joe, it's there. (Whether it's "hi-fi" is, of course,
    another argument...)
    
                         *   *   *
    
    Front to back placement in the soundstage is something I've been
    playing with recently and my observations have been counter to
    what I had imagined. If I want to bring an instrument forward I
    raise the level *increase* the reverb. (Pay attention to low-end
    mud when increasing reverb...)
    
    As I think about it, it makes some sense. If an instrument moves
    away from a back wall (towards me) the distinction between the
    original and the reflected waves becomes greater. If the instrument
    were against the wall, the reflected waves would bounce almost
    immediately and arrive virtually instantaneously with the source.
    
    Of course, there's more than one wall in most halls...
    
    Edd
1929.12ANT::JANZENMr. MSI ECL TestWed Mar 08 1989 19:525
    My idea was that if you are miked ORTF or crossed cardioids or a
    single stereo mike in a live hall, than more remote acoustic sources
    have more room in them, by ratio of amplitude of reflections to
    direct sound.  I you mike close, everyone can sound the same, muddy.
    Tom
1929.13DFLAT::DICKSONOne box, one bowl, one spoonWed Mar 08 1989 20:1111
To bring a sound "closer" or "forward" you would increase the volume, increase
the reverb TIME (farther from the back wall means longer round trip, assuming
that it is the *back* wall where most of the reverb is coming from. If most
reverb is from the ceiling and side walls, do not touch the time.) and decrease
the reverb LEVEL with respect to the dry level.  The reverbed signal off the
back wall travels farther than before, and the dry is travelling less than
before.  Inverse square law makes the reverbed path much quieter, the direct
path much louder.

[My best memory of the speed of sound is 1090 ft/sec at standard temperature
and pressure.]
1929.14Specify the circuit, pleaseFLOWER::JASNIEWSKIThu Mar 09 1989 11:2317
    
    	Well, as far as the "Stereo Wide" feature on some boom boxes
    and some graphic eq's, I'm pretty sure that the effect is done
    relatively *cheaply*; there's no tuned circuits, no delay lines,
    no phase shifting networks, it's simple. Like a couple of inverting
    opamps, and a resistor divider. Anyone who can specify the circuit
    exactly gets a cookie! I'm searching AUDIO right now for title words
    like "wide" and "sonic"...so far...Great - I get a couple of
    "panasonic" notes...I'll try holography. 
    
    	I appreciate the thoughts, however, especially the ones on the
    F/B dimensional aspect. Make that 4 delay lines, the additional
    one to accomodate the F/B dimensional difference in perception.
                                            
    	Joe Jas
    
    	
1929.15How National Semiconductor Did ItAQUA::ROSTShe's looking better every beerThu Mar 09 1989 12:0437
    
    I found a stereo enhancement circuit in the 1984 National Semiconductor
    Linear Supplement data book.  The part is LM1040, a "Dual DC Operated
    Tone/Volume/Balance Circuit with Stereo Enhancement Facility". 
    It looks like it is for low-cost stereo applications (like boom
    boxes, eh?).
    
    The circuit is shown and consists of two transistors, three resistors
    and a cap.
    
    I can send a copy of it to anyone interested.  Anyway, here's the
    text:
    
    "When stereo system speakers need to be closer than optimum because of
    equipment/cabinet limitations, an improved stereo effct can be obtained
    using a modest amount of phase-reversed interchannel cross-coupling.
    In the LM1040 the input stage transistor emitters are brought out to
    facilitate this.  The arrangement is shown below in basic form
    (schematic here).  With a monophonic source, the emitters have the same
    signal and the resisitor and capacitor connected between them have no
    effect.  Wiuth a stereo signal each transistor works in the grounded
    base mode for stereo components, generating an in-phase signal from the
    opposite channel.  As the normal signals are inverted at this point,
    the appropriate phase-reversed cross-coupling is achieved.  An
    effective level of coupling of 60% can be obtained using 4.7K in
    conjunction with the internal 6.5K emitter resistors. At low
    frequencies, speakers become less directional and it becomes desirable
    to reduce the enhancement effect.  With a 0.1uF coupling capacitor, as
    shown, roll-off occurs below 330 Hz.  The coupling components may be
    varied for alternative responses."
    
    What this suggests is that while this circuit could be used to "widen"
    stereo mixes, it does nothing whatsoever for mono signals, therfore
    using it to widen the soundstage to the positions #1 and 7 in the
    original drawing would not be possible.  However, by using this
    circuit, the stuff appearing at positions #2 and 6 would move to
    #1 and 7, respectively. 
1929.16I took mine apart, it's in there.CTHULU::YERAZUNISHaven't I met you before?Thu Mar 09 1989 14:4415
1929.17DFLAT::DICKSONOne box, one bowl, one spoonThu Mar 09 1989 16:173
Phased array antenna techniques are very sensitive to frequency.  It would
be just about impossible to synthesize an apparant position outside the
speakers for a broad-spectrum sound.
1929.18so it's a fake out of sortsNAC::SCHUCHARDLife + Times of Wurlow Tondings IIIThu Mar 09 1989 19:1415
    
    hum - the cheapo rat-shack eq that someone donated to me has this
    stereo expander thingie - too the ears, it just pans the seperation
    harder plus (probably cause it's cheap)  remove lot's of high end
    that force me to use the sliders to put back in. So far, i have
    only played with it and not saved the results.
    
    It would do the same thing to birthday that Edd describes. According
    to the fellow who donated it, it is suppose to add a little wetness
    in the process. As i mentioned before, it does something to the
    signal, but it is not something i'm terribly fond of either! Good
    discussion - once the physics are done, perhaps someone can explain
    what they feel the effect should be like to hear?
    
    bs
1929.19It's All In Your MindAQUA::ROSTShe's looking better every beerFri Mar 10 1989 11:5129
    
    The point of what I said in .15 was that while the stereo expanders
    will expand a *stereo* image, you can't just plug in a mono signal
    into such a device and "expand" it beyond hard left and hard right.
    
    If you place a signal hard left and then add an out of phase version
    of it hard right, it will simply lower the apparent volume due to
    phase cancellation, and if anything, move the sound towards the
    center!  If the gains are equal it will almost perfectly cancel
    and you will get just about zero volume.
    
    Stereo enhancement, works though, because in general things are
    *never* panned truly into just one channel, there's almost always
    signal in both channels.  This is true even of old stereo records
    like early Beatle albums.  Compare the separation of those with
    two mono tracks from a multi-track deck while listening through
    headphones to see what I mean.  They *are* severely ping-ponged,
    but not as hard as they could be.
    
    Therefore, when you cancel some of the right hand sound from the
    left, the left stuff sounds "more left" for lack of a better word.
    Also, psychoacoustics enters into it a bit.  Anyone who has ever
    used a quad setup will understand something of this, there are a
    lot of ways to fool the ear into thinking it's hearing something.
    Ever listen to binaural recordings through headphones?  There is
    quite distinct front-rear imaging, and sounds even sound as if they
    are coming from behind you!  Play the same thing over a set of speakers
    and it sounds practically mono.
    
1929.20More thoughts.FLOWER::JASNIEWSKIFri Mar 10 1989 11:5453
                                                                      
    	Besides my interest in creating a good soundstage when mixing
    "straight" electronic signals with a board, I'm also very interested
    in why my "DDD" disks *all* have this remarkable soundstage, with
    sounds *definately* coming from my "1 & 7" positions...
    
    	I assume and guess that it's part of the features available
    to the recording process that's afforded by Digital mixing. Having
    never seen a Digital mixing desk, I dont really know. But from what
    I do know about Digital, I'd expect that one could "copy" any signal
    at any relative phase with any attenuation from any input channel
    and place that anywhere in the final L / R soundstage via this process.
    
    	But, I only have an analog board. I wish to do as well as I
    can with what I have. Sometimes, I dont use all the channels, so
    I could make use of the unused few perhaps to effect soundstage
    positions.
    
    	The comment on the position 1 & 7 effect as being wrought in
    a similar manner as a phased array antenna is right on the idea,
    as far as the basic "how it works" goes.
    
    	I appreciate the reference given for the circuit intended for
    boom boxes. The low frequency cutoff is something I didnt consider,
    I spose you could just turn down the bass eq on the channel you
    bring the inverted signal into. At least I *know* now that it is wrought
    by a "cross coupled inversion with attenuation" process.
    
    	I've found that in my experience one can record in stereo and
    intend for different playback scenarios. For example, I'd do it
    differently if I knew I was recording for playback on headphones.
    I spose you could extend this idea to boom boxes, and make everything
    idea for playback on one of those.
    
    	Spatially, I find miking a drum set using the so called "far
    field" technique produces a much more realistic soundstage than
    does the "individually miked and panned" technique, which can put
    the sounds into unrealistic positions. Of course, to keep the drummer
    happy with his sound, I find that I *must* mike up the Bass drum
    and the Snare, mixing these in via pan pots. I EQ the piss outta
    those signals, to keep their monophonic pickup of the other sounds
    from mucking up the soundstage as created by the stereo "far field"
    mikes.                            
    
    	Matching this sense of spaciousness with an artificial soundstage
    created from a careful choice of where to put the electronic signals
    is what I want to be able to do. Of course, when using something
    like a Roland TR707 for the drum sounds, there is no sense of
    speaciousness to begin with. I spose you could hook individual
    amplifiers to each discreet output, and set those up in the drum
    room...but the idea here is to get to sound real...electronically!
    
    	Joe Jas
1929.21Down to an Experiment!FLOWER::JASNIEWSKIFri Mar 10 1989 12:1313
    
    	I disagree with .19! I guess it will take an *experiment* to
    find out what actually does happen. One way would be to record
    two sounds on a tape using both channls; one sound panned "mono",
    the other panned hard left. Play this back on the boom box with
    the Stereo Wide feature, and listen to the effect of the "Wide"
    switch, if any, on the sound panned left. .19 sez there will be
    no effect, because it's a monaural signal that happens to be sitting
    in the left channel. The mono recorded signal in this experiment
    gives the spacial referance point, so that any location change can
    be determined.
    
    	Joe Jas
1929.22Out of Phase beatles?DRUMS::FEHSKENSFri Mar 10 1989 13:1412
    re .19 - the early Beatles albums were recorded in mono.  The US
    versions were either "fake" stereo (by using differing EQ on the
    left and right channels) or simply panned the master's tracks hard
    left or right.  They were never intended to be mixed down to stereo.
    
    Also, out of phase mono signals sent to stereo speakers do not cancel
    uniformly.  The effect is strongly dependent on the listener's
    position, and is usually described as "diffuse or unstable
    localization".
    
    len.
    
1929.23AQUA::ROSTShe's looking better every beerFri Mar 10 1989 13:4316
    
    Re: .21 and .22
    
    Two last comments and I'll drop out of here.

    1. Phasing...yes, Len you're right about the effect of listener
    location, which is why I used wording indicating it would not be
    100% when discussing the volume loss.
    
    2. Experiment..if the *only* thing in the soundstage is a mono signal
    panned hard left, then there is no signal in the right channel to
    be subtracted (given the boom-box circuit).  However if you panned
    it to say your original position #3, I would expect it to move toward
    positions 1 and 2.  I.e. you need *something* in the other channel
    to get the cancellations that give the wide effect.
    
1929.24thxNAC::SCHUCHARDLife + Times of Wurlow Tondings IIIMon Mar 13 1989 20:292
    .19 - thank you
    
1929.25Regards,FLOWER::JASNIEWSKITue Mar 14 1989 11:2922
    
    	All told, these effects are subtle and are things that *most*
    people wouldnt even bother to listen for. They are also very dependant
    on the listening environment; when driving in your car, or listening
    to the boom box whilst painting the house, who cares whether "it
    seems like" the lead guitarist's echo comes from beyond the left
    speaker. You just want to listen to some music...
    
    	I would think, however, that part of the whole idea of recording
    *electronic* music is to effect the senses in an even more fantastic
    way they are in everyday life. But perhaps even that idea is just
    a cliche' with electronic music nowadays, and it's far better to get
    a good "garage band" sound via pure synthesis that it is to be
    deliberately dazzling with it. 
    
    	I rather enjoyed the conversation here and learned a couple
    of things too. I'll look forward to placing a note on another topic
    I *think* I know something about!
    
    	Joe Jas
     
    
1929.26But I wanted it over *there*!DDIF::EIRIKURTue Mar 14 1989 14:369
    While trying this extreme-pan business, I seemed to find a bug in my
    Y-word KM08 mixer.  The effect sends are mono, and when I increase
    the amount of effect send, I seemingly decrease the amount of stereo
    seperation--that is to say that turning up the effect send centered
    the source, overriding the panpot!  I didn't really chase this down,
    as it was late....
    
    	Eirikur
    
1929.27SALSA::MOELLERThis space intentionally Left Bank.Tue Mar 14 1989 17:1017