Soundboard Pre Mix

[Pete Cofrancesco]

The best method I've seen is:
Channel 1: Off camera mic on a mic stand. This will give you ambient and audience applause.
Channel 2: Mic at front of stage sent back to the camera via wireless. This gives you ambient with minimal audience. This is also a good safety source (let say there is a screaming baby at front of your other mic)

Mix in post will give you nice stereo sound.

You can also substitute board feed into channel 2. I let the Sound guy know in advance I need a board feed if the client requests it, I also come early and do a levels test. Despite doing all this, there is a good chance that the board feed will been unusable, so don't rely on it. If the board feed is that important another person has to be put in charge and test need to be done in advance of the performance.

Most clients think all that is involved is running cable from the board to your camera and I used to think the same thing. I now try to temper the clients expectations of relying on a board feed. Its an extra that if it works can enhance the ambient sound.

[Steve House]

Originally Posted by Pete Cofrancesco

The best method I've seen is:
Channel 1: Off camera mic on a mic stand. This will give you ambient and audience applause.
Channel 2: Mic at front of stage sent back to the camera via wireless. This gives you ambient with minimal audience. This is also a good safety source (let say there is a screaming baby at front of your other mic)

Mix in post will give you nice stereo sound.

You can also substitute board feed into channel 2. I let the Sound guy know in advance I need a board feed if the client requests it, I also come early and do a levels test. Despite doing all this, there is a good chance that the board feed will been unusable, so don't rely on it. If the board feed is that important another person has to be put in charge and test need to be done in advance of the performance.

Most clients think all that is involved is running cable from the board to your camera and I used to think the same thing. I now try to temper the clients expectations of relying on a board feed. Its an extra that if it works can enhance the ambient sound.

Mixing those two mics might give you two-channel sound but it won't be stereo sound. The mere fact something is coming from speakers on the left and right doesn't make it stereo.Stereo requires that the mics be arrayed in an arrangement that will recreate the sound stage of a person listening from the optimal "sweet spot" in the audience with sound from one side of the stage coming from one speaker and sound from the other side of the stage coming from the other.

[Pete Cofrancesco]

It gives a fuller sound with depth that is more pleasing to my ear than one mono mic. Whether it meets the definition of stereo is not of great concern to me.

[Greg Miller]

This is very slightly related to a technique that was proposed, tried, and abandoned back in the '50s in the early days of two-track recorders.

I forget who originally proposed this system... seems to me it was someone who had other audio credentials. A well-known name like Avery Fisher, or Henry Kloss, for example. But I don't think it was either of those two. If I research it and find the answer I'll update this post.

The idea was to place one mic front and center in the sweet spot, to record the main musical information; then to place the other mic far back in the hall to pick up mainly reverberant information. These would then be played back on a two-channel system, with one speaker front and center relative to the listener, and one speaker behind the listener. In other words, a mono main channel and a mono "surround" channel.

If you recorded it and played it like that, it was somewhat realistic. Yes, the orchestra was in front of you, but all 40+ musicians were located in one narrow point source. And yes, the hall's reverberation was behind you, but all in one spot, behind the center of your head. Indeed, it was "fuller" than mono, but it proved to be a lot less realistic than left/right stereo... that's why it was quickly abandoned.

(In fact I made some recordings like this as a test at the time. I was recording for AM radio broadcast, so I did my best to get a good mono recording with one front mic... we used that channel only for broadcast and that was fine. I recorded the reverb mic on the other track of the tape, and I played that experimentally at home on my L/R stereo system. If I sat between the speakers it was acceptable, but not nearly as good as real stereo. After all, when you listen to a symphony, you expect the violins on the left and the bass fiddles on the right and everyone else spread in between; hearing a realistic stereophonic image is much more important than hearing reverb coming from a single point behind your head.)

Of course if you listened to one of those front/back recordings over a normal left/right pair of speakers, it would sound as if you were sitting in the middle of the audience, facing directly toward one of the side walls, with the orchestra either to your left or to your right, depending on the channel assignment. That was very unrealistic. Yet it seems as if this is what Mr. Cofrancesco is recommending. Or do I misunderstand you, Mr. Cofrancesco?.

[Eric Olson]

Originally Posted by Greg Miller

Of course if you listened to one of those front/back recordings over a normal left/right pair of speakers, it would sound as if you were sitting in the middle of the audience, facing directly toward one of the side walls, with the orchestra either to your left or to your right, depending on the channel assignment.

I think such a configuration should be mixed as if it were a middle/side setup. The microphone near the stage gets mixed equal between left and right channels and the ambient mike is added to the left channel and subtracted from the right.

[Greg Miller]

Eric, that's a very interesting suggestion!

It would certainly have good mono compatibility, because the ambience mic would disappear in a mono mixdown. No problems there.

I wonder about having all that out-of-phase ambience information, though. In general, I don't think it's wise to have too much info with opposite phase in the L and R stereo channels. But if you mixed it in at a relatively low level, compared with the main front mic, it might work out OK.

In fact, a similar technique is sometimes used to create "artificial stereo" by using a short delay to produce the second track, then adding and subtracting it from the main track (which is mathematically the same as mixing it as M/S).

I'm wishing I had a pair of tracks here to try this right now... I think the result might, in fact, be audibly interesting. Next time I have a chance, I will try to capture a pair of tracks in this configuration.

I'm also wondering whether there's a way to process that ambience track, to get two resulting tracks that have a smaller phase difference between them... say 90º apart, rather than the 180º you'd get by simply using M/S math. Then you make one playback channel = front + unshifted ambience, and the other playback channel = front + 90º shifted ambience. That would lessen the out-of-phase problems, but would still give a nice stereo feel to the ambience. That way, some ambience would remain in the mono mixdown.

I don't recall seeing any constant-phase-shift features in Audition. (I have schematics for an old CBS analog circuit somewhere, I think it was part of the SQ matrix system.) Perhaps I'll have to dig deeper for a software equivalent.

[Richard Crowley]

Originally Posted by Greg Miller

I don't recall seeing any constant-phase-shift features in Audition. (I have schematics for an old CBS analog circuit somewhere, I think it was part of the SQ matrix system.) Perhaps I'll have to dig deeper for a software equivalent.

There is no "constant-phase-shift" except at a fixed frequency. Do the math. It seems horribly retrogressive and devolutionary to be discussing creating artificial "stereo" when we have the means to record the real thing. Artifical stereo was never convincing or even satisfactory except for the most undiscerning hearer.

[Seth Bloombaum]

Originally Posted by Greg Miller

This is very slightly related to a technique that was proposed, tried, and abandoned back in the '50s in the early days of two-track recorders...

In my experience with near-coincident stereo recording techniques for acoustic performance, in a good hall, a mic or array in the sweet spot can be wonderful.

OTOH, many "music" venues don't really have a spot for this kind of technique.

It comes back to the traditional chamber music and concert venues. In the early days of recording, smart and innovative people devised direct-to-stereo techniques that worked really well. Prior to general availability of sound reinforcement, jazz, pop, and folk performers all depended on achieving acoustic balance within their groups. Performance venue acoustics *had* to support this approach.

Today, most pop venues are supporting an entirely different sound, (mostly) meant to replicate close-miced studio recordings. Most venues aren't concerned with room acoustics unless it affects their bottom line.

As someone pointed out above, a sweet-spot recording might be better than some other methods, but usually will not reflect the close-miced aesthetic.

[Eric Olson]

Originally Posted by Richard Crowley

It seems horribly retrogressive and devolutionary to be discussing creating artificial "stereo" when we have the means to record the real thing.

Over half the televisions in use feature monaural audio through a single speaker. In today's world DVD audio is also played through the built-in speakers of a notebook computer, sound bars, and in theater setups consisting of 5 tiny speakers and one monaural woofer. The number of people who employ a well-positioned pair of high-fidelity speakers suitable for music reproduction is surprisingly small.

Sony has handicams that record 5.1 surround sound, but has still not produced a prosumer model with even 4 XLR inputs. If all you have is a monaural stage track and a monaraul ambience track, it seems quite resonable to center the stage track and use the ambience track to create fake stereo. It might even be reasonable to create a 5.1 soundtrack with such sources.

[Greg Miller]

Originally Posted by Richard Crowley

There is no "constant-phase-shift" except at a fixed frequency. Do the math.

I cannot do the math, that's way beyond me. But I did build the circuit (back around 1978) and tested it with a dual trace scope, and it did, in fact, produce a 90º phase shift across the audio band.

I remember this very distinctly. I was trying to build a variable speed drive for a hysteresis synchronous motor. The motor in question was designed, nominally, for 60 Hz. There were two separate windings, one driven directly from the power source (12 VAC in this case) and the other winding driven through a series capacitor, whose value was selected to produce a 90º phase shift... that's that makes the motor turn.

I was trying to figure out some digital way to do this, which would have been pretty complicated in 1978. Then I remembered reading some description of the Columbia SQ quadraphonic system, which encoded four audio channels onto two phonograph channels, using phase relationship to "steer" the playback signal. The system ran one of the channels (rear???) through an analog circuit that produced a constant 90º phase shift, before recording onto the LP groove.

I wrote to someone at Columbia, whose name I had found in the relevant magazine article. Luckily I got a reply from the right person. (Perhaps the fact that I had a prestigious-sounding government job helped with that.) After some discussion, he mailed me a set of the schematics, I built the thing, and it worked as described. I could feed in the signal from an audio oscillator, view the two outputs on a dual-trace scope, and while I swept the oscillator across the audio band the phase shift did -- miraculously, to my mind -- remain very close to 90º.

I went on to build a few prototype circuits. Incidentally, we fed the output of the phase-shifter into the two channels of a Crown D-75 power amp, and used the amp output to drive the two motor windings (directly, without the phase-shift capacitor, of course). The thing worked as designed. But by then the "higher-ups" had changed their mind about how the show would go, and we didn't need any of this stuff any more. Shortly thereafter I decided I'd rather work elsewhere. The last I knew, there were a dozen or more unused D-75s just gathering dust... your tax dollars at work.

If I can find any further documentation about this, I'll post it. I have some files that old stored in the attic, but no guarantee that I still have this particular info.

[Richard Crowley]

Originally Posted by Greg Miller

I cannot do the math, that's way beyond me...

If you could, you would realize that it is impossible unless you have invented time-shifting as well.

[Eric Olson]

Originally Posted by Richard Crowley

If you could, you would realize that it is impossible unless you have invented time-shifting as well.

Quadraphonic sound - Wikipedia, the free encyclopedia

[Richard Crowley]

Yes, and perhaps the primary reason quad sound never really made it off the ground was because of the quite poor performance of the "QUASI" phase shift scheme.

You could also have cited analog color television (both NTSC and PAL/SECAM) as popular communication channels that depend on phase-shift. But note very carefully that in THOSE cases, they were dealing with a FIXED frequency (3.579545 MHz for NTSC, and 4.43361875 MHz for PAL) which made phase-shift trivial.

And modern high-capacity digital communication (like ATSC and QAM) are also completely dependent on sophisticated phase modulation and demod. But AGAIN they are dealing with FIXED frequencies.

The people who developed SQ, et.al. knew that it was impossible, but they kludged a scheme the best that could be achieved in the real world. Alas, it was nowhere good enough even for casual consumer use. The same could be said for artificial stereo by the phase-shift method. Certainly there are better ways of simulating "stereo" from a monaural source here in the digital era. But you can't do broad-band "phase shift" with digital technology, either.

[Greg Miller]

Quadraphonic sound never made it off the ground for several reasons. For example, I suspect most non-audiophile "normal" folks didn't want to instantly double the cost of their audio system, by adding two more speakers and two more channels of amplification.

There were two quadraphonic-LP systems. The JVC system used a very high frequency subcarrier on the disc, which required a special stylus and preamp to play back. It was not reliable even with new discs, and of course repeated playings decreased the amplitude of the subcarrier signal, so that it became less and less reliable as the discs aged.

The Columbia-SQ system also wasn't workable, simply because music already contains a lot of random phase information. So simply using phase information to move sound between the two front speakers and two rear speakers really didn't work. There was a lot of "crosstalk" which was basically random in nature. yes, you got sound out of all four speakers, but it did not accurately reproduce the sound field.

Interestingly, years later, Dolby used a similar phase-encoding system with their analog sound tracks on 35mm release prints. Stereo music tracks were recorded, basically unchanged, to the left and right tracks on the film. Mono dialog was recorded equally on left and right tracks. Surround information was recorded out of phase on the left and right film tracks. The reason this worked (and Columbia-SQ did not) was that by the time Dolby was doing it, advanced DSP was available. The Dolby system performed a sophisticated phase-relationship analysis of the information on the two film tracks. If the system decided that the information was mostly in phase and equal, it "steered" it to the center stage speaker (by means of adjusting playback gains). If the system decided the information was mostly not correlated, it was "steered" to the left and right stage speakers. If the information was mostly out of phase, it was "steered" to the rear surround channel. I've listened to analog optical tracks played back over this system, and the steering worked quite well. (Columbia-SQ probably would have worked better than it did, if sophisticated DSP had done the steering. But maybe not. SQ was trying to reproduce four channels of continuous information: two "stage" speakers plus two "surround" speakers with reverberant audio at all times. That's a lot harder problem to solve than Dolby theatre sound, where you rarely have stereo music, mono dialog, and mono surround all at once. The typical motion picture track would be easier to "steer" than continuous quadraphonic music, like the SQ problem.)

Be that as it may, regardless of consumer acceptance of these flakey quadraphonic systems, and regardless of whether you can comprehend it and believe it, I assure you that the 90º phase shift filter did work the way it was supposed to. I confess that, when I first heard of it, my reaction was the same as yours: "that's not mathematically possible." What that really meant was that I didn't know enough math to comprehend how it worked. But after talking with some people who know a lot more about filters and math than I do, I was convinced to try it. And they were right... it did work. I believe it's related to an "allpass" filter, but it may be more complex than that. I'll try to find some more convincing information when I have time.

[Greg Miller]

Minor update:

I've just found a lengthy discussion of SQ-quad which includes this relevant sentence: "CBS encoders and prototype consumer decoders used precision aligned 10-Pole phase shift networks that were accurate +1° over a 20-20kHz bandwidth." The phase shift was, indeed, 90º.

It's getting rather far OT but for anyone interested in some history, the thread containing the discussion is here:
New Technology SQ Decoder discussion