Then we will have to agree to disagree. I am unwilling to accept "I can't explain the math" or "it's patented" or "it's magic". No, I have no respect for the technical competence of the USPTO. You would be amazed at what they have granted patents for.

Richard, we can agree to disagree. I know the filter I described is workable, and have found a few references to it in places like The Audio Cyclopedia, but I don't understand the supporting math (I checked out of math after first year calculus). But I built one, and tested it, and saw it work as predicted. No offense intended, but I give more weight to my own measurements and observations than I give to anyone's blanket statement that something is "impossible." After all, manned flight was once considered impossible. Just because something seems impossible to you does not make it so.

Having said that, I don't know why you keep trying to obfuscate the issue. First you wanted to use "SQ was never successful" to prove that a 90º phase shift filter is not possible. There is no logical connection between the economic success of the SQ system, and the feasibility of this filter (which was only one very small part of the SQ scheme).

Now you are invoking the USPTO to prove that a 90º phase shift filter is impossible. In the first place, I haven't made any claims, one way or the other, about the filter's being patented. Actually, I doubt that the filter, itself, can be patented, any more than a two-way 12dB/octave crossover can be patented. Ohm's Law isn't patented, but we accept it as workable. Thevenin's Theorem isn't patented but we accept it as workable. OTOH, at one point phonographs, motion pictures, and FM radio were patented and we know that they work, too. A given device can work, whether or not it is patented... there is no logical relationship between the two. So whether this particular filter is patented or not, there is no logical relationship between the patent status and the question of whether such a filter can exist and function as predicted. Your continued obfuscation simply makes me feel that you are arguing from a position that can not be logically supported. IMHO that doesn't strengthen your argument; if anything it weakens your credibility. Although you might make a good politician.

At the present time, this seems to be all we really have to go on: I have built one and have seen it work and believe it is possible; you have not built one and have not seen one work and believe it is not possible.

Wow this subject has gone off in a different direction than the OP must have wanted...

Phase from what I grasp is frequency and there fore time dependant, many years ago in the early days of stereo TV broadcasting we ran a "Howe phase chaser" in the control rooms when we were producing music clip shows, they corrected any head alignment errors that occurred in the VTR machines.
They were slow acting but worked VERY well for fixing small alignment errors but caused problems when used when with MS mics as there often sensed the side components as an error and tried to correct it, with some disastrous on-air audio results. (And hence my personal dislike of MS mics for broadcasting)
Here is a link to the patent for the Phase Chaser. www.google.com/patents/US4890065.pdf

The role of a patent office is NOT to determine the validity or practicality of an idea but purely a place to lodge an idea and to have it documented.

But I have to agree it would be difficult to produce a 90 deg phase shift on ALL frequencies of 20Hz -20k Hz

Brian,

Yes, we have gone off on a tangent. When I first briefly mentioned a 90º constant phase shift, I never could have imagined that someone would argue so vehemently and persistently that such a thing was "an impossibility," and would continue to do so when there is so much literature available online supporting their existence.

You're certainly correct in saying that phase is frequency- and time-dependant... no question about that. That much is intuitively obvious to me.

Yet there are a lot of complex engineering principles that are not intuitively obvious, which certainly are true. For example, I don't understand anything about thermodynamics. I can't prove mathematically that an internal combustion engine works; yet I've driven my car, so I know that it does work. I am certainly not a nuclear scientist and couldn't design an atomic bomb or atomic power plant; yet there is enough evidence to convince me that both of these things exist. Likewise, I can't begin to explain how a constant phase-shift filter works; yet I've built and tested one and I know that such a thing is possible. My point is simply that, although a given person doesn't understand the theory behind some device, it doesn't mean that device can't possibly exist. (Heck, I can't explain mathematically how we are here on the earth, orbiting the sun, which is in the milky way... etc. Yet obviously we are here, despite my limited mathematical comprehension.)

Since this discussion began, I've been reading many articles about constant phase-shift circuits. They are not intuitively obvious to me. I suppose if I had stuck with calculus for a few more years, and had taken some filter design courses, then I might begin to understand how the things actually work. I will readily admit that I don't have that deep of an engineering background. Nevertheless, there are plenty of relevant articles online. These filters seem to find a lot of uses; the most recent one I've read about was demodulating SSB (single sideband) radio transmissions. They are out there.

(I decided, at age 18, that I wanted to have a "hands-on" career. I did not want to spend my life playing with calculus and mathematical theory. But my hat's off to the people who do that for a living, and who are smart enough to design filters with this degree of complexity... filters that I really can't comprehend.)

Yes, Brian, you are clearly correct in saying that "it is difficult to produce a 90 deg phase shift on ALL frequencies of 20Hz - 20kHz." It is difficult, but not impossible. The detailed article I read about the SQ system (of which the phase shift is only one part) says that the encoders, and prototype decoders, needed 10-pole filters, with unusually precise components, to produce a phase shift which was accurate to 1º over that frequency range. A 10-pole filter is a very complex circuit, compared to most typical audio filters, crossover networks, and filters that I've worked with. Indeed, it is difficult.

I'll freely admit that I wouldn't have even thought of such a thing, if I hadn't read the article about SQ, and I certainly couldn't have built one if someone hadn't given me a detailed schematic. Yet here we are on the earth, orbiting the sun, flying through space... with cars, nuclear power, and constant phase-shift filters.

Thanks for your interest and input.

Desperation audio is a pretty strong term for something that's more reliable than the alternatives, but I would agree this is more for documentary type work than music quality.

Nothing can kill a documentary, or any other type of film, faster than crappy, amateurish sound. Just because the equipment availability has been democratized doesn't mean that the standards the finished product should adhere to have come down - if anything, the fact that everyone today has grown up listening to "that Hollywood Sound" means the standards are higher today that they ever were in previous generations. I don't care if you're recorder costs $500, $5,000, or $25,000, your finished sound track must adhere to the highest level of quality that is possible to deliver or you're not doing your job as a filmmaker properly. The only time it is acceptable to compromise on quality is when it's a situation so desperate that it becomes a choice of settling for second best or coming home with nothing at all.

Rather than micing a speaker, you need to bite the bullet and do whatever it takes to mic the stage properly for your recording mix. Don't have a second mixer? Buy one. Don't have someone to operate it? Hire one. Can't afford to do it right:? Don't do it at all and spare the world from the presence of yet another POC production.

Don't have a second mixer? Buy one. Don't have someone to operate it? Hire one. Can't afford to do it right:? Don't do it at all and spare the world from the presence of yet another POC production.

There is nothing done so well that it couldn't be done better. Not having an additional sound person for recording the band does not prevent most videographers from covering a wedding reception. The goal is to get the best sound possible given the constrains imposed on the shoot.

The original post considers what a single shooter using a single video camera can do to improve sound without going dual system or bringing so much equipment that it takes a crew to move it. Having said this, an inexpensive improvement could be made using reasonable recording techniques along with something like a $500 Zoom R24 which can record 8 simultaneous 48khz wav files through XLR using 6 AA batteries.

Zoom — R24 (http://www.samsontech.com/zoom/products/multi-track-recorders/r24/)

Thanks Eric,
That could be the best response to my original query.
So, $500 for the unit, another $180 for splitters and patch cords.
And probably an additional 3-4 hours of post for mixing and it should be the best possible solution.

A Couple questions though . . .
Say the performance is:
2 Vocalists (using hand mics & headsets for different songs)
Keyboard (piano + synth)
Drums
Bass
Brass (3pcs)

-do i request a submix from the board combining the different mics for each vocalist?
-same for drums
-same for keyboards

I guess all XLR cables aren't made the same . . . what should i look out for?
I don't specialize in this niche . . . but this is getting close to a solution. I'll have to raise my rates for this setup.

Thanks again.

So, $500 for the unit, another $180 for splitters and patch cords.

Consider also the trick of using channel inserts as recording outputs:

http://www.hosatech.com/product/0/DOC-106/_/

Amazon.com: Hosa DOC106 1/4 TRS (M) to 1/4 TS (F) Direct Out Insert Cable 6 Inch: Musical Instruments


This will look less suspicious than a bunch of Y-cables and then you don't need to worry about isolating yourself from the phantom power. I would also suggest using one well placed stereo microphone to make sure the brass players get recorded and as a backup. What's possible in terms of sub-mixes depends on the exact mixing desk and level of cooperation.

If you have a notebook computer and don't mind bringing it along, you may prefer using a USB interface such as the Tascam US-1800 for multi-track sound recording.

Product: US-1800 | TASCAM (http://tascam.com/product/us-1800/)

Thanks Eric,
That could be the best response to my original query.
So, $500 for the unit, another $180 for splitters and patch cords.
And probably an additional 3-4 hours of post for mixing and it should be the best possible solution.

A Couple questions though . . .
Say the performance is:
2 Vocalists (using hand mics & headsets for different songs)
Keyboard (piano + synth)
Drums
Bass
Brass (3pcs)

-do i request a submix from the board combining the different mics for each vocalist?
-same for drums
-same for keyboards

I guess all XLR cables aren't made the same . . . what should i look out for?
I don't specialize in this niche . . . but this is getting close to a solution. I'll have to raise my rates for this setup.

Thanks again.

Your biggest headache is going to be the drums. In many venues drums won't be mic'ed at all and their only presence in the mix for your recording would be through bleed into other performers mics. You're going to need a full set of mic's for the kit, 5 to 7 or even more mics just for the drum kit, each going to their own input on your mixer.

Electric guitars and basses might not be mic'ed either, depending on their cabinets to provide enough volume for the audience. DI boxes to split their signals to your recording mix might be necessary for them.

As been said before, mic'ing a group for recording and mic'ing them for a stage performance are two very different things

Dead End, for most budgets i'm associated with.

You're going to need a full set of mic's for the kit, 5 to 7 or even more mics just for the drum kit, each going to their own input on your mixer.

I've used an overhead boundary microphone with success. Rumor has it that some drummers will allow you to gaffer-tape a boundary microphone to their chest, but I don't know about that. Anyway, without a crew it needs to be simple and not require too much time to set up.

Crown International / Knowledgebase / Which PZM should I choose for drums? (http://www.crownaudio.com/kb/entry/174/)

Ben, everything you say there is entirely true. Mixing a close mono mic and a distant mono mic could give you some rather strange results. That is probably less of an issue if the "reverb" mic is mixed in at a relatively low level. For example if the level from the "reverb" mic is 1/10 the level (-20dB) of the "close" mic, then the comb filtering would not go to zero, it would go only to 90% (-.915dB). That doesn't make it a good idea, but if used very judiciously it would be less of a problem.

However, I notice that the scenario suggested above by Eric Olsen is a special case. He suggests mixing the "reverb" mic as if it were the "side" mic in an M/S setup. Let's look at the math. We'll call the mic close to the stage the M mic, and the distant (reverb) mic the S mic.

So if he mixes the two together in a typical M/S matrix,

L = M + S
R = M - S

Now if you further mix those together equally, to get a mono signal, you get:

Mono = (L) + (R) = (M + S) + (M - S) = 2M. In other words, in the mono mix the "reverb" mic, which we're calling the "side" mic or "S" in this equation, disappears completely. In this special case only you end up with complete mono compatibility, although with no reverb mixed in... just a close-up mono recording from the mic closest to the stage.

Be that as it may, I would not choose to do this, for three reasons. First, you're taking pressure-related signal from the "reverb" mic and putting it into the two stereo channels completely out of phase. If the level is at all significant, it could result in that "hollow" or "sound inside one's head" effect. Second, if the "reverb" level is at all significant, you will have some comb filtering. And finally, as mentioned above, the reverb will completely drop out of the mono mix.

IMHO if I wanted a close mono mic, with additional ambience mixed in, I'd use a stereo ambience mic, mixed L/R like a normal stereo source, and mix the close mic signal to the center, after delaying it so it is coincident with the ambience mic. Not ideal, but perhaps workable. Just my opinion... deposit 2¢ please.

Oh boy... where to even begin. I guess I'll leave it at this because there is so much wrong with this I don't know where to start. As an audio engineer, this really makes me cringe.

The problem with this whole line of thought is that this is not mid-side. Mid Side is a very useful stereo microphone technique, especially when it comes to mono compatability. A ambient mic does not make a side mic. the whole point of that mic technique is sum and difference. The middle mic has everything that a figure eight mic (as the side) does not have. To take a sum and a difference of those two mics makes a stereo image. Two individual microphones in different locations will never make mid and side.

I don't mean to be blunt or rude here, but I do need to try to clarify a correct usage of a technique. Two mics as originally described is mono only. It cannot be stereo.

--Ben

Ben:

I think you and I are actually in agreement. In my next to last paragraph (as quoted above) you'll note that I say I would not do what the OP suggests (namely using a mono mic -- presumably a pressure mic -- deep in the house, then mixing it in as if it were a "side" mic). And I give my reasons for disagreeing with that proposed technique.

Indeed, I think the only time the M/S matrix really produces true stereo results is with a coincident pair of proper M/S mics (i.e. the Side mic is figure-8 facing sideways). That is not what was proposed by the OP so the results will be entirely artificial and technically prone to numerous problems.

--

But I did find it an interesting sidelight that, if one followed the OPs proposed setup, the "reverb" mic would be added to the left channel, and subtracted from the right channel; therefore it would entirely drop out of the mono mix. (That's what I was trying to illustrate with the math.) So this particular proposed means of getting a bogus stereo mix, would in fact produce a clean mono mix, consisting of only the front mic.

--

Be that as it may, I agree with you, the proposed mix is not the best way to go about things.

I don't mean to be blunt or rude here, but I do need to try to clarify a correct usage of a technique. Two mics as originally described is mono only. It cannot be stereo.

No offense, however, I was not suggesting that mixing the ambient mic and monaural stage mike as M/S was correct usage, only that this might be better than mixing it as an X/Y pair. One way to create pseudo stereo is by adding a delayed signal to the left channel and subtracting it from the right. This is called the complimentary comb filter technique and similar in spirit to mixing the ambient mic as if it were M/S.

One way to create pseudo stereo is by adding a delayed signal to the left channel and subtracting it from the right. This is called the complimentary comb filter technique and similar in spirit to mixing the ambient mic as if it were M/S.

Yes, exactly. That is a known technique. If you cannot record more than two channels that are perfectly locked in time, and if you keep the level of the "reverb" mic relatively low, compared to the "stage" mic, you will get similar results: it will have some apparent stereo spread, although it won't be true stereo.

Of course it would be much better if you had a recorder with at least three tracks, so you could put a stereo mic out in the house. Better yet if you could then delay the stage mic to be coincident in time with the house mic.

Going back a tad to this..
Your biggest headache is going to be the drums. In many venues drums won't be mic'ed at all and their only presence in the mix for your recording would be through bleed into other performers mics. You're going to need a full set of mic's for the kit, 5 to 7 or even more mics just for the drum kit, each going to their own input on your mixer.

Electric guitars and basses might not be mic'ed either, depending on their cabinets to provide enough volume for the audience. DI boxes to split their signals to your recording mix might be necessary for them.
If the live sound people have not miked up the kit with individual mics, it's because there was no need - if this is the case, then for a recording there is no need to individually mic the kit - unless maybe it's a heavy rock sound, and if that was the case, then the kit would be miked for the audience too!

Boundary mics above work rather like an omni, or very, very wide cardioid - they cannot work as a boundary mic unless they're on a large surface - other than that, they're not boundaries. Sure - you can try clever placement on the drummer, but why not, if you really must mic the drums, just use an overhead, or an overhead with snare, or the verity common kick, snare overhead system. To complicated for a casual video without real sound people to work it!

Going back a tad to this..

If the live sound people have not miked up the kit with individual mics, it's because there was no need - if this is the case, then for a recording there is no need to individually mic the kit - unless maybe it's a heavy rock sound, and if that was the case, then the kit would be miked for the audience too!

...!

There's no need because the audience can hear the drums directly, true enough. But the RECORDING has no direct sound - if the kit isn't picked up by a mic somewhere it doesn't exist in the recording. Close mic'ing some instruments while relying on bleed or a house mic to capture others is asking for trouble. If you're recording the entire ensemble with a well-placed stereo mic then you are correct ... you're picking up whast the audience is hearing. But if you want high-quality concert footage you're going to have to mic up the group as if you were in the studio laying down a track destined for CD

That is NOT what I'm saying. I'm merely pointing out there is no absolute requirement to stick a mic on every drum. The style of music dictates the drum mic technique you use. Individually eq'd and processed rock drums with perhaps sampled sound slavered over need isolation, whereas many other drums styles just need a more modest amount of mics to sound good. I'm quite aware that there won't be drums or other instruments in the audience mix, so you need to replace them - but a 3 mic approach works really well for most styles, and for big band, just one or two is quite sufficient. The reality is that most video people do not have sufficient facilities to hand to manage multi-mic audio mixing themselves. As it happens, I do - and can record to hard drive 24 channels, or 16-32 on a computer based system. There is no 'standard' solution, it's worked on fluidly to get the best results for each project. My point is simply that you don't need to mic up every single thing just because you can.