gummblefish said:
If you have a singer that plays guitar at the same time, particularly acoustic, this should work for ya. <snip happens> So what worked for me was to make a perspex divider between the guitar and the voice. <snippity doo da> It was kinda awkward, but it did help in the separation of the tracks, plus the guitarist was able to see his guitar.
That's great!!
What I generally do is take two microphones that I would normally use for both applications and switch them into "Figure 8 pattern". If they don't go into "Figure 8 pattern" then they're not in the running for selection.
Lemme explain why [if you know why, then skip the next half dozen paragraphs... it'll only bore the snot out of you... if you don't know why (and give a shit), then take notes as there is going to be some good, useful theory ahead].
---------
WARNING:
Theory approaching [which may contradict some bullshit myth you may have heard or don't understand]
A microphone in cardioid pattern only picks up from the front... right? Well, to a degree, yes... to another, larger degree, no. As the frequency of a sound lowers it becomes more omni-directional by nature... which means that our cardioid microphone that only picks up from the front at 1kHz and above picks up everything from everywhere from say 400Hz and below... which means, that you will get a whole buncha non-specific mud building up in your recording... You could put bricks between the guitar and the singer's voice and still get bleed.
Now the way a cardioid microphone works is that you have either a single transduction element [capsule/diaphragm] that is in service at the time of use... so even with a "multi-pattern microphone" you're either acoustically [as in the case of many early RCA ribbons as well as cardioid and hyper cardioid dynamic mics... but I'm so not going into that expanation right now... we're going to stick to condeser mics for the rest of this explanation] or electronically [condenser mics] causing there to be a cardioid pickup pattern. In the case of a multi-pattern condenser mic you have a front and back diaphragm... turn off the back diaphragm/capsule and you have a cardioid mic.
If the MFG. doesn't put in a rear diaphragm then you have a TLM-103 or any of cornucopia of cheap assed 'cardioid only' "Maocrophones".
Because of the nature of sound, it is impossible to achieve a fully directional pickup pattern across the frequency spectrum. Anything below 250Hz is fully "omnidirectional" and it becomes less fully omni the higher you go in the sound spectrum so that when you get up into the "speech intelligibility" range the mic is fully "directional"... but there is a lot of 'mud' and crap underneath the 'speech intelligibility' section of the pogrom... and you can't filter out the low shit or your recording will sound "thin" and nasal instead of 'full and rich'.
The way "Omnidirectional" directional condenser mics work is that you have two diaphragms [or two capsules back to back] that are 100% in polarity so they pickup in both directions... front and back.
Now, if you take the polarization voltage to the rear diaphragm/capsule and flip it 180' [so it's out of polarity to the front capsule] the net result is that you get a distinct, full bandwidth line around the microphone where the front side and the back side cancel 100%. No other pickup pattern has this "full bandwidth" area of rejection. This is how you get a condenser microphone to go into "Figure 8 pattern" as you're electronically cancelling the information [sound] common to both the front and rear diaphragm/capsule.
End of theory... you may now return to the original retort already in progress
---------------------------------------
So... if you have two mics that go into "Figure 8" and you point the "null" [no sound] band at the singer's mouth and the business end of the mic at the singer's guitar... and take a second mic in the Figure 8 pattern and point the "null" [no sound] part at the guitar player's guitar and the business end at the guitar player's vocals... then you get the singing in one mic and the guitar playing in the other mic with no bleed [because it's electronically cancelled, which means a 'full bandwidth' sound rejection along the line where both capsules pick up common information.
Yeahbut... Figure 8 mics pick up from the front and the back so I'm going to get too much "room sound" you might say... to which I would respond: "not if you take the 'inverse square law' into account".
---------------------------------------
Warning:
More theory
The inverse square law is that for every frequency there is set piece of real estate where the sound pressure level drops 3db... so, if you have a mic 1ft. away from a guitar [or vocal] and a second guitar/vocal two feet away from the mic on the other side, it'll be 3db down from the signal on the front side.
Which means that when using a microphone in the figure eight pattern you will get some 'room' sound from the back side of the mic... but it will be sufficiently attenuated in relation to the original sound that it won't be a factor [yeah, kind of a lame description of the law... but you get what you pay for... suck it up or crack a book].
--
end theory--
--------------------------------------
The bottom line is that using two mics in 'figure 8' pattern and positioning them carefully will give you far better separation than any construction project of which you can conceive that will do little more than make the player uncomfortable and thus, less prone to give up an amazing performance. We're in the business of selling emotions here... and while "pissed off/inconvenienced" is kinda an emotion... the fact of the matter is that we don't really strive to listen to music to reinforce the stresses in our life but to distract us from the stresses in our life... which means that the music should be recorded from a happy/comfortable performer.
....and thus concludes our sermon for today.