Recording in stereo...
Jan 12, 2004 - by barpi75Why make things complicated when … Our ears are great: they hear better than technology, last shout or not, could restore. Yes, apart from the complexity of the means, the stereophonic recording will be an "interpretation" of the resonant space. Some sound recording techniques, tested and coming from ancestral times, have succeeded in approaching the quality of our auditory perception. They are listed here, from the most familiar to the most obscure, these techniques will allow you to achieve quality recordings of solo instruments, bands, and environments, all without getting your feet trapped in wires.
|Recall the definition of Stereo:|
A set of recording, reproduction and diffusion processes, allowing one to give the feeling of acoustic relief. Adding a word after Robert, do not forget that it is the brain that “buildsî the acoustic relief, not your monitoring device or hi-fi system. I won’t go into how the brain works, but here a few essentials to understanding this article.
Dephasing: it measures the delay between two sounds. Assuming the speed of sound in the air is constant, this delay, for us, will be the time difference for a given sound to reach the various sensors. Dephasing is an essential notion to our brain for acoustic relief reconstitution. However, our brain doesn't distinguish the phase of one sound, only the dephasing between the right and left ears.
Thus, you can easily guess that a sound doesn't arrive at the same time to our right and left ears (so, there is dephasing), and that you can vary the dephasing by turning your head (or if you are a little insane, try turning your head from left to the right very quickly! You are a phaser now. Just watch your step …). Of these observations, one can deduce the guiding principle for "measuring" dephasing, and therefore reconstitute the stereo: the captation system (i.e. microphones) must create the (artificial) conditions so that the sound coming into its left and right doesn't arrive simultaneously.
However, the difference of phase alone is not sufficient, frequency and intensity are also important. These associated phenomena are complex but less so in the context of talking about the distance travelled to the ears. For example, if the sound comes from your left, your head absorbs part of its intensity, and the highest frequencies will be more affected by this loss of intensity: thus, the sound in your right ear (compared to your left one) will be late, not as loud, as less acute. (particularly for those with big noses…)
Intuitively and according to the aforementioned guiding principle, the different methods discussed below will be crystal clear, and if you always think about the distance that different sounds will travel, you’ll be more likely to invent your own techniques (don’t believe the experts that claim that all have already been studied). If I’ve already lost you, just think about the distance that sound will travel. In this introduction, I have to say that I don’t know all the methods and that some beautiful diagrams have been picked up here and there (ORTF, AB, Decca Tree ones).
|1. The "ORTF pair" or "spaced out XY pair" and its variants.|
This technique will reproduce level and phase differences between right and left. It is therefore, at the same time, a phase and intensity stereo.
For mixing all in mono, the ORTF used to think it was OK, but because of the phase difference, some sounds will be deleted. So, be careful.
A 110° angle and 17cm spacing between the capsules are considered like the best human audition approaching parameters (think about a human head).
This technique will give a good spacialization and an accurate instrument location every time, and the resonant picture will be clean.
There are a few variants for the arrangement of the microphones:
DIN technique: The 2 microphones are placed 20cm apart to form a 90° anlge.
NOS technique: 30cm distance between the microphones with a 90° angle.
|2. The "coincident XY pair"|
This technique uses two cardioïd microphones whose "sleeves" are arranged in an angle of approximately 90° (the capsules should be the nearest, microphones can be one on the other).
This technique reproduces level differences but not phase differences between right and left (a sound arrives at the same time to the two capsules, which is why it’s called coincident; one will also notice that the guiding principle is not adhered to). It is therefore an intensity stereo, but not a phase one.
Spatialization is a little bit worse than with the ORTF technique, but the stereo picture stays accurate (i.e. there is still a good level difference between L/R).
Mono mixing can be done without risk, as L/R phases are the same.
The angle between the axes of the capsules can vary between 45 and 180°. It is necessary to determine the angle by testing it before the recording, according to the direction of the microphones and the size of the scene (in all meanings).
As a rule, a 90° angle gives the best result, but then the spatial effect is not very pronounced.
|3. The "AB pair"|
This technique necessitates two omnidirectionnal microphones (but cardioïd ones will do in a pinch) arranged in parallel, separated by a current distance of 40 to 60cm (see above for this distance). The greater the gap, the larger the “holeî in the middle (and the less beautiful the recording becomes).
The AB pair reproduces the level and L/R phase differences, but a little less so than with the ORTF technique (you can guess why using the guiding principle). It is an intensity and phase stereo.
Mixing in mono presents less risk than with the ORTF technique (the microphones being straighter). This technique offers good depth and stereo picture, although the middle field will have a tendency to be less clean (thus, the famous hole in the middle).
The primordial parameter for this technique remains, you guessed it, the distance between the two microphones. Since the human ear (and the brain) has some difficulty isolating the direction of the source for the lowest sounds (under 400Hz), the distance between microphones should be at least 40cm.
The microphones will be normally placed between 1 and 3 meters facing the source, at the same height as the group or stage; raising them at more than 3m may improve the recorded ambience (always test it).
An evolution of the AB technique is the "Decca Tree", which includes a third central microphone in order to fill the hole in the middle (see more below).
|4. Decca Tree|
As everyone knows, this technique was created in the sixties by Wilkinson Arthur, engineer of the famous company, in order to remedy to the "hole in the middle" problem of AB technique (truth be told, I didn't know the name of the guy either).
It consists in adding to the AB pair a third microphone, omnidirectionnal or cardioïd, disposed in front of the two others.
When you’re looking at the microphones from above, they should be situated to comprise an equilateral triangle.
The qualities are the same as with the AB technique, but, as you might guess, with better definition in the middle.
The microphones used by Decca were M50 Neumanns, which are considered mythical by many sound engineers.
|5. The "MS pair" or "Middle Sides"|
The MS technique necessitates a cardioïd microphone (or omnidirectionnal) and a bidirectionnal (a figure 8) microphone arranged one on the other, facing the source.
The bidirectionnal figure 8 microphone captures the sounds of the two sides while rejecting those arriving from 90 degrees to the axis of its sides.
The omni will pick up middle sounds, the bi the sides. This technique will permit some very precise stereo pictures with an excellent mono compatibility.
For decoding the signals, you must have three tracks on your recorder (I’ll try to be clear, because it can get a little twisted):
The separation of the channels serves to combine:
(a) the central and in-phase lateral signal in order to create one side of stereo signal ;
(b) the middle signal and the phase-inverted lateral signal in order to create the other side ;
(c) the central signal to the middle.
(a) the in-phase lateral signal in order to create one side of stereo signal ;
(b) the phase-inverted lateral signal in order to create the other side ;
(c) the central signal to the middle.
The mono sound is in the central signal.
Anecdotally, the microphones sold with MD tape-recorders are MS (rotted, but MS): there are two capsules in the same body (you can open them, but be careful). One can choose the 90° or 120° position for the vertical angle of sound taking.
The “sound-polesî used in cinema often have a MS pair hidden under the “hairsî. This method is very often used in audio-visual applications.
|6. The "Jecklin" disk|
Attention, we have now arrived in the land of experimentation.
Discovered by Jürg Jeckling of Switzerland, this technique consists of inserting, between the two microphones of the AB pair, a disk (20cm diameter) made of an absorbing but not reflecting material (e.g., cork). Microphone capsules should be at a 17cm distance (measuring through the disk).
It is a patented system (this is ridiculous : it looks like a good old human head, but OK, he was Swiss). However, if you put the microphone capsules 18cm apart, it still works and doesn’t infringe upon Mr. Jecklin’s patent.
The dephasing increase (since the resonant waves must get round the disk) adds to the mask effect for the sounds coming from the sides (more or less intensity is “eatenî by the obstacle in function of the sound frequency) and increases the stereophonic separation (definitely, it is a human head).
|7. The Schoeps sphere|
This process will give some ideas to the more handy of you readers.
It will be necessary for you to have a half sphere with a diameter of 20cm.
Two omnidirectionnal microphones, diametrically opposite, are placed in the interior of the sphere. The capsules’ extremity should extend a little bit out of the sphere.
The signals could be treated as in the case of MS technique.
I believe this is also a patented system (again, ridiculous).
|Some convenient advice|
For the choice of cardioïd or omnidirectionnal microphones, don't forget that the human ears have a small diameter (except for some extreme cases seen in Star Trek), so the smaller the capsule is, the better. So a tubular mic will give a better result than an SM 58.
Honestly, the difference of reproduction between a cardioïd or an omnidirectionnal is mostly subjective in a lot of cases. You can usually use one or the other, except in the case of a room where you must place the microphones close to a wall: the reverberations can be undesirable, so cardioids are recommended.
Always make a test before recording (but I didn’t need to remind you again, now did I ?)
The end result will ultimately depend on the microphones’ quality.
Some techniques will function better with dynamic microphones, others with the condensors (which are much more precise and sensitive). Giving the preference to one or the other depends upon your budget, the expected/desired recording quality, and the environment (resonant level, mechanical strains). It’s up to you.