3 Recording in Stereo
When recording with stereo techniques, the panning decisions are being made before the editing and mixing process begins. This automatic panning is caused by the interaural differences being recorded by the microphones, which will then be replayed to listeners over speakers.
There are different types of stereo recording technique, which are outlined in this post along with examples of particular techniques following later. It is worth nothing that different technique types will make use of interaural differences in different ways, so the sound can differ, dramatically.
Spaced microphone arrays are multi-microphone techniques which incorporate a spacing between them. The significance of a spaced technique, is that sound will ‘hit’ one microphone before the other. With the sound captured by both microphones, the interaural timing difference between the two will also be recorded.
When the recording is played back over a good stereo speaker setup, the sounds will ‘leave’ the speakers with the same timing gap which the brain will interpret as directional cues. Depending on the technique being used, sound intensity differences can also be a significant source of directional cues. For example, a spaced technique using cardioid microphones will be more sensitive to intensity differences than a spaced omni-directional technique. A general spaced technique is shown in Figure 3‑1, where the direct sound path from the green source travels a longer distance to the right microphone which introduces a timing difference.
where A < B
Figure 3‑1 –Direct Sound Distance Differences between in Spaced Microphones
Coincident recording arrays are multi-microphone techniques where the microphones are placed as close as physically possible. As the microphones are so close, there will be no significant timing differences between the sounds reaching the microphones. The differences which are recorded are intensity differences. An example of coincident technique is shown in Figure 3‑2 where two microphones are placed, one on top of the other, with the capsules as close as physically possible. This technique would only suit a unidirectional set of microphones such as cardioids, as amplitude differences are the focus.
where A = B
Figure 3‑2 – Direct Sound Distance Differences between in Coincident Microphones
The difference in sound between these two types of techniques are generally quite distinct. Spaced techniques tend to sound softer, rounded and spacious which can be partly put down to the phasing which spaced techniques record, just like our ears. Coincident techniques generally sound sharper, less spacious and more direct as there are little or no phasing issues, which can tighten up the stereo image.
With so many different techniques used by engineers, these general sound characteristics can vary. For example, some spaced techniques can sound sharper than others. There is also no right or wrong stereo technique, so it is worth experimenting or being open to as many techniques as your microphone collection allows for. Certainly, experimenting with the specifications of a single technique can of great benefit, such as adjusting angles or using difference size microphone diaphragms.
A common phrase which can be heard in the area of audio production is that “there is no right or wrong way” of doing certain things. Where an engineer was tasked with making a recording which replicated the way in which humans hear, using a spaced technique with many metres between microphones would certainly be the wrong way to do it; however, more often than not, the task of an engineer is to create a sense of hyper-realism.
This is a concept where certain sonic aspects of a mix are emphasised in order to keep listeners interested or to provide them with a ‘larger than life’ sound, or more basically, a better sound. Sound design is a very clear example of this. If you listen to dramatic series, footsteps or breathing would perhaps be louder than they would be in real life. The classic movie punch is made up of a number of layers such as a basic punch, a thump of the chest and a whoosh sound of a stick moving through air.
What this means is that engineers are commonly tasked with creating an unrealistic sound, but in a positive and constructive way. In the context of mixing, using EQ to clear up part of a mix is something humans cannot naturally do so recording with an array which offers ‘unhuman’ levels of clarity could be just what is required.
4 Stereo Recording Technique Examples
In this section, a number of stereo recording techniques will be outlined. General considerations will be given which will tie in to a lot of what has been covered earlier in this guide. All techniques use condenser microphones, as dynamic microphones do not have the higher frequency response required for detailed recording of delicate instruments. Experimentation is key. All these techniques will have specific measurement, angle and polar pattern requirements. In some situations, you could decide that a combination of techniques could work for your particular situation.
Spaced omnidirectional stereo arrays have long since been a popular way of recording classical music, especially classical ensembles. By using omnidirectional polar patterns, the two microphones pickup ambient sound as well as direct sound. The spacing between the microphones can vary; however, a starting point of around half a metre can work well. If the spacing is widened to approximately a one metre or more, a common issue known as the “hole in the middle” can occur. This is a situation where the stereo image is so wide that sound sources will be bunched on the left and right sides of the image, with nothing of significance being perceived in the middle.
For example, dealing with larger ensembles would result in a wider spacing between the stereo pair so an additional microphone placed between them can be used. A single extra microphone could be placed in the centre which would fill in the hole in the middle when it is panned equally to the left and right.
Figure 4‑1 – Three Microphone Omni Stereo Array
If the ensemble is wider again, a fourth microphone can be used. In this case, the original left and right microphones would be panned hard left and hard right, and the additional microphones would be panned part way left and part way right. The amount which all the microphones are panned is down to what sounds best for the performance, but care must be taken to ensure the difference between the L to CL and CR to R are the same.
Figure 4‑2 – Four Channel Omni Stereo Array
The spacing between the microphone array and the musicians is dependent on the balance of direct to ambient sound that you want. For instance, if you want more ambience, then the microphones should be moved further away from the musicians. A closer distance will give more direct, clearer sound.
Having a good monitoring setup with test recordings of different positions can help decide how far the microphone array should be from the musicians, but be sure to use your own ears and stand in various spots to find the balance you want.
Spaced omni-arrays tend to sound rounded, warm and soft. The omni polar pattern inherently allows ambience into the recording which adds to this perception. If there is too much ambience, changing the polar patterns to figure of 8 or a form of cardioid can help, with the use of room microphones providing a controllable source of ambience reinforcement.
The Decca Tree was developed by Decca Records in or around the 1950s. The tree can be configured as a three microphone stereo array, so the centre microphone is routed in equal amounts to the left and right speakers which adds stability to the centre of the image. It can also be used as a basis for surround sound recording arrays.
Interestingly, the Decca Tree uses all omni-directional microphones which would mean that no microphone angles are required; however, the original Decca Tree specified Neumann M-50 microphones, which were notoriously directional from around 1kHz upwards. This was a known characteristic of the microphone, so the Decca Tree actually has microphone angles specified; however, exact angles are difficult to come by with the technique being so old,. It is also very popular, with engineers often tweaking the specifications, so make sure you take sample recordings to reference off when making decisions on how to angle microphones and their distance from the musicians.
The general rule of thumb which can be used is to place the array far enough from the musicians to allow for a nice balance of direct to ambient sound, with the left and right microphones angled such that they point towards the left and right edges of the ensemble. The distance between the Left and Right microphones is generally two metres with the centre microphone placed one and a half metres in front. These distances can be experimented with for smaller ensembles.
Figure 4‑3 – Basic Decca Tree layout for stereo recording
Recording engineer Akira Fukada re-specified the Decca Tree to use cardioid microphones and fixed spacing, which has become known as the Fukada Tree. The use of unidirectional microphones allows for more control of ambience in the mix.
Implementations can vary between the Decca and Fukada trees as the engineer sees sonically fit. The full Fukada Tree is actually a surround sound recording array, but it can be used for stereophonic recordings by just using the front section of the array as shown in Figure 4‑4.
Where distances and angles of the Decca Tree can be somewhat unclear, The Fukada Tree states that the distance between the left and right microphones is twice that of the distance between the centre microphone and the left/right microphones. This relationship is shown in Figure 4‑4 where the green line is half the length of the red line. The distance between the left and right microphones can be anywhere between 180 and 360cm, with the green line half that figure. The angle between the left and right microphones is between 110 and 130 degrees.
The ORTF technique was developed by the Office de Radio-Television Diffusion Française which is a French broadcasting organisation. It was designed to combine the image sharpness of a coincident technique with the spaciousness of a spaced technique hence its characterisation as a spaced or more importantly, a near-coincident recording technique.
This is achieved in two ways. The first is the use of a small spacing of 17 to 18cm between the microphone capsules. The second way is the utilisation of unidirectional polar patterns such as cardioid, super-cardioid and hyper-cardioid which reject ambient sound to a significant extent when compared to omnidirectional microphones.
The small spacing allows for a more human use of the Interaural Time Difference while the unidirectional polar patterns help simulate the directionality of our ears with respect to the Interaural Intensity Differences.
Figure 4‑5 – ORTF with 18cm spacing
The NOS technique was developed Nederlandsch Omroep Stichting which is the Dutch equivalent to the ORTF. This array uses a larger spacing of 30 centimetres with a microphone angle of 45 degrees from centre. Note that the bigger distance results in a tighter angle.
Figure 4‑6 – NOS with 30cm spacing
The spacing and angle differences between these arrays mean that when recording the same source and when placed in the same position, there will be differences in the stereo image. Figure 4‑7 shows that the ORTF technique will evenly space sound sources across the stereo image. The NOS technique will slightly bunch sources away from the centre of the image.
Figure 4‑7 – Stereo Images: ORTF in yellow and NOS in red
Although the NOS image may not seem desirable, CL and CR being pushed away from the centre can be very useful in giving important centre instrumentation a clear place in the mix.
Alan Blumlein developed a stereo recording technique where two figure-of-8 microphones are placed coincidentally. This generally means one microphone would be placed upside down, on top of the other. microphones are angled 45 degrees left and 45 right from centre respectively, so be sure to note which is pointing left and which is pointing right. As seen in Figure 4‑8, each lobe of the figure-of-8 patterns overlap with two others which is very important for the front portion of the array.
Figure 4‑8 Blumlein Array
The figure-of-8 microphone will be most sensitive to sound which comes directly in front of it, or on-axis. If the sound slowly moves off-axis, the sensitivity drops. When a sound is placed in the very centre of the array, both the frontal lobes will be 45 degrees off centre and thus equally less sensitive. With the left and right recorded signals are combined acoustically, the resulting centre signal will be 3dB lower in amplitude compared to full left or right sources, which is exactly how the pan-pot law would work. The result of this is a perfectly uniform amplitude pickup of sounds from the full left to full right.
This acoustical summing is important as the 3dB rule in analogue and digital mixers will only ever apply to a signal when it is panned somewhere between full left and full right. As stereo arrays will generally be panned fully left and fully right, the Blumlein array achieves the 3dB panning rule acoustically.
This array provides excellent localisation and focus, as well as ambient pickup which is somewhat unusual for coincident stereo arrays which generally favour direct sound rather than ambient. Also, the side lobes of the array cancel each other out as they are out of phase with each other. For example positive one added to negative one will equal zero. This can be helpful when recording where there is an audience or side room reflections which are not sonically beneficial.
Figure 4‑9 – Blumlein overlap addition
The Mid/Side or MS stereo technique is a coincident technique which uses a figure of 8 microphone which is placed perpendicular to the sound source, along with a cardioid microphone placed directly towards the sound source. This is shown in Figure 4‑10 and Figure 4‑11.
With the two signals recorded, the engineer must duplicate the figure-of-8 recording, which is called the side signal. The phase of the duplicated track must then be inverted. The two side signals should then be panned fully left and fully right. With the mid and side faders up, the resulting ‘phantom’ signal is equivalent to a set of cardioid family microphones in a coincident XY configuration as shown in Figure 4‑12.
Figure 4‑12 – Physical microphone placement on the left with the resulting XY phantom array on the right (Chris Corrigan, Queen’s University Belfast)
If the side signals are muted, the result will just be a mono mid signal. As the side signals are unmuted and slowly raised in the mix, the mono signal will split and form a coincident stereo array. With the side signals up full, lowering the mid signal will widen the angle of the XY array further until there is no mid signal in the mix, and the fig-8 signals are all that is left. This provides engineers with a powerful post-recording manipulation of recording angles, which can greatly help when mixing decisions are being made. For example, a mainly mid signal could be used in a busy mix; but, a wider stereo image can be introduced in more sparse sections of the song by raising the side signals.
126.96.36.199 Extra Mid Side Configurations
This technique is not limited to just using a cardioid microphone as the mid signal. Figure 4‑13 shows two variations of the MS technique. The top of the figure shows how a coincident 180 degree cardioid stereo array can be generated when using an omnidirectional mid signal.
The bottom of the figure shows how a Blumlein array can be generated when using a fig-8 as the mid signal. In this case, all the engineer would have to do is move the MS array 45 degrees to create a physical Blumlein array; however, the power of the MS technique means the engineer just has to switch the mid signals pattern to create a variety of recording options; which can be very easy to do with switchable pattern microphones.
Figure 4‑13 – Different MS variations (The Microphone Book, John Eargle)
This technique was originally proposed by Alan Blumlein, which was later developed by Jurg Jecklin. Two omni-directional microphones are placed 16.5cm apart. Normally, a stereo omni array spaced this close would not provide any useful stereo information; however, the OSS system uses an acoustically absorbent disk shaped baffle which is placed between the microphones. Approximately 20 degrees of an angle is used between the microphones to account for any microphone directionality, such as with the Neumann M-50. This technique can be useful for smaller ensembles such as duets or quartets.
This technique is also a baffled recording system; however, instead of a disk between the microphones, each microphone is placed flush with the surface of the sphere and where an ear would be. This simulates the effect which the human head has on the sound that reaches it. Both this technique and the OSS are considered pseudo-binaural techniques which means they operate in a binaural way, but can be played over loudspeakers.
This technique uses a dummy head which has a very realistic set of rubber ears which mimic in great detail how the human hear interacts with sounds. This creates a binaural recording of the sound space, which means that headphones should be used to best appreciate the effect. Its usefulness is generally restricted to binaural research, or commercial applications like shown in Figure 4‑17.
Every human with healthy hearing will hear things in ways which are specific to them; sound will interact differently with people who have different facial and upper body features. For example, people with long thick hair will attenuate sound from behind more than someone with short hair or no hair at all. The significance of this is that if you asked these two people to pinpoint a sound, they may point in the same direction but if you were able to feed their brains with the hearing signals of the other, then their perception of direction would be off. This means that each individual has their own personal set of hearing characteristics, which our brains learn and adapt over time. These are called Head Related Transfer Functions, or HRTFs.
The dummy head uses the HRTF concept as a basis for creating binaural sound, but it can only ever be used as an average; for example, average sized head, sized ears, sized nose etc. This means that when dummy head recordings are played to listeners, the surround sound effect can be better or worse for different people. You can have your own personal HRTF function created by having special measurement microphones placed in your ear canals, while test signals are recorded and noted.
The area of stereo recording and by extension, surround sound recording, is a very interesting one and I hope you found this guide useful. Although the theory can be heavy going, with a core set of techniques understood, some fantastic recordings can be made which will be a great addition to your skill set and portfolio. It is well worth checking out what is available on the internet on the subject. The short list of recommended reading is well worth looking into, too; especially The Microphone Book, by John Eargle. If you have any questions, please contact me at firstname.lastname@example.org.
Eargle, J. (2005). The Microphone Book. Focal Press.
Fukada, A. (2001, June). A Challenge in Multichannel Music Recording. Audio Engineering Society, Paper Number 1881.
Fukada, Tsujimoto, & Akita. (1997, September). Microphone Techniques for Ambient Sound on a Music Recording. Audio Engineering Society, Paper Number 4540.
Howard, D. M., & Angus, J. A. (2009). Acoustics and Psychoacoustics (4th ed.). Oxford: Focal Press.
Moylan, W. (2007). Understanding and Crafting the Mix – The Art of Recording. Focal Press.
Recording Hacks. (2014). Neumann M50. Retrieved March 2014, from Recording Hacks: http://recordinghacks.com/microphones/Neumann/M-50