Monday, 8 September 2014

The Recording Session

In june Bob Whitney and I recorded the Cameo Quartet for this project. They played really well and Bob was his usual excellent self. He was always one step ahead of a quite fast moving session and we couldn’t have been so productive without his valuable input.

The set up was relatively complex. I spent the evening before searching in the roof of the Stevenson Hall for multiway boxes. The boxes and cables were primarily installed for BBC broadcasts and it seems that one has been faulty and I think may have been taken away by them to get fixed/replaced! This meant I had to bring down the only remaining box in the roof and drop cables for the ambient mics down to the floor which is a little bit of a pain as this limited the number of available channels and we had to make some set up adjustments (discussed later).

The rest of the set up was relatively painless, although an amount of time was lost sorting out the headphones. Unfortunately there was some slight user error on my part. But once we got started the session ran smoothly.
While we were checking levels and tweaking mic positions Bob and I were struck by how easy recording can be when done in a good space. Bob and I were talking about various engineers over lunch and their approaches and I was telling him about Trygg Tryggvason’s book chapter “Classical Music” in Sound Recording Practice Bob read the chapter after the session and the passage he picked out most clearly was:

“The criteria by which a suitable location for a recording is chosen will not be dealt with in detail here. It should be stressed, however, that a good complementary acoustic is the most fundamental requirement for music recording, without which all subsequent operations are necessarily compromised.”

This isn’t a revelation for Bob or I, more a confirmation. We often talk about the importance of the room sound and how fundamentally classical recording is relatively simple, but infinitely subtle. The ideal is always to capture the sound on as few mics as possible. This will reduce phase issues and will keep the sound from muddying. I believe that if engineers spent more time tweaking main arrays than adding spot mics more recordings would sound clear and clean.

Capturing a stereo image comes down to making a number of decisions and answering a number of questions:

  1. What is the final playback media? (headphones, speakers, multi-channel sound system,  electroacoustic diffusion, iPods)
  2. What kind of image is most appropriate for the music?
  3. What kinds of sound sources are present? (Essentially mono – a solo instrument, properly stereo – ensembles, piano, tuned percussion)

So for this recording the playback medias will be stereo speakers and headphones. I wouldn’t expect anyone to want to listen to this on headphones in the street so it doesn’t need overly compressing – if at all.

The image, as discussed earlier, is wide. However, distance is the most important spatial parameter.

A quartet is properly stereo. An individual clarinettist would be a good example of an essentially mono source. Capturing a single player in stereo will create a stronger, richer image, however, the clarinettist will still be positioned in the centre of the stereo field. Individually, therefore, the clarinettists are essentially mono, when combined they create a stereo image.

Stereo sound stages are created in a number of ways. Coincident techniques provide strong lateral imaging whilst spaced techniques capture depth more effectively. For coincident techniques the difference in amplitude (loudness) of the sound at each mic capsule determines its position in the stereo field whilst time arrival differences create a stereo image using spaced microphone techniques. Spaced techniques commonly use omnidirectional microphones which means they capture more of the room sound than coincident techniques that use cardioid microphones.

Adding more mics to these main arrays creates time differences between the different mics (spot mics will receive the sound earlier than main arrays and ambients will receive it last). These micro delays in time arrival can both help establish a sense of depth or create instability in the stereo image. It is the job of the engineer to manage these issues, selecting which recorded perspectives can be combined effectively.

Having said all this, to make a great sounding recording it is really a very simple equation. Whilst carefully selecting a space and recording method the most important aspect is, of course, the actual performance. Perhaps great recordings can be summed up by the following equation:

(Great Mics + Great Posotioning + Great Acoustic) X Great Performance = Great Recording

The space we recorded in is the Stevenson Hall at the Royal Conservatoire of Scotland. Below you can see a picture of the quartet and I, the mics and stage. This should give you some sense of the scale of the room. A medium sized concert hall, good for large chamber ensembles, but it can work well for smaller ensembles (from solo instruments) and can accommodate a symphony orchestra with a sensitive conductor. It is a good size for this piece. Whilst there are only four clarinettists, when the fixed-media part is added the scale of the piece feels relatively large. The vertical space in the hall, in particular, allows the sound to breathe.

Whilst we were setting up we decided to change a few aspects of the system plan. We decided not to use the Decca Tree for a variety of reasons:
  1. We felt the distance between omnis in the Faulkner array and the decca tree was too small to make the wider omnis useful as outriggers.
  2. The tree does not offer the ability to balance the relative proximity of the clarinets in the mix.
  3. Outriggers seemed to be more crucial rather than simply adding another kind of main array.

We ended up with this set up:

Here we have the main Faulkner Array on the wide stereo bar using B&K 4006 microphones (omnidirectionals) and Microtech Gefell M930 microphones (cardioids) These also make up AB and ORTF systems.

In front, the rather larger mic is an AEA R88 mk2. This is a stereo ribbon mic. Ribbon mics are all bidirectional and in stereo ribbon mics they are arranged at 90 degrees to one another. This means you can use them for Blumlein or mid-side techniques. I’ve never been a fan of mid-side, and I’m slightly skeptical about the artificial rendering of the stereo image (you can read about that here Blumlein has its own issues, but as it is a coincident technique offers excellent spatial positioning (for the person positioned in the sweet spot) coupled with a nice rendering of spaciousness afforded by the bidirectional polar pattern.

To these arrays we added spot mics, outriggers and ambient mics.

The Cameos had worked hard in the rehearsal and the more they are playing the piece the better they are engaging with it. In the session we took a full take and then worked through it section by section.

In later posts you’ll hear more examples, but here is a very quick mix of the edit.


TRYGVASSON, Trygg. (1996). Classical Music. In: BORWICK, John Sound Recording Practice. Oxford: Oxford University Press. 210-228.

Sunday, 22 June 2014

Why Classical Stereo Techniques?

In my previous post I talk about what kind of sonic imagery is present in the piece. It is essentially stereo:

1.      The tape is in stereo, exploring depth and having a sense of width.
2.      The clarinets sit symmetrically and therefore can be captured evenly in stereo.

Capturing the piece in stereo makes sense because:

1.     Stereo techniques capture depth, to a greater or lesser extent depending on the technique.
2.      Stereo techniques capture the lateral positioning of instrumentalists, again to a greater or lesser extent depending on the technique.

Stereo techniques often come with quite fixed rules. ORTF for instance is a pair of cardioid microphones with the diaphragms 17cm apart with the capsules splayed out 110 degrees apart.

All stereo techniques can be adjusted to taste, although there are schools of thought that are more scientific about this – this is not within the scope of this research. In order to give myself a variety of options for mixing it is important to plan which techniques I will record. (n.b. I have a limit of 16 channels.) 

Some Terminology:
I don't want to assume that all the readers are audio experts so I think it is worth defining some terms that I will be using frequently: 




Polar Pattern

Microphones do not simply pick up sound. They accept and reject sound depending on design factors. A polar pattern is defined by how it picks up unidirectionally - or not. These patterns are then plotted onto a graph to represent a microphone model’s unique pickup characterises.



A polar pattern that picks up principally frontally and tends to reject sounds to the rear of the mic. These mics tend not to pick up the ambience of an acoustic very well.

Image of the the polar pattern for a microtech gefell M930 microphone.

(Example 1)


A polar pattern that picks up equally right around the microphone. In practice omnis are not always strictly omnidirectional and the uni-directionality of the pick up varies by frequency. These mics tend to pick up the ambience of a room very well.

Image of the polar pattern for DPA 4006a microphone. This diagram shows how the pick-up of the microphone changes depending on the frequency of the sound (essentially, omnis tend to pick-up higher frequencies more directionally)

(Example 2)


A polar pattern that pick up in a figure of 8 shape. These microphones reject sounds coming from the side of the mic. These mics tend to pick up the ambience of an acoustic quite well and the side rejection can be advantageous in rooms where sounds echo (slap) between the side walls.

Image of the polar pattern for one capsule of the R88 stereo ribbon mic.
(Example 3)

Which techniques and why?


What is it?


Faulkner Phased Array

A 4 microphones system, using a near coincident pair and a pair of omnidirectional microphones, slightly splayed out, 66 cm apart.


  •  Shoeps CM2 Omnidirecitonal microphones
  • Microtech Gefell M930 cardioid microphones (possibly swap out for M950s, wide cardioids, if the bass response is poor)

This array itself allows one to adjust the balance between closeness and spacious-ness. In theory this will allow me to mix a more intimate version and a more ambient version depending on the context and relative attractiveness of the acoustic.

The piece is more about depth than width. Whilst the piece is wide sounding and there is plenty of lateral movement the main spatial parameter consciously composed with is relative depth between the tape and live instruments.

The ability with this array to discreetly balance the perceived depth makes this potentially an ideal stereo system.


A two-microphone system. ORTF is one of the near coincident techniques using two cardioid microphones. In ORTF they are spaced 17cm apart and angled out 110 degrees.

In a practical sense it is a part of the Faulkner array and it makes sense to record it separately.

ORTF is charaterised by clear lateral imagery but often a slightly weak bass response. Swapping the M930 cardioids for M950 wide cardioids may improve the bass response - although using wide cardioids will make this no longer an ORTF array.


A two-microphone system. Microphones of any polar pattern arranged so they address straight ahead.

In this case, two omnidirectional microphones spaced 66cm apart.

In a practical sense AB omnis are a part of the Faulkner array and it makes sense to record it separately.

AB omnis will have a very even frequency response and excellent spaciousness, but perhaps a slightly indistinct lateral stereo image.

Decca Tree

A three microphone system:

  • 2 parallel omnidirectionals 2m apart
  • 1 omnidirectional centre mic 1.5m ahead of the main pair.


2 Shoeps CM2

1 other omnidirectional microphone

The decca tree is a classic stereo technique. It captures a beautiful sense of spaciousness and may allow me to simply capture the balance in the room. It may also simply capture the tape part in the room.


Typically Omnidirectional microphones arranged in line with the main array(s) spaced reasonably widely (perhaps as wide as the ensemble itself).

(The decca tree omnis may be spaced wide enough to replace these.)

Outriggers typically add width to a recording and a sense of ambience to a main array.

Ambient Pair

Omnidirectional microphones arranged at the rear of the hall, very widely.

Ambient mics add room resonance to a recording. I will be able to use these to add ambience to any of the main arrays.

Spot Microphones

Individual microphones for each instrument. Typically cardioid microphones.

These will be used to add detail if needed and to provide a frame of reference for the other systems.

Saturday, 21 June 2014

Compositional Materials in Fata Morgana

Some Assumptions

As Fata Morgana is my own piece and I know it reasonably well (!) I have many preconceptions about recording it. Most of which are probably fairly close to the mark but are worth exploring further. These include:

  • Expecting classical micing techniques to produce the best results.
  • Recording the tape part in the concert hall, after the quartet have played, will allow me to blend the ‘dry’ fixed electroacoustic part and the fixed part in the space more effectively with the live clarinets.
  • Using close mics may add extra detail and intimacy, highlighting any breathiness present.

Whilst these thoughts may be perfectly valid it is worth thinking about the piece in more depth before making any decisions.

The General Images in the Piece:

The piece, whilst being in 4-channels, is essentially in stereo. The four-channel tape is a means to distribute sounds across a stereo sound stage. The clarinets themselves are distributed symmetrically. The tape part in live performance should not be distributed around the audience and so stereo recording techniques become the obvious choice.

Spatial Planes:

In this piece spatial width is important, but it is not discreetly composed with. There are no large left-to-right or right-to-left gestures that articulate dramatic change or sounds being panned extreme left or right in counterpoint or opposition.

Depth however, is discreetly explored, especially the relative depth between the clarinets and the fixed electroacoustic part. Fata Morgana are especially vivid mirages and the title became a frame for shaping the illusions I was creating between the live clarinets and the fixed part. The programme note sums this up:

“The image you see and hear is always changing. The clarinets, both real and recorded, come in and out of focus in a fluid and illusory way. As the image dissolves, like a mirage, the real clarinets may appear closer or more distant than they actually are. As they melt and mix together the materials they carry are re-imagined creating new perspectives that whilst different, retain the same character and essence throughout. You can never be quite sure that what appears to be on the horizon is really even there. 

'Fata Morgana' are unusually vivid, and complex mirages named after the Arthurian Sorceress Morgan La Fey. In my Fata Morgana the illusions are subtle; they are not grand tricks like a Houdini escape.”

So, because of this, one of the most important considerations is not to flatten the stereo sound stage. This tends to imply that not using too many mics at any one time will be important as often flat stereo sound stages are a result of over use of spot microphones. Whilst I will record a relatively large number of mics it will take discipline to be discerning about which ones I use in final mixes.

Localisation of Sounds

It is very easy for me as a recording engineer to assume that discreet localisation of the sounds in this piece then is less important that in other pieces. However, as the composer of the piece this is harder to assume. If depth is the most important consideration this might imply the use of omnidirectional microphones to best capture this aspect. However, whilst width is not a discreet parameter in this composition the sounds were still carefully placed to create a balanced stereo image. In order to capture the location of the sounds an XY* system will create the most accurate lateral image. However, the lack of depth in XY stereo pretty much immediately discounts it for this purpose. This leads on to a variety of stereo systems that will discussed in more detail, with relation to localization and depth, in a future blog post.

The Clarinet’s Materials

When considering what a certain kind of material is it is also worth considering what it is not. I have found it useful to define the clarinet parts with regard to temrinilogy developed in Smalley’s Space form and the Acousmatic Image where he describes performance space as
“…gesturally rooted. A human agent, using the sense of touch, or an implement, applies energy to a sounding body, producing spectromorphologies. Alternatively, internal physical energy generated from inside the human body can be applied to an external sounding body, like a wind instrument.”
Whilst this is of course true of the way the players interact with the instrumental material this piece does not foreground these actions. This is not a piece highlighting key clicks or other tiny sounds. The clarinet’s might appear “zoomed in” but it is more zooming in on the natural sound and timbre, rather than the mechanics of the insturments.
Moving away from performance space Smalley goes on to describe ensemble space as:
Ensemble space, within which individual gestural spaces are nested, is the personal and social space among performers: a group of performers produces a collective performed space. This is revealed both visually (seeing proprioception at work, and knowing how it works) and in the music (hearing proprioception at work). In duos or small ensembles the space is more personal…”
This then is the principal “space” that the clarinets occupy in Fata Morgana. Generally they work as a unit in order to create fully formed textures of their own in order to texturally interact with the fixed part. For much of Fata Morgana the clarinets are all engaged in playing the same kinds of materials at any one time although occasionally they break off into pairs.
Smalley continues
“Sound can bind spatial zones when like behaviour creates a sonic contiguity where we cannot separately identify individuals, as with a violin section, for example. In ensemble space we witness the fabric of the music in the process of articulation – the synchronisation of a collective gesture, collaboration or competition within texture, and exchanges of materials.”
So it is this kind of space that I hope to capture. This would tend to imply that main arrays will be more effective; after all they are designed to capture a well balanced performance made by well rehearsed musicians. However, many very good recordings have also been made using individual spot mics and it is perfectly possible to create a blended sound that does not pick up unwanted key clicks or other sounds. It will be important to carefully analyse just what the different arrays capture in order to best select which microphone positions to use.
The final principal space that Smalley defines is arena space:
“Arena space is the whole public space inhabited by both performers and listeners.  An audience may be conscious of the personal and social spaces within its own zone, and certain kinds of audiences (rock concerts, clubs), through their behaviour, may seek to enhance personal and social contexts.
For our acousmatic purposes the most important aspect is that we perceive arena space as an acoustic setting, as a bounded and enclosed space produced by the nesting of gestural and ensemble spaces within it.”
Arena space is important to my research. My aim is to bring both electroacoustic and live parts into a perceivable space that they both inhabit, rather than trying to bring the instrumental part into the acousmatic fixed part. Identifying these kinds of spaces also affords me the opportunity to try to creatively move between them. For example, a more distal section could be mixed to use a more ambient recording sound and a more intimate section could be mixed to use just spot mics and the dry fixed part.
* XY stereo uses two cardioid microphones with the capsules on top of each other angled at 90 degrees.

SMALLEY, Dennis (2007), Space-Form and the acousmatic image, Organised Sound, Volume 12, pp 35-38, DOI 10.1017/S1355771807001665,