How to get it right or very nearly right the first time By John Eargle

1
How to get it right - or very nearly right- the
first timeBy John Eargle

• A close look at stereo and surround recording in
  both live and studio settings

2
2-track jazz recordings

• Arrange players, left to right, to match the
  stereo sound stage
• Lay out console faders the same way
• This gives you intuitive visual cues and
  consistent leakage between panned images
• Heres an example

3
Live to 2-track jazztypical studio setup and
target sound stage
4
2-track chamber music

• When the players agree to it, have them depart
  from the usual concert setup
• Arrange them in circular fashion so they can see
  each other, and mic them accordingly
• This gives both engineer and producer greater
  control for on-the-fly balance changes

5
Classical example 1
6
Classical example 2
7
Classical large forms

• When a conductor is present, there are fewer
  opportunities to make fundamental changes
• Some left-right/fore-aft adjustments can easily
  be made in the orchestra
• The most likely major change that can be made is
  when chorus and soloists are involved
• Here is a good example

8
Put chorus behind the conductor
9
This has the following advantages

• Gets the chorus away from the back of the
  orchestra, where percussion and brass can leak
  into the chorus mics
• Lets the chorus communicate directly with the
  live ambience of the hall, as opposed to the dry
  ambience of up-stage

10
Live classical recording of orchestras

• This has become common in recent years because of
  escalating costs
• Conductors, players and soloists almost always
  prefer it
• There are certain musical advantages (i.e.,
  performance spontaneity and avoidance of
  over-exertion of players)

11
What are the disadvantages?

• Audience noises
• Acoustical matching (4 concerts are the norm,
  with a finishing session scheduled after the
  audience has left the last concert)

12
We normally handle these by

• Having orchestra management personnel TELL the
  audience to be quiet (beepers, watches, cell
  phones, rustling programs and coughs are all
  problems)
• No applause until the conductor has lowered the
  baton
• Coming up with a strategy for matching the sonic
  signature between the hall, empty and occupied

13
What are these strategies?

• Modern halls usually have variable acoustics
  drapes can be deployed to deaden the room when
  the audience leaves
• Some halls have adjustable reverber- tion for
  making a good acoustical match between full and
  empty halls
• Have a good reverberation unit on hand to make up
  the differences

14
How well can recordings be matched?

• Two examples
• Full-to-empty house (DSO Mahler 2 McDermott
  Hall)
• Same house, from 1987 to 2001(OSO Feste Romane
  trumpet insert Schnitzer Hall)

15
Session efficiency

• Keep a detailed log of all professional recording
  assignments/activities
• Log all pertinent console settings, mic models,
  channel/pan assignments, EQ, reverb send/receive
  settings, calibration values, on-stage mic
  positions, piano serial numbers, etc.
• Dont forget to list all pertinent personnel,
  including recording, artistic and general
  support technical credits are important

16
How to reduce the time required for a sound check

• In todays commercial world you should be up and
  running in no more than 5 minutes after down-beat
• The industry goal is to get 10 minutes of usable
  recording for each scheduled hour of studio time
  (each studio hour includes 20 minutes break time)
• You must have a template in mind for each new
  project
• These templates define a basic microphone array
  which you determine to be best suited for the
  project, based largely on your experience
• Lets look at some of these

17
Microphone array hierarchy

• Main arrays determine the basic perspective on
  the ensemble
• Spot mics are used for internal highlighting
  (delay often required)
• House mics are used to define, as needed, the
  overall room signature
• Remember, we are still talking about
  direct-to-stereo recording

18
The aim is to provide stereo image specificity
and spatiality in the recording

• Specificity refers to hearing instruments where
  they are placed on the stage
• Spatiality implies that there is a sense of
  location, large or small, in which these
  instruments are played
• This is a tall order for only two loudspeakers in
  the listening space

19
Basic microphone layout for DSO recording of
American in Paris
20
Examples of basic orchestral perspectives

• ORTF pair (cards at 7-inches splayed at
  110-degrees)
• Flanking omnis alone (spaced 1/3 stage width
  about the center line)
• House mics alone (cards at 24 facing back)
• ORTF plus flanks
• ORTF plus flanks and house
• The above 3 pairs plus internal spot mics -- the
  final stereo mix

21
Why use spot mics

• Spot mics are used
• To correct fore-aft balances in the orchestra
• To correct balances for weak instruments (e. g.,
  harp, celesta)
• To correct for spectral imbalances (e. g.,
  basses)
• To add presence (primarily percussion and back of
  orchestra)
• Delay of spot mics always correct, but not
  always necessary
• Always pan spot mics to their natural positions
  in the basic stereo pickup
• Roll off LF slightly

22
Approximate guidelines for delaying spot mics
23
Other mic arrays

• Spaced omnis (good sense of space slightly
  compromised sense of specificity mainstay of the
  early US stereo catalog)
• Decca tree flanks (rich spatial cues hallmark
  of Decca and EMI recording, even today)
• Heads, baffles, etc. flanks (largely
  experimental)

24
Spaced omnis

• These tend to emphasize spatiality, largely
  through in increase in time cues and pickup of
  reverberation
• Number depends on width of ensemble
• Separation of about 12 is the maximum

25
Excellent for organ and large choruses

• To get a broad sound front with emphasis on
  diffuseness, not image specificity
• To produce a big sound suggesting more
  performers than are actually present
• Example of organ pickup
• Examples of chorus pickup
• 2 spaced omnis Piccolo Hear us, O Lord
• 3 spaced omnis Mahler Symphony 2
• 4 spaced omnis Lauridsen O Magnum Mysterium

26
Other microphone arrays

• Decca tree makes use of Neumann M50 microphones
• These are omni at LF, increasing in both output
  and forward directivity at HF
• A flanking pair com-plements the main array

27
Omnis and a pair of subcardioids

• Flanking omnis as before
• Splayed separated subcards at center
• Imaging rich in time cues
• Excellent in not-too-live halls

28
Franssens data on stereo localization
29
The signal path microphones

• Salient specs sensitivity noise floor
  impedance 0.5 THD level pattern integrity
  axial response and limits
• How good can a microphone be?
• Consider the BK 4179 noise floor -2 dB(A)
  however,max level is about 102 dB LP,and
  frequency response is erratic above about 10 kHz.
  Polar response is that of a 1-inch diaphragm

30
The eternal triangle

• Dynamic range, bandwidth, and polar response. Any
  improvement in one of these attributes comes with
  a compromise in one or both of the others
• Dynamic range Make diaphragm larger (compromises
  polar response) Increase spacing and reduce
  tension (compromises frequency response)
• Polar response Make diaphragm smaller (reduces
  output level and increases noise)
• Bandwidth Make diaphragm stiffer (reduces output
  and decreases dynamic range)

31
The signal path consoles vs. stand-alone preamps

• Phantom powering options
• Impedance adjusting
• Maximum input capability
• Noise floor
• Common mode rejection
• Padding/gain adjustment

32
The signal path line level

• Keep as simple as possible
• Output busses vs. direct outs aux outs
• EQ capability
• Monitoring flexibility (line out 2 machines for
  stereo what about surround?)
• Communications interface with stage and conductor

33
The signal path recorders

• Metering setting reference level (typically,
  -14 to -20 dBFS)
• Monitoring all tracks? How often?
• Accurate note taking log sheets
• How many operators are needed?
• The producer should be given a set of DATs and
  complete log indicating PNOs, start/end times and
  score locations

34
Metering confusion

• The various PPMs are all electronic and have very
  fast attack time (10 ms) and slow recovery time
  (3 seconds)
• The VU meter typically has 0.3 second attack and
  recovery times
• In this modern era we need digital meters

35
Audio monitoring

• Loudspeakers vs. headphones
• (producer and engineer should agree beforehand)
• Nearfield vs. large loudspeakers
• (You rarely have room for large loudspeakers)
• Monitoring levels
• (85 - 90 dB maximum is recommended)
• How to calibrate monitor levels
• (a set dBFS level a set LP)

36
What can go wrong? The following things have
actually happened to me

• On-stage video monitor showing a mirror image of
  what was being monitored over loudspeakers. A
  noisy mic developed -- and guess what I did!
• Noisy microphone due to improper seating of XLR
  cable at microphone
• RF noise (not present during morning/afternoon
  sessions but evident in the evening)
• Organ solenoids inducing clicks in mic lines
• Problems with installed house mic wiring (noise,
  polarity, mislabeling) Check carefully before
  using!!

37
Digital timeline I

• Soundstream (1976)
• 14 bits at 32 kHz
• Eventually 16 bits at 50 kHz
• Stereo recording segments 20 minutes
• Three editing centers worldwide - a real
  bottleneck

38
Digital timeline II

• Early 80s JVC and Sony U-matic models plus
  editing systems
• SPDF and AES-EBU standards
• 3M and Mitsubishi reel-to-reel
• DAT machines
• 3rd party converters
• Evolution of DASH and Pro-digi standards

39
Digital timeline III

• Rise of DAWs and interface standards
• Rise of DSP and internet based tech
• Rise of tape based MDMs bit-splitting
• Rise of disc based MDMs
• Eventual relegation of reel digital recorders to
  dinosaur status?
• Multitrack analog recorders may outlive us all

40
Surround sound consumer media

• Dolby AC-3 on DVD-V
• DTS encoded Redbook standard CDs (with decoder)

41
New consumer media

• DVD-A versus SACD,or, PCM versus DSD
• PCM sample rate/bit depth related
• Signal is quantized at each sample sinx/x filter
  reconstructs the signal exactly, if it is below
  the Nyquist rate. Noise shaping improves
  resolution at low signal levels
• Easy access to EQ and 3rd-party DSP equipment
• DSD 1-bit/high sample rate
• Signal is defined as the moving average of many
  1-bit approximations. Noise shaping essential for
  system operation
• Access to EQ by way of PCM chain used only when
  necessary

42
Surround sound rationale

• Stereo provides
• image specificity at the sweet spot
• conveys a sense of spatiality
• Surround sound
• frees the listener from the sweet spot
• adds a new dimension of envelopment
• Use more channels if
• More image specificity is needed, or
• Larger audience area is required

43
Surround mic techniques

• The Soundfield microphone samples air pressure
  and particle velocity in 3 dimensions
• These four elements can synthesize first-order
  cardioids aimed in any direction
• These directions normally correspond to the
  locations of loudspeakers in the listening space

44
A seven microphone array

• Presented at 2000 AES Convention, LA
• Loudspeaker locations should match mic angles,
  but normal 5.1 performance is quite good, with
  vertical signals summed and fed to the 5
  loudspeakers at reduced level

45
Schoeps KFM 360 array

• The array, as seen from above, consists of both
  front and back pairs of MS microphones
• Patterns, as shown, can be chosen differently for
  front and back
• A front-center channel can be derived from the
  front L/R pair
• Control unit provides additional processing,
  including delay of back channels

46
The SPL (Sound Performance Lab) array

• Telescoping arms can be adjusted
• Microphones can be rotated and patterns adjusted
• Note resemblance to Decca tree

47
Frontal arrays I

• These are 3-channel arrays that optimize frontal
  pickup
• Goal is to minimize double phantom images
• Used in conjunction with a back pair of
  microphones

48
Frontal array II

• Proposed, but not yet designed
• Will provide 15 dB isolation between A and E
• Polar equation r (.5 .5cosq)(cosq)