 |
| Multichannel
Sound Mixing Practice for Broadcasting |
| Abstract |
It will describe the procedures of producing
multichannel surround sounds based on NHK's experiences in the
production of HDTV programs and radio dramas. Specific methods
are reviewed in the following sections regarding:
1) Acoustics of the control room;
2) Surround sound design methodology;
3) Mixing considerations, and
4) Drama and music production examples. |
| |
| 1 Design of control room acoustics
and monitoring environment |
| For the optimal mixing of surround sounds
and their precise reproduction in the home, more appropriate
guidelines for the design of the control room environment, and
its acoustics in particular, need to be drawn up. Table 1 shows
the guideline specifications proposed by HDTV Multichannel Sound
Study Group (HDTV-MSSG) which operated from 1992 to 1995. These
specifications were intended for sound mixing work in small-capacity
studios, an environment quite different from that for motion
picture sound production. Table 2 lists the test items at that
time. Table 3 shows the specifications of NHK's surround sound
post-production studios, mix-down studios, music recording studios
and TV studios built and used since 1987. We started with rather
vague ideas and have tried various design approaches, leading
to the primary design methodology in 1999 as follows: |
(1) Each channel must
undertake even work in accordance with the number of channels
in the multichannel system. This means that increased sound
field homogeneity is needed for the control room.
To achieve such homogeneity, NHK has designed the control room
to absorb sound in its front section and evenly diffuse it at
the rear ceiling and sidewalls. We adopted this idea to secure
a broad and homogeneous sound field over the entire band in
the optimal listening area, which is different from the conventional
method of combining different acoustic conditions, such as live
end and dead end. The control room's interior is an irregular
heptagon that is right-left symmetrical. |
| (2) The front section
of the floor must be made of sound absorbing materials to prevent
the primary acoustic reflections of L-C-R speakers from the
floor. In the conventional construction, monitor speakers are
built into high-rigidity walls, but this construction tends
to generate rear reflections in multichannel surround sound
production and disturb the frequency characteristics near the
mixing area. Therefore, the present front surface has a structure
that transmits sound. |
(3) Diffusers must be
used for homogeneous diffusion of medium and high frequencies.
The installation location will depend on the studio's capacity
and space arrangements.
In studios built so far, the diffusers are installed on the
ceiling, side walls and back wall. |
| (4) The control room's
own low-frequency processing function and unnecessary resonance
must be inhibited since the control room has higher Low-frequency
reproduction ability than conventional rooms. Sound traps are
installed in the air behind the speakers for Low frequency processing
near the front side. The walls behind monitor speakers are made
of burnt bricks to increase the articulation of low-frequency
sound. |
(5) Fire-resistant stone
blocks and 'Acoustone'TM porous heavy bricks are used to facilitate
low- and medium frequency processing in the studio's production
area.
This improves low-frequency absorption ability and eliminates
unnecessary resonance with the studio's base structure. |
| (6) The studios have common
parameters: RT = 0.2 to 0.3 sec. with a 'dead' tendency; NC
= 15 or less. The front L-R speakers are installed 3.6 to 3.2m's
apart in consideration of the alignment of L-C-R speakers compatible
with stereophonic production. The distance to the effective
acoustic center, which is at the apex of an equilateral triangle
formed with the line connecting L and R speakers as its base,
is 3.4 to 3.5 m's, slightly behind the mixer seat. This arrangement
is intended to enlarge the listening area. When more than one
pair of speakers are installed in the rear section, one pair
is at 60dgree and the other at 120dgree. When a single pair
is installed, it is set at 130degree, somewhat laterally. |
| (7) In fact, it is difficult
to install all speakers at the same height in an actual production
studio as specified in the ITU-R Standards, which place greater
emphasis on the reproduction environment. For this reason, allowances
must be taken into account and, particularly in the case where
the same listening area, as that in the front section cannot
be secured in the rear section, an electrical correction method
shall be used. Figure -1 shows an example of the time-base correction
of monitor speakers at the mixer seat in NHK's HVD-520 post-production
studio. Because of this, the same distance is maintained on
a plan-view drawing between respective channel speakers and
the acoustic center in the control room. A cross-sectional view
shows that the speakers are some 30% higher than the control
room. This configuration, however, is a practical way to ensure
a broader listening area because there are also production staff
on duty in the rear. Since the monitor balance of each channel
may vary due to secular changes, it should have a trimming function
to enable fine adjustments to be made. |
|
 |
| Figure1:SUPEAKER
TIME ALIGNMENT |
|
| (8) No LFE speaker dedicated
for low frequency processing is installed. The CA-421 screening
room is only intended for reproduction. In the use of LFE components,
the output is chiefly delivered to the L and R speakers. The
LFE components must be retrieved on the reproduction side as
necessary and matched with the main-channel components. |
(9) All the front speakers
must be of the same model, while small types of the same model
must be used for surround sound production due to physical restrictions.
This unifies the quality of sound in all the channels. |
 |
| 2 Primary positioning
for sound design for TV dramas |
| Figure 2 shows the positioning for dialogue,
music and sound effects for HDTV dramas. Typical patterns of
surround sound design are shown in Figure 3. |
| |
 |
| Figure2:SURROUND
COMPONETS PLAN |
 |
| Figure3:SURROUND
SOUND DESIGNs |
|
|
| 2.1 Recording of dialogue |
Since dialogues that accompany images
are positioned in principle in the center channel, they may
be recorded as monaural even when recorded on the spot. Dialogue
spoken in a crowd scene or party scene where the voice needs
to generate a sensation of expanse are recorded in stereo as
material to be processed in post-production.
If a portable multichannel recorder is developed in the future,
it will become possible to record 3-2 surround material locally.
Sound collages and dialogue used as special effects may be positioned
in accordance with producer's ideas. |
| 2.2 Foley and sound effect materials |
| Foley may be recorded in principle as monaural
because it is positioned in the center channel. It is recommended
that the material be recorded in three L-C-R channels and processed
if a sensation of expanse, like doors being fully open over
the whole screen, is required. In case of using the material
for special effects, it can be positioned so that all the channels
may be used efficiently. Basically, the front speaker configuration
for sound effects may be any of L-C, C -R, or L-R combination,
while the rear alignment may be any combination of SL-SR, SL
only and SR only. The size of sound image and the sensation
of expanse should conform with image shots. |
| 2.3 Music |
| As regards scoring music, usually the front
L-C-R speakers are used for main music and the rear configuration
is for ambient components. The center channel is primarily used
for source music in dramas and all the channels effectively
undertake ME. |
|
| 3 Surround sound designs |
| 3.1 Six basic patterns of surround
sound design for dramas (see Fig-3) |
| (1) Surround ambience
|
| This is the most fundamental surround sound
design for either music or dramas. For music, it produces an
environment space behind the audience so that they perceive
a stronger sense of reality or atmosphere. In dramas, the environment
sounds enable the audience to better perceive how the story
is proceeding. The difference between drama environment sounds
and environment music is that surround components used for the
former are not necessarily those recorded simultaneously. |
| (2) Fly-over |
| As suggested by the name, specific sound
flows longitudinally between the front and rear sections of
the studio. A sharp snap of sound effect adds a strong impact
to the scene. |
| (3) Whirlpool |
| The audience is thrown into a spiral whirlpool
of sound and so feels as if the place is swinging in every direction.
|
| (4) Proceeding sound
field |
| Sounds which may predict what is about
to happen in subsequent scenes are reproduced, not merely generating
a sensation of reality or feeling of unity as with the method
in Item (1) above. The sounds produced here must be short, have
a punch and allow the audience to guess what's coming next. |
| (5) Sound shower from
above |
| A shower of sound comes from above the
audience. It is theoretically impossible to reproduce vertical
relations using any configuration of the current horizontal
6-channel speakers. The method, however, makes use of the physical
advantage of surround speakers, which are typically installed
higher than the audience seats. |
| (6) Big, closer sound
feeling |
| Most of the sounds come horizontally instead
of from above; the main components are reproduced with the front
C-channel, and supplementary components with the L-R/SR-SL channels.
This method is useful for emphasizing a specific human voice
in dialogue or monologues or for big sound effects representing
gunshots and explosions. In this method, sound can be boosted
larger than that reproduced from a single channel. Using more
than one channel can strengthen the drive to a higher level
than representing all with a single channel, while securing
the peak margin. |
|
| 3.2 Three basic patterns of surround
sound design for music(see Fig-4) |
| (1) Stage layout |
| The main music components are positioned
in the front section, while spatial information, such as reverberation
of the hall and indirect acoustic components, are reproduced
in the rear. |
| (2) Discrete layout |
| This layout is not intended to reproduce
theatrical performances, but is suitable for the musical representation
of something unrealistic by actively using more of the assigned
channels. It is aimed at the front section but reproducible
sound can be laid out freely over the audience's surroundings.
|
| (3) Omni-directional layout |
| The audience's front axis is not fixed
so that they can receive the sound from all over. Music artists,
such as Japan's Isao Tomita and Britain's Allan Parsons, have
created 'sound walls' by making good use of such Omni-directional
acoustic space. |
 |
| Figure4:3MUSIC
SURROUND DESIGNs |
|
|
| 4 Considerations for
surround sound mixing |
| In this section, we discuss the following
subjects and review usage examples by rule of thumb: |
1) Optimal monitor level settings and practicable
monitoring circuits;
2) Operations of the center channel; and
3) Application of heavy bass components (LFE). |
| 4.1 Monitor level alignment |
| Since the production of broadcasting programs
involves a variety of areas including motion pictures, large
monitor level settings are corrected at 85 dB per channel in
pink noise in order to make them compatible with movie sounds.
Medium and small speakers are also corrected at such monitor
levels as 82 dB, 80 dB, 78-76 dB/channel so that monitor levels
can be mutually switched over where necessary to optimize the
balance of sounds reproduced in the home. NHK's CR-602 studio,
remodeled in March of 1998 to produce radio dramas, is equipped
with various digital oscillators, thus facilitating the correction
of speaker levels. Figure 5 shows monitor speakers in operation
and typical monitor balances according to size. |
| |
 |
| Figure5:REFERENCE
MONITER LEVEL |
|
|
| 4.2 Mix-down and dynamic range
control |
| A monitoring function compatible with both
surround mixing and 2-channel stereo mixing is important for
maintaining compatibility in mix-down. This function must allow
the 2-channel stereo balance to be checked during surround mixing.
Figure 6 shows the mix-down circuitry installed in the CR-602
studio. There are large differences in the audiovisual environment
and reproduction levels between 3-2 surround mixing and 2-0
stereo mixing. To resolve these problems, a processor capable
of audibility compensation and appropriate dynamic range compression
will be necessary. Future developments of multichannel music
software production rather than drama production will increasingly
require such processors. |
| |
|
| 4.3 Representation of the center
channel (See Figure 7) |
|
 |
| Figure7:HOW
TO USE CENTER CH |
|
|
| a) Positioning for the phantom
center |
The same as with the conventional 2-channel
stereo method, this position is used to emphasize high-quality
sound blending between the front L-R speakers or when no articulates
sound image is necessary.
This is enough provided there is no pressing need to use the
center channel or the speakers in the reproduction area have
narrow intervals. It should be noted, however, that phantom
center sounds would form a diffused sound image and would sound
to be out of balance if the image screen size was 50 inches
or larger or the L and R speakers were placed 2.5 m or more
apart. |
| b) Positioning for the hard center |
| This is used to clearly distinguish a real
sound image from the images of other channels or when an articulate
sound image of center position, instead of a phantom center
image, is needed. This has an advantage that it allows a mix-down
using coefficient, which approximately match theoretical values.
This method helps stabilize overall positioning when used for
the main vocals or a specific solo instrument as well as for
narration and monologues. Even if out of the sweet spot area,
the sounds would largely be kept in equilibrium. |
| c) Mixture |
| By combining the foregoing two methods,
this approach sets positions by combining the specific hard
center in the center channel and the supplementary phantom center
between the L and R speakers. It is useful for smoothly blending
sound images in the whole front together while articulately
placing center components in position. For this purpose, cross
talk between the L/C/R channels must be controlled using what
is called the divergence function. Typical examples are as follows:
for monologues, position the main sound in the hard center channel
and also in the L-R channels but with the level reduced by 3
to 4 dB, or avoid the risk of over level by concentrating the
bass and kick drum parts only in the hard center channel. The
latter example requires special care in mix-down to prevent
any difference in the balance between surround sounds and 2-channel
stereo sounds. |
| 4.4 LFE control |
| Heavy bass components of 120 Hz or under
offer a useful means of representing motion picture and drama
sounds. In general, music performed with acoustic instruments
contains little heavy bass components unless bass drums, 'cannons'
and contrabasses are played in concert. Adding a slight flavor
with a limited amount of heavy bass components may be enough
except when they are intentionally added in bulk for some purpose. |
|
| 5. Production |
| 5.1 HDTV |
Here is an example of actual surround
sound mixing. Figure 8 shows the audio system of the HVD-520
studio. The main channel console AMS/NEVE Logic-2 has a maximum
of 8-channel output and can prepare various monitoring environments.
For sound, the following mixing buses should be configured in
the console:
L-R-C-SR-SL
L-R-C-S (-2 dB* SR + SL) *Variable by content and material
LT = L + 0.7C + 0.7S
RT = R + 0.7C + 0.7S or Dolby encoding LT/RT
Since the HVD-520 studio's console monitor system allows free
selection from 8-system monitors, the following combinations
are assumed to be prepared:
Monitor A = L-C- R
Monitor B = SR-SL
Monitor C = S (L +R)
Monitor D = LT/RT |
| |
 |
| Figure8: HD520
system |
|
|
| These combinations can be freely added
together for monitoring, so well balanced mixing on three systems
is possible by switching A-D for pre-mixing or final mixing
procedures. As shown in Figure 9, materials are taken by DAW
and stored in magneto-optic discs. |
| |
 |
| Figure9:HD0TV
DRAMA LAST BULLET 1995 MIXING FLOW |
|
|
| The Logic-2 console incorporates an audio
file capable of 16 channels which is responsible for sound rectification,
ADR and base noise, as well as for dialogue including monologues
and narration. It is recommended that music be also stored in
this file if there is sufficient track space to facilitate the
delicate timing procedure. |
| For the surround mixing of sound effects,
pre-mix output from the Fairlight recorder that stores material
and records finished sound effects on the PCM-334 8. In the
final mixing, mix dialogue, music and sound effects, in that
order, using the Logic-2's automation mode. If everything is
all right, record them by segments and then check and retouch
respective scenes, as needed, by monitoring them by blocks.
|
| Finally, three formats are simultaneously
recorded on Tracks 9 to 16 of the PCM-33 48, as shown in Figure
10. Lay them back in the master VTR to complete the whole mixing
procedure. Optimal tracks may be chosen from the master VTR
in accordance with the medium and format used. |
| |
 |
| Figure10:
LAST BULLET TRACK-SHEET |
|
|
| 5.2 Music.Ambient sound recording
~ Fukada-Tree microphone arrangement |
What is the best way to record spatial
environment components to reproduce natural sound space from
them? In recent years, the Omni-directional microphone has been
used as the main microphone in many cases of stereo recording
because it can take in rich reverberation components. However,
if surround microphones are added to this stereo method for
surround recording, an excessive amount of reverberation components
may produce sounds such that the environment is overly enhanced.
For reverberation component energy containing spatial information
to be perceived as natural by the audience, the total sound
levels in both stereo and surround-recording procedures must
be equal. Recording reverberation components with an Omni-directional
mike will result in the reproduction of rich reverberation.
However, since the mike picks up delayed direct sound components
as well, rear sounds will become unnatural and the sound positioning
in the front will be unstable.
Installing surround microphones apart from the sound source
would reduce direct sound energy, but would also lower reverberation
energy and thus aggravate the reverberation texture.
|
| To resolve all these problems, Fukuda has
devised a new microphone arrangement plan. The plan, in order
to clarify mike positioning and record rich environment sounds,
suggests the installation of unidirectional mikes for main and
environment sounds at around the critical distance where the
concentration levels of direct sound and indirect sound components
become equal. The Fukuda plan, which basically consists of seven
mikes and supplementary environment mikes, may be indicated
in the form of a tree as shown in Figure 11. The microphones
marked LL/RR are Omni-directional, while all the other microphones
are unidirectional. For recording an orchestra performance,
for example, the main mike and environment mikes are set at
2 m or smaller intervals. If they must be installed further
apart; insert a delay in the front main microphone to adjust
the time axis. |
| |
|
| Figure 12 shows the sound images which
may be covered with this mike arrangement. The LL/RR microphones
on both sides are intended to pick up the orchestrated sound
expanse and a smooth sound envelope covering the front and rear
sections of the hall. The unidirectional mikes are B&K 4011
and the Omni-directional mikes, B&K 4006. The configuration
of the tree can vary depending on the hall's acoustic characteristics,
while the mike intervals may be changed conforming to the orchestra's
size and formation. The horizontal directivity angles of microphones
are important in order to pick up the orchestra's sounds precisely.
They are usually installed horizontally. This mike arrangement
is chiefly intended to record a sensation of expanse of the
hall. The random energy efficiency of the unidirectional mikes
is 4.8 dB lower than that of the Omni-directional mikes, which
provides appropriate separation between direct sound and indirect
sound. Since the unidirectional mikes have a distance factor
of 1.7, if they are installed at the same positions as the Omni-directional
mikes, they will offer a close and 'dry' feeling. With this
taken into account, the microphones are placed at adjusted distances
from the sound source. Since unidirectional mikes come in a
variety of types with different characteristics, their positions
vary accordingly. Examples include the Neumann M-50 which is
Omni-directional but actually has directivity characteristics
from 5 kHz and up, as well as the Schoeps MK21 capsule featuring
a wide directivity range and the B&K 4006 which has directivity
when used together with an acoustic pressure equalizer. All
these mikes are compatible with the tree system. The center
mike is set for a balance level lower by 6 dB and 4 dB than
the L and R microphones, respectively. Fig13 shows actual example.
|
| |
 |
| Figure12:FUKADA-TREE |
|
| |
| |
 |
| Figure13:FUKADA-TREE
97’ SUPER CONCERT |
|
|
| 5.3 Radio drama ~Concept of sound
design for 'Yume-no-Hitsugi' |
This section introduces a sound design
example called 'Yume-no-Hitsugi' or Coffin of Dreams, a 1998
Radio drama. The design concept is itemized below:
* Two different worlds are depicted: a warped jungle where God
lives and sobers everyday life. Each line of information inserted
in the scenario is supposed to be converted into full-range
sounds from fine to dynamic. One of the issues is how to effectively
use the heavy bass area or LFE with frequencies of not higher
than 120 Hz.
* Another challenge is how to successfully represent totally
different environments, the inside of a closet or elevator and
a vast jungle, in a 3-2 surround sound field.
* The 3-2 surround sound field also requires more sophisticated
music representation than usual. So, Mr. Fukada, the NHK mixer
in charge of verifying his own 'Fukada Tree' mike arrangement
plan, undertook the mixing of the music score. |
To realize images envisaged by the composer,
we first carefully considered which sound should be resounded,
the desired sound-origin directions and the timbre balance of
instruments, and then determined the positions of instruments.
We made a basic sound field in compliance with the Fukada Tree
and created a sensation of expanse by positioning a harp and
vibraphones in the rear. The percussion was separated from the
other parts with a sound barrier to prevent sound overlap with
the Tree. This was particularly necessary to prevent sound confusion
with the Tree when producing something like sound effects in
the rear.
A prepared piano was fine-tuned to its optimal condition in
cooperation with the composer. In consideration of the music
coming down to a background level, we carefully mixed music
sounds so as not to attenuate their reverberation energy and
bass power. |
| * The method of dealing with dialogue in
dramas in the L-C-R front in the 3-2 surround recording procedure
is still a problem. Also, how to represent center components,
as in the case for 3-2 music, and possible microphone arrangement
for the recording of dialogue must be clarified. In this 'Yume-no-Hitsugi'
example, we positioned ordinary conversations in a phantom position
in between the L-R channels and monologues and narration in
a hard position in the center C channel. |
| The following sections introduce the flow
of production processes from recording of dialogue, to mix-down.
|
|
| 5.3.1 Recording of dialogue |
| The dialogues of the drama were recorded
in the CR-601 studio that was built adjacent to the CR-602
studio in 1997. The new studio, which is dedicated for
recording dialogue, consists of three booths, each incorporating
a main area to reproduce virtual space, a semi-dead room
and a full-dead room, respectively. An X-Y pair of U-87Ai
mikes, which have a versatile quality of sound, is usually
used to record dialogue. This time, however, we employed
the U-87Ai for monologues and a pair of M-149 new tube
mikes for dialogue. Conversations over the telephone were
recorded using a SM-58 microphone with the other party
in the semi-dead room. The dialogues of the God of Forests
were recorded through an MKH-40P mike placed in the full-dead
room. The recorder was Fairlight MFX-3 Plus using a M.O
disc, which was then edited, in an editing room. |
| 5.3.2 Foley recording |
| We took special care regarding two
points: one was to record at a good S/N ratio weak sounds,
such as a teacup being put down on a saucer or the rustling
of a dress, and the other was to record materials that
could add to effective texture in later sound effect pre-mixing.
For these procedures we mainly used an expander and Program
8-1 pitch shift included in Lexicon 480. Materials to
be added later with LFE's heavy bass components were somewhat
reinforced in their low-frequency area during recording.
Typical examples were a door or a closet being opened/closed,
a monster bird flying up into the sky, a man running up
a steel staircase, etc. Sounds representable in the front
R-C-L as a whole, e.g. a door being opened/shut, were
recorded through an MKH-60P mike into the monaural center.
Also, we positioned the expanding and resounding door
sounds through a pair of U-87Ai mikes into the L-R channels.
We also represented the flying monster bird appearing
at the start of the drama by recording artificial sounds
and synthesizing them in pre-mixing and produced a sensation
as if the bird emerged from behind the audience using
a joystick panning method. Delicate rustles of bedclothes
were recorded with a very sensitive MKH-60P mike in a
close-up position. The sound of the alarm clock that resounded
in the hero's bedroom was made from 5 alarm clocks with
different tones so that the sounds could be positioned
in the whole surround channels in pre-mixing. |
| 5.3.3 Recording of music
in CR-509 studio |
Music, put through a 3-2 mix-down
process, was then recorded on the Fairlight MFX-3 Plus
recorder using a magneto-optic disc. It was downloaded
in the CR-602 studio.
Fig 14 shows scoring music recording. |
| |
 |
| Figure14:FUKADA-TREE
97’ SUPER CONCERT |
|
| |
| 5.3.4 Pre-mix |
| Pre-mixing and subsequent processes
were performed in the CR-602 studio. Table 1 shows track
assignments for dialogue, music and sound effects. The
introduction, lasting about 8 minutes from the start,
incorporated a variety of sound designs. They are explained
below. |
| FAME NO-1 |
|
|
FAME-2 |
|
| 1 |
DIALOGU-A |
L |
MUSIC |
L |
| 2 |
|
C |
|
C |
| 3 |
|
R |
|
R |
| 4 |
DIALOGU-B |
L |
|
SL |
| 5 |
|
C |
|
SR |
| 6 |
|
R |
|
|
| 7 |
DIALOGU |
SL |
|
|
| 8 |
|
SR |
TEMPMIX-1 |
L |
| 9 |
SFX-A |
L |
|
C |
| 10 |
|
C |
|
R |
| 11 |
|
R |
|
SL |
| 12 |
SFX-B |
L |
|
SR |
| 13 |
|
C |
TEMPMIX-2 |
L |
| 14 |
|
R |
|
C |
| 15 |
|
SL |
|
R |
| 16 |
|
SR |
|
SL |
| 17 |
SFX-C |
L |
|
SR |
| 18 |
|
C |
FINAL-MIX |
L |
| 19 |
|
R |
|
C |
| 20 |
|
SL |
|
R |
| 21 |
|
SR |
|
SL |
| 22 |
|
|
|
SR |
| 23 |
|
|
|
Lt |
| 24 |
|
|
|
Rt |
|
| |
| table1:track
assignments for dialogue |
|
| 5.3.5 Sound design for the
jungle |
| We will describe how these acoustic
scenes were assembled following the illustrated continuity.
The scenes, coming at the start of the drama, had to impress
the audience. We imagined the following visual scenes:
|
| A close-up view of a big parrot resting
on a tree somewhere, but it is not known that it's a jungle.
The parrot flies up going from SR to FL. |
| The camera sinks with a jungle unfolding
in a long shot. The camera on a dolly further sinks until
it catches a girl walking through a bush with a cat in
her arms. The camera sinks to the girl's foot level and
follows her moving feet. FL-SL and FR-SR lateral sounds
are also moving together. The camera rises and pans to
catch the God of Forests who spreads his dark wings saying
in a loud, resounding voice, 'Buenas tardes!' Fascinated
by the voice, the girl sighs. |
| Sound materials used for this were
the flapping of wings; the girl going pit-a-pat through
the bush, trotting and then halting; the big bird flapping
up into the sky; and the rustling of grass and plant leaves.
|
| To blend these sounds with sounds
in the jungle, we prepared four types of stereo sound
as the jungle sound base, collecting close-up sounds of
shouting monkeys, twittering birds, running water and
rustling grass from the Imai Library. God's resounding
voice was made by putting an original monologue in the
center channel then reducing it 5% on the L-channel side
and 3% on the R-channel side using the pitch change program
of Lexicon 480 and putting the results in position. To
produce a sensation of entire vastness using the Big Voice
program of DSP-4000, we positioned another sound effect
component in the SR and SL channels. Also, we added sub-woofer
components to the FR and FL channels using a sub-harmonic
synthesizer to produce a heavy bass feeling in the front. |
| 5.3.6 advantage of 3-2 mixing
procedure |
| Production of the radio drama revealed
the following features of the 3-2 mixing |
1.The presentation capabilities of
the 3-2 discrete layout are attractive also for
Radio dramas because all the channels have equal frequency
characteristics.
|
2.The center channel should be used
for parts of dialogue, such as narration and
monologue, while ordinary conversation should be positioned
in a phantom
position. |
3.The production-monitoring environment
is fairly good if sub-woofer components,
called LFE, are distributed to the front L-R channels.
|
4.Switching the monitoring environment
between 3-2 and Dolby 3-1 in pre-mixing can
maintain an approximate balance. Sound stage deviation
toward center due to
Dolby steering logic and sound field bleeding on account
of crosstalk must be
tolerated. |
5.During production, monitor levels
were set for 85 dB/channel in pre-mixing and 82
dB/channel in mix-down. No problem is likely at these
levels for reproduction in
the home |
|
|
| Conclusion and Pending Issues |
| We have discussed the application of a
multichannel surround sound recording method for drama and music
production. Development of such recording and production methods
has just started in Japan and overseas. However, the following
issues remain: |
1)The methodology has
evolved from 2-channel stereo to 3-channel stereo and then to
3-2 surround. How many channels will eventually have to be installed
for ideal
spatial reproduction? |
2)What effects do physical
rear reverberation and indirect sound components have on
psychological sound? |
| 3) What are the rear equal
loudness characteristics? |
| 4) Does the sound field
of multichannel music reproduction have a broad sweet spot?
|
5) What level of deviation
from the appropriate arrangement should be allowed in the
home music reproduction environment? |
| 6) Is there any adequate
down-mix method to secure compatibility with stereo? |
| 7) How to ensure an appropriate
dynamic range for various audibility levels |
| 8) Development of a practical
method of using the center channel for music |
9) Is there any way of
effectively using the sub-woofer band, called the LFE channel,
for music? |
| We believe that multichannel surround sounds
offer tremendous potential as a presentation means since broadcasting
is not only an entertainment medium but also can be used to
create new arts and other possibilities. In conclusion, we would
encourage software producers to continue their work. |
|
| REFERENCES |
| 1 S.Yoshikawa ［HDTV Multichannel Sound
Study Group］ 1996 AES Copenhagen [Proposal for the Specification
of Control Rooms for HDTV Multichannel Sound Program Production] |
2 A.Fukada 1996 AES Copenhagen
[Surround Mixing technique on Dolby&3-1 surround sound for
live broadcasting] |
3 Mick.Sawaguchi 1996 AES Copenhagen
[HDTV Drama Multicannel Sound Productio with 3-2/3-1/2 CH] |
4 A.Fukada 1997 AES N.Y
[Microphone Techniques for Ambient Sound on a Music Rcording] |
5 N.Kurihara 1993 AES N.Y
[The First Full Scale Surround Sound Sweetening Room for HDTV] |
6 H.Suzuki Sam Toyoshima H.Shinbara 1993
AES N.Y
[Study on Optimum Rear Loudspeaker Height for 3-1 HDTV Audio] |
7 Y.Tkahashi A.Fukada H.Takada H.Shumid
T.Jung 1997 InterBEE
[The Frontiers of Multichannel Music Production] |
8 T.Wakatuki 1990 J.Acoust.Soc Jpn 11-3
[A basic study of acoustical design for a 4-CH sound control
room] |
9 G.Steinke 1996 AES Copenhagen workshop
[Surround Sound -The New Phase An Overview] |
10 Mick.Sawaguchi A.Fukada 1998 AES Munich
[New Radio Broadcasting Studio with Disc Based Mixing Facility
for Producing Multi Channel Audio Dramas] |
11 Mick Sawaguchi M.Fushiki 1991 AES Detroit
[Surround Broadcasting in Japan] |
12 Akira Fukada T.Kamekawa S.Akita H.Kikuta
1997 AES TOKYO
[Our Challenges for Multichannel Music Mixing&the subject
of expression] |
13 K.Hamasaki 1991 17th ITS Montreux
[How to handle Sound with Large Screen] |
14 Mick.Sawaguchi Akira Fukada 1999 IBC
Amstrdum
[Multichanel Sound Production Practis for Broadcasting] |
|
|
|
