Jag strövade runt i min dator och hittade detta inlägg på Usenet (rec.audio.high-end) som jag inte vill undanhålla er. Inlägget är skrivet av Gabe Wiener, som jobbade i inspelningsindustrin då, runt 1996. Speciellt intressant tycker jag hans beskrivning av "mastering" är. Jag har fetstilat det stycket nedan.
Gabe skrev:> Could somebody post here the complete recording process from
>microphone to pressing. I'm curious about the ins and outs of the
>whole process. What does it mean to master? And why do you have to
>do it? If you could, discuss what teh differene is between two track
>and multi-track recordings. How is it all put together? Stuff like
>that.
Gulp. You might want to order in a meal. This may take awhile...
For the most part I will stick to the recording process most-commonly
used for audiophile quality recordings today (i.e. high-resolution
digital minimalist recording), but I'll put in a few sidebars about
how the process differs for variants such as analog recording and
editing, or vinyl pressing.
The first order of business is to find yourself a venue. The general
guideline is that you should record a particular genre of music in a
place where you'd ideally want to attend a concert of that repertoire.
The ideal venue should be quiet and well-isolated. An otherwise
perfect venue is useless if it's over a subway line or across the
street from a playground.
Audiophile recordings, particularly those of classical music, are
recorded on location rather than in studios. Studios are fine for
creative recordings (i.e. pop music) where one is using recording as a
creative palette to make sounds that do not occur naturally. But when
one records with the audiophile ethic in mind, the general aim is to
re-create a musical event, not create one. To re-create a musical
event, one should ensure that the event takes place in the best
possible acoustical surroundings.
Next, microphones are chosen to render the pressure-waves of sound as
electrical undulations. There are many different types of microphones
and many different makers, each one of which has a wholly different
set of performance parameters. My ideal microphone adds no editorial
perspective of its own, and serves only as a transducer in the
original meaning of the word.
There are many ways to set up microphones so that they produce
stereophony. Stereophony (Greek for "solid sound") refers to any
technique for creating audio with a three-dimensional quality to it.
The stereophonic effect is, in my opinion, best achieved using
techniques that provide auditory cues based both on intensity and
time-arrival information. Intensity refers to techniques whereby two
directional microphones are placed in proximity to one another while
pointing in different directions. Sound originating from one side
appears louder in one channel than in the other. Time-arrival refers
to techniques whereby spaced microphones are used, and a stereophonic
effect is created because of differing arrival times of musical
signals.
Generally, intensity stereo produces seemingly-accurate soundstaging
with strong localization. Time-arrival produces a pleasing sense of
ambiance and warmth, but lacks the accurate localization of intensity.
The Frendh and Dutch broadcasting agencies (ORTF and NOS respectively)
have each proposed hybrid techniques that emply both intensity and
time arrival information using directional microphones. Similar
approaches using baffled omnidirectional microphones have been
suggested in the writings of Jecklin and Schneider. In 1986, Guenther
Theile working at the Institut fur Rundfunktechnik in Munich proposed
a spherical microphone employing omnidirectional capsules.
Each type of stereophonic array has different requirements for
placement in order to replicate a convincing soundstage. Microphones
based on pressure transduction (e.g. spaced omni, Theile sphere,
Jecklin disc) must be placed close to the musical ensemble because the
capsules are equally sensitive to direct and reverberant sounds.
Those techniques based at least in part on pressure gradient
(velocity) microphones need to be placed further back, since the
directionality of the mics means that they're not as sensitive to
reverb, and thus the sound needs more real estate in order to blend.
Rough placements of microphones can be worked out trigonometrically so
long as you have an understanding of how sound behaves and how a given
pattern creates stereophony. But in the end, the recordist hones the
placement by ear, making the aesthetic judgment as to what sounds
best.
Next, the signal must be preamplified immediately to prevent
degradation over the inevitable cable run from the venue to the
control room. As a general rule, audiophile recordings preamplify
right at the base of the microphone stand and send the signal back at
line level. In more conventional pop recording, the signal is brought
back at microphone level and preamplified in the recording console,
often with deleterious effects on the audio.
Upon receipt in the control room, the signal is fed directly into
an analog-to-digital converter whereupon it emerges as a clocked
bitstream. This bitstream is what is ultimately stored to tape
or hard disk.
We should pause for a moment to examine the various permutations that
can occur here. In an analog recording session, the preamplified
signal would be stored to analog tape, usualy 1/4" or 1/2" 2-track.
If this were a multi-mic session and not a minimalist stereo session,
the preamplified signals would either be routed through a mixing desk
where they would be panned and mixed down to two tracks, or else would
be stored on a multi-track tape machine, either analog or digital.
Multi-track machines come in all varieties. The most ubiquitous are
the 24-track 2" analog machines that have been a staple in pop music
for two decades. Among the audiophiles, multitracks are rarely used,
as direct 2-channel stereo is the way of things. Most audiophile
recordings that employ >2 mics will mix direct to two-track in the
field. A popular machine for ultra-high-end audiophile recording is
the Nagra-D 4-track 24-bit digital machine. Tascam and Alesis both
make a new generation of cheap 8-track 16-bit digital recorders, but
the converters in these are nothing to write home about, so when these
are used at all in audiophile recording, they are always used with
external converters. Further, Rane and Prism make adaptors that let
one take the 8 channels of 16 bits and regroup them as 6 channels of
20 bits or 4 channels of 24 bits, thus achieving the same performance
as the Nagra-D at a somewhat (though not substantially) lower cost,
albeit with a great deal of multi-chassis clunkiness.
Some recordists have decided to skip the tape idea altogether and to
instead bring a digital audio workstation to the sessions, recording
directly to the hard drive. Other manufacturers such as Studer and
Genex are manufacturing digital recorders that take those
magneto-optical packs, such that one could load the pack directly into
the workstation.
There is other gear present at a recording session, such as a control
room monitoring system of any number of flavors, and a talkback system.
The latter is usually an intercom with a set of lights such that the
control room personnel can talk to the musicians and instruct them with
lights what to do (begin the next take, stop playing, stand by for a
comment, etc).
Once the session tapes are done, it's on to editing. Nowadays, one
copies the data from the digital tapes directly to a hard drive for
editing and assembly. Without a doubt, the most popular editor for
audiophile recordings is the Sonic Solutions system, though other
systems such as SADiE are encroaching on the market and promise to
offer improved flexibility. In the analog world, tape editing is
performed the old way, using a razor blade, grease pencil, and
splicing block. Few could argue that this method is preferable to the
astounding possibilities proffered in the digital editing world.
Often an analog recording destined for CD will be transferred
digitally at this point and edited on a DAW.
Upon completion of editing, a new digital runout is made, encompassing
the edited form of the music. From there, it's on to mastering.
Digital mastering (more properly called pre-mastering, since the
actual master is made later on...see below) is the time when you
finally need to construe the album as a whole. How loud do you want
the individual tracks to be? Do you _really_ want the album to go
out to 0 dB digitally? What about a clavichord recording? Would it
be wise to place it on the CD such that 0 dB is achieved? What happens
to the person who just played a Beethoven symphony on their stereo?
They'll be treated to a mighty loud clavichord.
Level differences, timing issues, and importantly, track start and end
IDs are all dealt with during mastering. Further, if the recording
was made at >16-bit resolution, mastering is when the recording is
redithered to 16 bits. The brilliant researcher Dr. Stanley Lipshitz
teaches us that by weighting the dither such that it is concentrated
in areas where the ear is least sensitive to noise, we will be able to
achieve linearity and resolution far, far below the least significant
bit due to the ear's ability to pick out non-random weightings among
randomness. Further work in this field has been done by Michael Gerzon,
Bob Stuart, Rhonda Wilson, and others.
An aside: This idea...noise shaping...is perhaps the greatest use of
96 kHz sampling. A 96 kHz sampling rate will give us a 48 kHz
bandwidth. Whereas now we place the noise into regions where we are
less sensitive, once we have a 48 kHz bandwidth we'll be able to shape
the noise into regions where we are profoundly deaf (i.e. ultrasonics).
96 kHz may, ultimately, do more for dynamic range detail than it will for
frequency response. But I digress..
Once the IDs are in place (i.e. when does the counter count up, when
does it count down, how many frames of silence go before the music
when you press the search key), the PQ burst must be generated. The
PQ burst is an electronic take-sheet that is used to cut the CD, and
actually contains all the display information for the CD player. When
you see the counter moving on your CD player, that isn't an internal
clock. That's actual data being read off the disc and displayed. It
is during mastering that all that data is generated. When you put in
a CD and the player tells you the total time and the track count,
that's all in the PQ burst.
Once the PQ is done, the data and PQ have to be stored on a master
delivery format so that they can be sent to the plant. The original
CD mastering format was 3/4" videotape using the PCM-1600-series
encoding scheme...a clunky format which is at last suffering a long
overdue death at the hands of more robust formats. The current
favorite plant delivery formats are CD-R and Exabyte DDP. Both have
provisions to hold both audio and PQ mastering data.
Upon receipt at the plant, the delivery format is scanned for any
noncorrectable errors. In the case of a CD-R, it is put on a readback
drive which extracts the PQ data as a serial stream and the audio as
yet another stream. The data is modulated through an encoding process
called Eight to Fourteen Modulation (EFM). The output of the EFM is
used to drive a laser cutter which strikes a glass master which is
coated with a photoresist. The glass master is then developed and
the image is used to make the metal stampers that will ultimately
be used to press the disc.
The stampers are used to impress the pit pattern into a layered
polycarbonate substrate which, after pressing, is metalized, and
layered over with another thin layer of polycarbonate. The label is
silkscreened onto the disc, and the discs are put on a spindle and
sent to in-packing where they are stuffed into jewel boxes, the
booklets inserted, and the product shrinkwrapped.
As Dick Pierce would say, so there you have it.
Vi behöver lösa din tråd då och då, när diskussionen kommer upp! Tack för din insats att utbilda oss!