Preparation - Technical Explanation
Voice messages will be cleaned of surrounding
“noise” using various filters, and translated into
a format compliant with a spectrograph analyzer
(SpA).
The SpA will process the sound file and display
visual interpretations according to various parameters such as
sample rate, frequency range, bands per octave, bandwidth
filter, rows per second, etc. Slight adjustments to these
parameters allow us to highlight certain aspects, increase
output clarity, and choose the most suitable image from an
artistic perspective.
At this stage, we have different types of black
and white images, as illustrated in the examples below.
sample 1: 3D graphic
(sound level-frequency-time)
|
sample 2: 2D graphic
|
sample 3: 3D graphic
|
The three samples above and the one below show
four different spectral representations of the recording of
Marilyn Monroe’s famous line “Come on now,
everything's fine” (from “The Seven Year Itch”,
directed by Billy Wilder in 1955).
|
|
Each gray dot in these
diagrams corresponds to a precise value. Thus, sample 1
and 4 are 3-dimensional sound wave graphics where the
x-axis (left to right) corresponds to the time, so that
the first word Marilyn pronounces is on the left, and the
last one is on the right. The y-axis (bottom to top)
corresponds to the frequency of her voice at a given
moment, while the different shades of gray translate the
sound level.
|
The tiny gray dot at the
point of the red arrow in the picture above translates the exact sound level of the
3.5kHz frequency in the “v” of the word
“everything” in Marilyn Monroe's
recording. These raw images straight out of the
spectrograph might seem rather dull, but they contain all the
information we need to interpret the voice message as curves,
patches and shades of gray.
Graphical treatment
We apply various filters and colors to
emphasize the pattern and dynamics of the voice message. With our
top-notch graphic programs the possibilities
are almost infinite. By defining precise filter sets, we obtain an
optimized rendering and still keep full control of the
content. In theory we can reverse the process at this stage and get back to
the original sound file that was recorded.
From encoding to art
The image on the right is a simple graphic translation
of Marilyn's recording. The process we used here is very
similar to that used to encode a message. If you
translate the characters and spaces in the message
“Come on now, everything's fine” using a
precise code table, such as: space = 0, a= 1, b=2... k= $
etc. you will get something like this:
Even if this colored gibberish doesn't look very artful, it illustrates the
principle of encoding. Thus, by using the same code table, you can
easily trace back to Marilyn's initial line “Come on
now, everything's fine.”
|
 |
|
Using a computer has this advantage: up to a certain point you can reverse the action, or even record what you have done so far. Such a record can be seen as a “code table“, which would mean that we “encode”
voice messages in a similar way to Marilyn's line. But only the painter decides, he will make many subjective choices all along: how to transform, deform or merge, which colors to use, what to ad or eliminate, what parts of the whole should be emphasized, etc. And don't forget that this is still raw matter. Now the painter will use this raw material to produce his art. Sometimes he will continue working on the computer in free inspiration, until the painting is finished, sometimes he will start directly on canvas.
Thus starting from physical ingredients that are invisible to our eyes - sound waves - we "translate" them into visible building blocks of human art, that will be used to reveal the hidden beauty of the human voice.
|
|
