Triptych on the Weight of Words. Part 3: Sushi
Parameters
Description
Falling characters create a stream of text generated in real time from the user’s speech.
The speech recognition system converts words into pulsating code, reminiscent of the symbolic notation of the world known from The Matrix.
Sushi combines the pop-cultural motif of digital code with a reflection on the place of human speech in a reality shaped by artificial intelligence.
Read more
Sushi
Perhaps the most ingrained pop culture “falling” text is the distinctive pattern created for the Wachowski sisters’ “Matrix” film series. Therefore, the Miniatures included a recreation of this pattern. As with the other installments of “Triptych,” this time, the Miniature utilizes the browser’s built-in speech recognition engine (it’s worth noting that all three thumbnails within “Triptych” allow you to select a recognized language from among the many languages and dialects built into most popular browsers). Apparently, the original text displayed in the film came from the menu of a Japanese restaurant where the filmmakers met. However, “Sushi” focuses on the language capabilities of AI—after all, “The Matrix” is a story about a conflict between humans and thinking, conscious machines.
“Triptych on the Weight of Words” combines three Miniatures into a larger whole that examines and plays with the titular issue, and is based on two sources of inspiration.
The first is the (meta?) metaphor of speaking about language, and particularly about words, which, when described, are given physical properties, especially weight and mass. Such metaphors are produced by both literature and everyday language (perhaps the tendency to carve words considered important into heavy, massive objects—tombstones, columns, monument pedestals, etc.—and other similar activities are also an emanation of the same phenomenon). Giving words physical attributes is a cross-cultural practice and gives the impression of a natural human predisposition that does not require special training, coming with the development of linguistic competence. The power of such a practice seems to correspond to the importance of the brain system of intuitive physics with which evolution has endowed humans (and with concepts of evolutionary aesthetics in general)—in this case, it would be a beautiful example of the culture-forming potential of the side effects of natural selection mechanisms. Falling and sinking words, subjected to gravity and the resistance of dense media in which this falling takes place, or simply encountering hard obstacles, anchor the reading process in the parameters of the physical world.
The second source of inspiration concerns the physical nature of information (and thus the “content” of language and words) and the mass/energy/information (M/E/I) equivalence hypothesis. According to the M/E/I hypothesis, information is a physical quantity that can be associated with the energy necessary to change or erase a single bit of information. In very simple terms, the hypothesis can be explained by invoking the so-called Landauer’s rule (a physical principle stating that erasing a single bit of information requires at least a certain minimum energy, which is proportional to the temperature of the system—erasing information therefore entails an increase in entropy, e.g., at room temperature, erasing a bit “costs” 3 x 10 to the power of -21 joules). Landauer’s rule determines the theoretical efficiency of every conceivable computer, but above all, it relates information to the energy values required to change it. By employing a few more formulas from thermodynamics, Shannon’s information theory, and remembering the matter-energy equivalence rule (E = mc2), we can calculate the mass of a bit of information at room temperature (300°K, or 26.85°C), which is 3.19 x 10 to the power of -38 kg. This is not a large mass, and experimental confirmation (or refutation) of the mass/energy/information equivalence hypothesis is currently not possible. The M/E/I hypothesis is widely discussed among researchers exploring the intersection of physics and biology, for example, those attempting to define the characteristics of life—there are concepts that emphasize the exceptional energy efficiency of some life processes, approaching the barrier defined by Landauer. James Clerk Maxwell’s famous thought experiment (“Maxwell’s demon”), in its modern version, also expands on the demon’s task of acquiring and processing information about the world and the associated energy expenditure.
Technical documentation
Parameters passed via URL
- pointer 0 or 1; default 1; hides/shows the mouse cursor
- worker 0 or 1; default 0; allows blocking the browser’s built-in mechanisms that suspend the program when the window is not visible
- stt 0 or ISO language mnemonic; disables (0) the speech detection system or sets the language (from those available for the used browser) in which this system operates (ISO mnemonic); more information: http://www.lingoes.net/en/translator/langcode.htm
- items default 170; sets the number of “streams” with text phrases
- repprob 0.0 … 1.0; default 0.7; probability of reusing a text phrase (in such a case, the phrase is passed again to the system for display)
- dens default 2; “density” of passing new phrases
- flip 0 or 1; default 0; switches the text display direction (top to bottom or bottom to top)
- strings default 7; maximum number of text phrases in the buffer
- minfontsize 0.0 … 1.0; default 0.01; minimum normalized size of the displayed text
- maxfontsize 0.0 … 1.0; default 0.04; maximum normalized size of the displayed text
- minspeed 0.0 … 1.0; default 0.1; minimum normalized falling speed of a phrase
- maxspeed 0.0 … 1.0; default 0.3; maximum normalized falling speed of a phrase
- maxrndstrl default 15; maximum length of randomly generated, stylized text streams
- continuous 0 or 1; default 1; disables/enables processing of continuous phrases captured by the speech detection system
- interim 0 or 1; default 1; disables/enables processing of interim phrases captured by the speech detection system
- blkdur (ms) sets the default “flash” time of the text after changing the stream’s content to new content
- blk 0 or 1; default 1; (“blink”) disables/enables the “flash” at the moment of replacing the text stream’s content with new content
- cri 0 or 1; default 1; (“change random item”) disables/enables the mode where new content is assigned to random text streams
- blktxtcol text color at the moment of “flashing” in standard hexadecimal notation 0xRRGGBB
- txtcol text color in standard hexadecimal notation 0xRRGGBB
- font name of the font used to display text
- resetitempos 0 or 1; default 0; disables/enables resetting the horizontal and vertical position of the text “stream” when changing its content to new content
- emit 0 or 1; default 0; disables/enables broadcasting the STT engine results via web socket
- ws address; disabled by default; enables and sets the address (usually localhost:9001) for WebSocket communication (allows external control of Miniature and customization)
- xws address; disabled by default; enables and sets the address (usually ws://localhost:9001) with protocol specification for WebSocket communication (allows external control of Miniature and customization)
Messages passed via WebSocket
Messages to which Miniature responds
- str text phrase (string) from an external source
Messages sent by Miniature
- sushi stt text phrase (string) from Miniature’s internal STT system
MIDI
Messages to which Miniature responds
not applicable
Messages sent by Miniature
not applicable
Communication with MaxMSP
Messages to which Miniature responds
not applicable
Messages sent by Miniature
not applicable