SUB.PRESSURE — Sound as Investigation
SUB.PRESSURE — interactive media on dubstep, bass music, and sound perception.
The project is founded on the idea that a music track can be studied not only through its technical characteristics but also through listener reactions.
Instead of conventional reviews and playlists, the media offers small experiments. Users listen to a musical excerpt, mark interesting moments, and compare their reactions with those of other participants.
The central question of the project is: why can the same sound evoke completely different feelings in different people?
The media’s unique feature is that listeners generate the data themselves, which forms the basis for future publications.
Media Format and Structure
The SUB.PRESSURE content is built around musical experiments. Each experiment involves four stages:
LISTEN — Listening: The user receives a short musical excerpt without prior explanation.
REACT — Reaction: During listening, they mark moments where they expected a drop, felt maximum sound power, or lost interest.
MAP — Reaction Map: Participants' responses are compiled into a shared timeline, allowing viewers to see the most frequently marked segments.
DECODE — Analysis: The editorial team analyzes the results and explains which compositional features might have influenced the perception.
From a single experiment, several publications can be created: an interactive quiz, a visual reaction map, a short video, and a full musical analysis.
Data Sources
For SUB.PRESSURE, data is sourced from four main areas.
Musical Materials from Artists Producers submit demo versions of tracks, individual sound snippets, and alternative composition drafts.
These materials form the basis of the experiments.
User Feedback This is the primary source of information. Participants provide notes on anticipating a drop, the most impactful moments, waning interest, and a desire to re-listen to specific parts of a track.
Playback Analytics When using the interactive player, the system can log user actions: replaying, switching between versions, pausing, and jumping to a different segment.
Professional Sources This includes musician interviews, sound design resources, studies on musical perception, and educational publications about electronic music production.
These sources are used to validate and enhance our editorial critiques.
Data Flow Diagram
In SUB.PRESSURE, information processing is shared between the editor and the automated system.
Stage 1. Preparation
Human: Accepts the demo from the musician. Verifies permission to use the material. Selects the compositional element to be studied. Formulates questions for the listeners.
Machine: Loads the audio file. Determines the track duration. Creates a timeline and a visualization of the sound wave. Prepares the experimental interface.
Stage 2. Experiment Execution The human participates in listening and notes their own sensations.
The machine logs the selected moments, ties them to the track time, and saves the responses.
Stage 3. Analysis The machine aggregates the reactions, calculates the proportion of responses, and generates an activity map.
The editor reviews the statistics, compares the results with the musical structure, and prepares an explanation.
Stage 4. Publication After review, the material is formatted into an interactive article, card, or short video.
SUB.SCAN — Automating Musical Experiments
SUB.SCAN — a designed service for testing musical demos within SUB.PRESSURE. Its purpose is to transform listener actions into actionable data for editors and musicians.
What data the service processes: DROP POINT — the moment the user expected a drop. PEAK — the segment that seemed the most impactful. INTEREST LOSS — the moment of declining interest. REPLAY — the fragment the user wanted to listen to again. A/B TEST — preference between two versions of a track. For each experiment, a card is created that includes the demo title, its duration, the version number, the number of responses, and the results.
SUB.SCAN Results An interactive composition map where the musical wave is combined with audience reactions. For example, a musician can compare the original and reworked versions of a single drop and see how the participants' responses have changed.
Guidelines for SUB.SCAN and Automation
For automated tools to function correctly, standardized rules must be established.01. Do not alter results. The system must not spontaneously invent reactions, inflate metrics, or suppress unsuccessful experimental outcomes.02. Ensure accurate comparison. In A/B tests, versions must be compared under identical conditions. Differences in volume, duration, and playback method must be accounted for.03. Maintain precision. Each user rating is tied to a specific moment in the audio file. Data from different track versions are not mixed.04. Protect information. The service collects only the data necessary for the study. Participants are informed in advance about the collection of reactions, and publications use anonymized results.05. Separate facts from interpretation. Automation shows how many participants rated a specific moment, but it does not independently determine whether a track is good or poor.06. Review publications. All graphs, captions, and conclusions must be editorially reviewed before publication.
How Data Transforms into Content
The primary goal of automation is not only to speed up information processing but also to create new publication formats.
DROP MAP Visualization of the moments when listeners expected the drop to start.
PRESSURE TEST A comparison of two bass variations, displaying participant preferences.
REPLAY ZONE A map of the composition sections that users most frequently replayed.
BEFORE / AFTER A comparison between the original demo and the version modified after receiving feedback.
Each piece of material can be published across multiple platforms. Telegram is used for short tests, the website for full-scale interactive experiments, VK for visual results, and Shorts for quick comparisons.
Thus, a single musical experiment becomes the foundation for several content formats.
Final SUB.PRESSURE Model
SUB.PRESSURE is built on the interaction of three participants. The musician provides the material and receives feedback. The listener participates in the experiment, generating data through their own reactions. The editorial team and SUB.SCAN process the information, create visualizations, and explain the results. Automation handles data collection, storage, and calculation, while the editorial team is responsible for verification and interpretation. As a result, a continuous cycle is formed: Musical excerpt $\rightarrow$ Audience reaction $\rightarrow$ Data map $\rightarrow$ Analysis $\rightarrow$ New experiment. The system allows listening to music to transform into an interactive process, enabling the resulting reactions to be used both for content creation and for improving musical compositions.
