Longread translated automatically
big

Лейбл «ШОВ»: источники медиа

Media Concept

Title:

SHOV—a music label and media outlet focusing on experimental music, sound art, and sonic explorations.

Short Description:

«SHOV» releases music, discusses how it is created, and connects listeners, musicians, and artists through releases, lyrics, audio materials, and events. The project works with music that doesn’t always fit into conventional genres or formats: ambient, noise, drone, field recordings, modular synthesis, and audio-visual practices. The label doesn’t just publish releases; it builds context around them: author notes, interviews, analyses of sonic processes, playlists, small exhibitions, and listening sessions.

Mission:

To create a space where experimental music can be heard, understood, and presented not as an insular practice «for insiders, ” but as an accessible experience of attentive listening.

Media Goal:

To support independent musicians and sound artists by helping them find an audience and cultivate a culture of conscious listening through digital publications, physical releases, and offline events.

Data sources for media

Internal Sources (Human-Driven):

1. Open Calls and Demo Submissions: Audio files (WAV/FLAC), conceptual descriptions, and setup diagrams submitted by musicians via website forms or Telegram.

2. Interview Audio Recordings: Unedited dictaphone recordings of conversations with sound artists and acoustic researchers.

3. «SOHA» Author Archive: A private library of field recordings complete with geolocation metadata and equipment specifications.

External Sources (Machine-Driven / API & Scraping):

1. Bandcamp API & Web Scraping: Monitoring new releases tagged field-recordings, experimental, drone, noise, sound-art, and soundscape.

2. Freesound.org API: Automated search for open samples and noise textures based on key parameters.

3. RSS Feeds and Parsing Specialized Media: Automated collection of news and announcements from The Quietus, KVR Audio, Chain D.L.K., and Neural.it.

4. Spotify Web API & Last.fm Trends: Analysis of micro-trends in niche avant-garde genres (popularity indices and similar artists).

Data Diagram: Human vs. Machine Processing

  1. Data Collection (Automated)Parsers collect new releases from Bandcamp and Freesound.A bot accepts demo recordings and audio interviews from artists.Data is saved to a shared database (Notion/Google Sheets).↓2. Preliminary Processing (AI Services)Whisper: Transcribes audio interviews to text.Librosa: Analyzes the tempo (BPM) and spectrum of the tracks.LLM: Summarizes the interviews and extracts key insights.↓3. Curatorial Selection (Human)Editor: Reviews the interview transcripts, refining style and meaning.Curator: Listens to the selected tracks and decides what will be released.Art Director: Approves the visuals and cover art.↓4. Publication and Distribution (Automated)FFmpeg: Automatically generates video previews with a sound wave.Poster Bot: Formats the post for Telegram and includes a Bandcamp link.Script: Uploads the final release to streaming platforms.

Guidelines for Automated Services in the Scheme

1. AI Transcriber and Summarizer (Whisper + LLM)

System prompt for the LLM editor:
Role:

You are a professional media music editor for «SHOW,» specializing in experimental music and sound art.
Task: Reformat a raw, unedited voice interview transcript with an artist into a structured draft article.
Rules:
Do not alter the core meaning or specialized terminology (e.g., «generative synthesis,» «spectral density,» «field recordings,» «patching»).
Remove filler words («um,» «like,» «so») while preserving pauses and the author’s intonation.
Divide the text into thematic blocks with concise headings.
Avoid marketing jargon and clichés (e.g., «unique,» «stunning,» «masterpiece»). The tone must be reserved, inquisitive, and respectful.

2. Acoustic Analysis Audio Script (Python / Librosa)

Purpose: Automated initial analysis of submitted demo tracks for database tagging.

Processing Algorithm:
Peak & RMS Loudness: Determining the dynamic range. If RMS is above -8 dBFS (over-compressed sound), the file is tagged hard_compressed.
Spectral Flatness: Determining the character of the sound (noise vs. tonal). High values = noise / textural; low values = ambient / drone.
Zero-Crossing Rate (ZCR): Identifying high-frequency noise and clicks for automatic marking as lo-fi or field_recording.

3. Auto-Media Asset Generator (FFmpeg + Telegram Bot)

Purpose: Automatic creation of preview videos for Telegram and Instagram from an audio file and a cover image.

Rules for card and video layout: Video format: 1:1 (Square, 1080×1080px), length — 45 seconds (the most expressive snippet of the track).

Wave visualization: A minimalistic oscilloscope line at the bottom of the frame (color: orange FF5500 or monochrome FFFFFF on a black background 000000).

Post text formatting: Title: [ARTIST] — [RELEASE / TRACK NAME] (bold font).

Metadata block: tags, recording format, equipment used.

Description: No more than 400 characters (an author quote or a brief summary from the AI summarizer).

Mandatory link to the full release on Bandcamp.

Лейбл «ШОВ»: источники медиа
Project created at 08.11.2026
We use cookies to improve the operation of the website and to enhance its usability. More detailed information on the use of cookies can be fo...
Show more