Envion Ontology / Sampling Without Awareness

In the Envion ecosystem, sound is not sampled — it is harvested, stripped, and reborn. The act of recording or fetching is not documentation but ontological mutation. Each sonic fragment becomes detached from its author, context, and origin, re-entering the system as pure gesture or potential form.

This process embodies the principle of sampling without awareness: the deliberate erasure of provenance, an ethical stance that refuses ownership and recognition. The fetched or processed sound no longer represents anything — it exists, suspended between signal and meaning, as an anonymous event.

Ontologically, Envion operates as a cybernetic ecology where material and code co-produce each other. In this system, gesture ≠ material — gesture is the relational behaviour that brings material into being. Through its Seeder and Foundry components, Envion orchestrates this metabolism between network, machine, and ear.

EnvionSeeder / Network-Based Sound Processing Ecosystem

EnvionSeeder represents the networked counterpart of the Envion ecosystem — a living infrastructure that unites fetching, transformation, and ontological depersonalization of sound within one integrated flow. It connects Pure Data’s gesture-based environment with an online backend for procedural harvesting, normalization, and erasure of source identity.

• Fetches random or query-based sounds from curated APIs (currently Freesound)
• Performs automatic slicing and normalization
• Provides a live web console with query logs
• Builds an ontological archive of depersonalized sound fragments

Ontologically, EnvionSeeder extends the concept of sampling without awareness — the system fetches audio atoms scattered across the network and rewrites them through automated processes that dissolve authorship, semantics, and identity.

• Short random slices
• Non-readable filenames
• Extended fade-outs and short fade-ins
• Pitch-up: asetrate=384000, aresample=48000, atempo=1.6
• Pitch-down: asetrate=16000, aresample=48000, atempo=0.7

EnvionFoundry — the core local environment that performs advanced audio processing, ffmpeg-based transformations, and manifest generation — represents the future research-tier of Envion.

Access & Contribution

EnvionFoundry is currently distributed through a private Git repository. To gain access, you must request an invitation directly from the author.

Because EnvionFoundry is part of a self-funded artistic-research project, access to the private repository requires a small one-time contribution that helps cover server, maintenance, and development costs.

• Request access via email or message (GitHub username required)
• Receive an invitation link to the private repository
• Clone or download EnvionFoundry once approved

Your contribution directly supports the continued growth of the Envion ecosystem, ensuring that its tools remain independent, sustainable, and publicly documented.

1) What is EnvionFoundry

EnvionFoundry is a small local lab that takes any audio file and prepares it for Envion — automatically and effortlessly.

• Cuts the audio into 10 slices of 3 seconds
• Applies pitch-up and fade
• Makes fragments louder and more stable (boost + limiter)
• Saves everything in a folder ready for Envion

No technical skills required — open a page, choose a file, click Process Audio.

Upcoming & Extended Processes

Beyond the basic slicing pipeline, EnvionFoundry already integrates and tests a growing set of FFmpeg and CDP (Composers Desktop Project) transformations — processes that expand Envion’s sonic vocabulary and bring it closer to the language of musique concrète and algorithmic morphology:

• FFmpeg spectral modes — bitcrush erosion, reversed lowpass masking, spectral drift and time-domain blurring
• CDP spectral transformations — morph, blur, stretch, grain, and re-synth operations mapped to Envion’s procedural architecture
• Dynamic convolution with cassette and domestic IRs (Imprintape engine integration)
• Entropy modulation — randomised amplitude, duration and envelope mapping for each slice
• Manifest generation — automatic metadata and structure logs for research or sample-pack export

These processes are continuously evolving — early supporters who contribute to access the private repository receive all future FFmpeg and CDP expansions as they are released.

3) Folder structure

EnvionFoundry/
├── app.py
├── index.html
├── style.css
├── script.js
├── EnvionFoundry_Guide.html   ← this guide
├── uploads/                    ← temporary inputs
└── output/                     ← processed results

Recommended: place EnvionFoundry inside Envion-Algo-Score/tools/ and exclude it from Git with .gitignore.

5) Using EnvionFoundry

1. Click Select audio file and choose a .wav or .mp3
2. Click Process Audio
3. Watch the green log and, when finished, open:

   output/<filename>/slices_pitchup/

Each fragment is already sliced, pitched, faded, boosted and limited — ready for Envion.

6) What happens inside

• Slicing: 10 × 3-second WAVs (48 kHz / 16-bit)
• Pitch-Up: resampling + smooth fade-in/out
• Boost & Limiter: loud, clean signal with no clipping

The pipeline matches the “v5” script you’ve tested earlier.

7) Troubleshooting

“No module named flask”

pip3 install flask

“ffmpeg: command not found”

brew install ffmpeg

Page opens but nothing happens

• Check the Terminal window — FFmpeg errors are shown there.
• Try again with a different audio file.

Output sounds too quiet

The pipeline includes boost + limiter. To increase gain further, edit volume= in app.py (section “boost + limiter”).

Additional Troubleshooting — Seeder (Online Fetch Console)

Invalid JSON response / Empty result

• Check Internet connection.
• Verify Freesound API token validity.
• Ensure stats.php and fetch/ have write permission.

PHP error on peamarte.it/envion/

• Hosting must support PHP 8+ with curl enabled.
• Check fetch_stats.json exists and is writable.

File fetch works but no audio downloaded

• Verify sound ≤ 50 s duration (longer files are ignored).
• Retry with a different keyword.

Cache issue (same sound repeatedly)

• Clear browser cache or refresh with CMD + SHIFT + R.
• Reset fetch_stats.json manually.

If problems persist, test the local EnvionFoundry pipeline to confirm FFmpeg and slicing work independently.