DAJAI · Glossary
Glossary
The language of the catalog — mastering, ownership, and the DARK Library universe, defined.
35 terms
- 31-Band Reference Match aka 31-band EQ match, spectral match, deterministic reference matchMastering
- A 31-band reference match analyzes a track and a reference across 31 frequency bands (mirroring a graphic-EQ layout) and computes per-band corrections to align the track's spectrum to the reference. The process is deterministic and rule-based rather than guesswork, so the same input and reference always yield the same correction. It is the spectral-matching core of DAJAI's mastering engine, paired with reference profiles to dial in tonal balance band by band. The result is a transparent, repeatable way to make an independent master sit alongside commercial releases.
- Catalog Compounding aka catalog growth, back-catalog revenue, catalog stackingOwnership
- Catalog compounding is the effect by which each new owned release adds to a growing back catalog that keeps earning streaming and licensing revenue indefinitely. Because the artist owns the masters, older tracks continue paying out while new ones stack on top, so total earning potential rises with every release. A planned multi-volume body of work like the DARK Library is designed to compound this way over years. It is the core financial logic of owning your masters as an independent artist.
- Cloudflare Named Tunnel aka cloudflared tunnel, cloudflare tunnel, named tunnelSovereign Tooling
- A Cloudflare named tunnel is a persistent outbound connection from a private machine to Cloudflare's edge that exposes local services to the public internet without opening inbound ports or renting a server. The cloudflared daemon dials out and Cloudflare routes specific hostnames to the local ports behind it. DAJAI uses one named tunnel to publish multiple domains from a single home Mac, with no VPS in the path. It is what makes hosting a real website on hardware you own both safe and practical.
- Concept Album aka narrative album, thematic albumDARK Universe
- A concept album is a record whose tracks are unified by a single overarching theme, narrative, or framework rather than standing as unrelated songs. DARK I is a concept album, with each of its ten tracks mapped to a chapter of Outwitting the Devil and positioned within an esoteric structure. The format rewards listening front to back and supports deep worldbuilding across a multi-volume cycle. It is the storytelling form the entire DARK Library is built on.
- DARK I — Outwitting the Devil aka dark i, outwitting the devil album, DARK Volume IDARK Universe
- DARK I, subtitled Outwitting the Devil, is the first volume of the DARK Library, released April 7, 2026, on Apple Music and Spotify through distributor UnitedMasters. It comprises ten tracks, each chapter mapped to Napoleon Hill's 1938 manuscript Outwitting the Devil, and was mastered by Solana Conejo. It sits at Kether, the crown, on the Kabbalistic Tree of Life within the cycle's symbolic map. As Volume I, it establishes the conceptual and sonic foundation for the entire ten-volume DARK cycle.
- DARK II — Too Dark aka dark ii, too dark, Too Dark 2024DARK Universe
- DARK II is the second volume of the DARK Library, released as Too Dark in 2024 and distributed via CDBaby (a separate distributor from DARK I's UnitedMasters). It appears in DAJAI's discography on Apple Music and Spotify alongside the rest of the catalog. Though released before DARK I in calendar time, it occupies the second position within the cycle's conceptual ordering. It extends the DARK universe's themes and contributes to the compounding owned catalog.
- DARK Library aka dark library cycle, the dark cycleDARK Universe
- The DARK Library is DAJAI's planned ten-volume album cycle, a unified body of work releasing across years under a single conceptual architecture. Each volume is a chapter in a larger thematic and esoteric framework, beginning with DARK I and continuing through DARK II and beyond. It is structured as an owned catalog meant to compound in value as the cycle completes. The Library doubles as a worldbuilding universe linking the records, their artwork, and their lore.
- Discography aka catalog, body of work, releasesCatalog
- A discography is the complete catalog of an artist's released recordings, including albums, EPs, and singles across all distributors and platforms. DAJAI's discography spans the DARK Library volumes plus earlier works such as Flight to Vegas 2, Dark, Daj's Dedication, and Neoworld, alongside thousands of catalog tracks on SoundCloud. For a master-owning independent, the discography is both an artistic record and the compounding revenue base. Keeping it accurately registered with ISRCs, The MLC, and SoundExchange ensures every release is collectible.
- Distribution vs Label aka distributor vs record label, DIY distribution, independent distributionOwnership
- Distribution delivers your finished masters to DSPs like Apple Music and Spotify in exchange for a flat fee or percentage, while a record label typically funds, controls, and owns recordings in return for a larger share. A distributor is a pipe; a label is a partner and usually an owner. DAJAI releases through direct distribution (UnitedMasters for DARK I, CDBaby for DARK II) with no label intermediary, keeping master rights and creative control. Choosing distribution over a label is the structural decision behind the sovereign-artist model.
- Headroom aka dynamic headroom, gain stagingMastering
- Headroom is the available margin between the loudest part of a signal and the maximum level a system can handle without clipping, usually expressed in decibels. Leaving adequate headroom during mixing and mastering prevents distortion, gives limiters room to work transparently, and keeps inter-sample peaks under control. A mix delivered with healthy headroom (and a true-peak-safe ceiling at the master) survives loudness normalization and lossy encoding intact. It is a foundational concept in gain staging across the whole production chain.
- Host-Header Multi-Tenancy aka multi-tenant by host, host-based routing, host header routingSovereign Tooling
- Host-header multi-tenancy is an architecture where one application serves several distinct sites by inspecting the incoming request's Host header and rendering tenant-specific content and branding. A single Next.js codebase can present entirely different sites for different domains without duplicating the app. DAJAI's three domains share one codebase and diverge purely by host header, simplifying maintenance and deployment. It is the software side of the one-machine-many-sites approach.
- ISRC (International Standard Recording Code) aka international standard recording code, recording codeCatalog
- An ISRC is a unique 12-character global identifier assigned to a specific sound recording, used to track plays, sales, and royalties across every platform. It stays with the recording for life, so a track keeps its ISRC across reissues, playlists, and DSPs. Distributors assign ISRCs at upload, and they are the key that ties streaming data and royalty statements back to one exact master. Accurate ISRCs are essential for collecting from DSPs, SoundExchange, and neighboring-rights bodies.
- Kether aka keter, the crown, sephirahDARK Universe
- Kether (the Crown) is the topmost sephirah on the Kabbalistic Tree of Life, representing the point of origin and unity from which everything emanates. In the DARK Library's symbolic architecture, DARK I is positioned at Kether, marking it as the crown and source of the cycle. Mapping each volume to a position on the Tree of Life gives the ten-album structure an esoteric scaffold that parallels its literary framework. It signals the project's deliberate fusion of music, mysticism, and concept.
- launchd aka launch agent, launchd job, macOS service managerSovereign Tooling
- launchd is macOS's native service and job manager, used to start, supervise, and restart long-running processes and scheduled tasks. Defined by plist files, launchd jobs keep web servers, tunnels, and automation running across reboots without manual intervention. In the sovereign stack, each website process and the Cloudflare tunnel run as launchd jobs so the whole operation is self-healing on one machine. It is the process-supervision layer that makes a home Mac behave like reliable server infrastructure.
- LM Studio aka lmstudio, local model serverLocal Inference
- LM Studio is a desktop application for downloading and running large language models locally, exposing an OpenAI-compatible API on a local port. It lets developers point existing tooling at a local model instead of a paid cloud endpoint, with no code changes. In the sovereign stack it serves general LLM tasks on owned hardware, complementing MLX-served vision models. It is a practical bridge between off-the-shelf models and a self-hosted, cost-independent workflow.
- Local Embeddings aka on-device embeddings, sentence embeddings, vector embeddingsLocal Inference
- Local embeddings are vector representations of text or media generated by a model running on your own hardware, enabling semantic search and similarity matching without a cloud API. A small sentence-transformer model can produce hundreds of dimensions per item at high speed on a single machine. In the sovereign stack they power search and recommendation over a catalog at zero marginal cost. They are the retrieval layer of a fully self-hosted, local-first AI system.
- Loudness Normalization aka streaming normalization, playback normalization, loudness levelingMastering
- Loudness normalization is the process by which streaming platforms adjust playback gain so all tracks sound roughly equal in loudness, typically targeting an integrated LUFS value. Because the platform turns loud masters down rather than up, mastering past the target gains no loudness advantage and only sacrifices dynamics. Understanding normalization is why modern masters aim for a sensible LUFS and a safe true-peak ceiling. It reframed mastering away from the loudness war toward dynamic, codec-safe delivery.
- LUFS (Loudness Units Full Scale) aka loudness units full scale, integrated loudnessMastering
- LUFS is the standardized unit for measuring perceived loudness of audio, accounting for how the human ear weights different frequencies over time. Streaming platforms normalize playback to a target integrated LUFS (commonly around -14 LUFS for music streaming), so a master that is too loud is turned down rather than rewarded. Mastering to a sensible LUFS target preserves dynamics and avoids the distortion of the old loudness war. It is the primary loudness reference in a modern mastering chain.
- Master Rights aka master recording rights, owning your masters, sound recording copyrightOwnership
- Master rights are the ownership of the actual sound recording of a song, separate from the composition (publishing) copyright. Whoever holds the masters controls licensing, streaming payouts on the recording side, sync placements, and reissues. Traditionally a label owns an artist's masters in exchange for funding, but independent artists who release without a label retain 100% of their master rights. DAJAI owns the masters across the DARK catalog, which is the foundation of catalog compounding and long-term independent revenue.
- Mechanical Royalties aka mechanicals, reproduction royaltiesCatalog
- Mechanical royalties are payments owed to songwriters and publishers for the reproduction of a musical composition, including physical copies, downloads, and interactive streams. On streaming, a portion of each play generates a mechanical that is distinct from the recording-side payout. In the United States these are administered for streaming by The MLC, which collects from DSPs and pays the rights holders. For an artist who both writes and records, mechanicals are a separate income stream from the master royalties.
- Mid/Side Processing aka M/S processing, mid-side EQ, stereo widthMastering
- Mid/side processing decomposes a stereo signal into a mid (mono, center) channel and a side (stereo difference) channel so each can be EQed, compressed, or widened independently. It lets an engineer tighten a center-panned vocal and kick while widening reverb and instrument ambience in the sides, sharpening clarity without smearing the image. In mastering it is a precision tool for stereo width control and mono compatibility. Overuse can collapse the center or cause phase problems, so it is applied surgically.
- MLX aka Apple MLX, MLX frameworkLocal Inference
- MLX is Apple's open-source machine-learning framework optimized for Apple Silicon, using unified memory to run models efficiently on M-series chips. It lets a single Mac perform vision, language, and embedding inference locally without a discrete GPU or cloud service. In DAJAI's stack, MLX powers local model serving such as Qwen2.5-VL for vision tagging on an M4 Max. It is a key enabler of the run-on-hardware-you-own, self-hosted approach.
- One-Mac-Three-Domains aka multi-site single machine, one machine many sitesSovereign Tooling
- One-Mac-three-domains is the pattern of serving multiple public websites from a single Mac, each domain mapped to its own port and routed by host header through one Cloudflare tunnel. DAJAI runs three distinct sites, dajai.io, hellcatblondie.io, and sovereignagiasi.com, from one codebase on one machine. launchd supervises the processes and the named tunnel exposes them, with no VPS or cloud host. It is a concrete demonstration of running a full web presence on hardware you own.
- Outwitting the Devil Mapping aka napoleon hill mapping, chapter mapping, manuscript mappingDARK Universe
- The Outwitting the Devil mapping is the conceptual scheme by which each track on DARK I corresponds to a chapter of Napoleon Hill's 1938 manuscript Outwitting the Devil. Hill's text frames a dialogue about fear, drift, and self-determination, and the album translates those ideas into song-by-song themes. This one-to-one structure turns the record into a narrative concept album rather than a loose collection. The mapping is a defining feature of the DARK Library's literary and esoteric worldbuilding.
- Qwen2.5-VL aka Qwen VL, Qwen2.5 vision-languageLocal Inference
- Qwen2.5-VL is an open-weight vision-language model that can interpret images alongside text, used for tasks like captioning, tagging, and visual question answering. Running it locally via MLX or LM Studio allows automated image understanding without sending media to a third-party API. In DAJAI's operation it tags and classifies media assets on-device as part of the self-hosted pipeline. It exemplifies how open models make sophisticated AI ownable rather than rented.
- Reference Mastering aka reference matching, reference-based mastering, reference masterMastering
- Reference mastering is the practice of measuring a target track's tonal balance, loudness, and dynamics against one or more professionally finished reference songs and adjusting the master to match. Instead of mastering by feel alone, the engineer matches frequency curves, LUFS, and true-peak behavior to a proven benchmark. DAJAI's chain uses this approach through a reference_master.py pipeline that compares against curated reference profiles. It produces consistent, competitive, release-ready masters without a commercial studio.
- Reference Profile aka reference target, mastering profile, REF profileMastering
- A reference profile is a stored description of a target sonic signature, capturing the per-band tonal balance, loudness, and dynamic characteristics of a benchmark sound. The mastering engine loads a chosen profile and matches the track to it via the 31-band reference match. Maintaining a curated set of profiles lets an artist consistently hit a desired character across an album or a whole catalog. It turns subjective mastering taste into a reusable, deterministic asset.
- Self-Hosted Inference aka local inference, on-device AI, local-first AISovereign Tooling
- Self-hosted inference runs AI models on hardware you own rather than calling a paid cloud API. For an independent artist this means tagging, transcription, embeddings, and vision tasks happen locally at zero marginal cost and with full data privacy. DAJAI's stack runs models on a Mac with Apple Silicon using MLX and LM Studio, keeping creative and operational AI in-house. It is the AI counterpart to owning your masters: control and cost-independence over the whole pipeline.
- SoundExchange aka digital performance royalties, neighboring rights USCatalog
- SoundExchange is the US nonprofit that collects and distributes digital performance royalties for sound recordings played on non-interactive digital radio, including SiriusXM and internet radio. It pays the featured artist and the master rights owner directly, rather than the songwriter or publisher. These royalties are separate from interactive-streaming payouts and from mechanicals, and many independent artists leave them uncollected by not registering. Registering recordings and ISRCs with SoundExchange captures a distinct, often-overlooked revenue stream.
- Sovereign Artist aka independent artist, artist-owned operation, DIY artistSovereign Tooling
- A sovereign artist owns the full stack of their career: master rights, distribution choices, tooling, and infrastructure, without ceding control to a label or relying on rented platforms. The model pairs owned catalog economics with self-hosted production and operations so the artist captures the value they create. DAJAI runs the website, mastering pipeline, and release operations on owned hardware as an expression of this philosophy. It is the organizing principle behind both the music and the sovereign-AI infrastructure.
- Sovereign Mastering Chain aka reference_master.py, local mastering pipeline, Fable mastering engineSovereign Tooling
- The sovereign mastering chain is DAJAI's self-hosted, scriptable pipeline that takes a raw mix to a finished master entirely on owned hardware. It runs stem separation, a vocal RX chain (iZotope-class VST3 plugins), and a deterministic 31-band reference match through a reference_master.py engine. Because it is rule-based and local, it produces repeatable, release-grade masters with no studio fee and no cloud dependency. It is the production backbone of the independent, owns-everything model.
- Stem Separation aka source separation, stem splitting, demucs, BS-RoFormerProduction
- Stem separation uses machine-learning models to split a finished stereo mix back into component stems such as vocals, drums, bass, and other instruments. Tools like Demucs and BS-RoFormer make it possible to isolate a vocal for re-tuning, remix a beat, or repair a single element without the original session. In an independent workflow it enables vocal-clean passes, warmth rebuilds, and remastering of older tracks that exist only as two-track files. It is a cornerstone of AI-assisted, no-studio production.
- The MLC (Mechanical Licensing Collective) aka mechanical licensing collective, MLCCatalog
- The MLC is the US organization created under the Music Modernization Act to collect and distribute mechanical royalties from interactive streaming services to songwriters and publishers. DSPs pay The MLC a blanket royalty, which it then matches to compositions and pays out to registered rights holders. Self-administered independent songwriters must register their works and claim them to be paid, or the money sits unmatched. It complements SoundExchange (which handles the recording side of digital radio) on the composition side.
- True Peak aka true peak limiting, dBTP, inter-sample peakMastering
- True peak is the actual maximum level of an audio signal including the inter-sample peaks that appear when a digital file is reconstructed into analog or re-encoded to a lossy codec. It is measured in dBTP (decibels true peak) and can exceed the sample-level peak, causing clipping after streaming compression. Mastering engineers typically set a true-peak ceiling around -1.0 dBTP to leave safe margin for codec conversion. Controlling true peak is essential for clean playback across Apple Music, Spotify, and other DSPs.
- Vocal RX Chain aka vocal cleanup chain, vocal restoration, iZotope chainProduction
- A vocal RX chain is a sequence of restoration and polish plugins applied to an isolated vocal stem to remove noise, harshness, and artifacts before mixing. DAJAI's chain runs iZotope-class processors (RX, Nectar, Neoverb, Ozone) plus a pitch corrector, typically as VST3 instances. It cleans the vocal, controls sibilance and resonance, and adds presence so the vocal sits forward in the mix. Pairing it with stem separation lets even a phone-recorded vocal reach a release-grade result.