Files

T

Antigravity Agent f8b21af4be Wiki cleanup: error-doc removal, dedup merge, link normalization

10_Wiki/Topics 대규모 정리:
- 오류 캡처/미완성 stub 문서 227개 제거
- 교차폴더 중복 43클러스터 병합 (63파일 → redirect)
- 링크명 정규화: 깨진 링크 수정·redirect 직결·개념 매핑 ~2,400건
- 카테고리 MOC 6개 신규 생성
- Graph 섹션 미해결 related-keyword 링크 10,058건 제거

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-20 23:52:15 +09:00

4.9 KiB

Raw Blame History

id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack

title

The Evolution of Music Distribution

매 한 줄

"매 130-year arc — 매 vinyl → cassette → CD → MP3 → streaming → AI-curated → AI-generated". 2026 매 Suno/Udio AI track의 Spotify chart entry, 매 generative-music subscriptions, 매 artist + AI co-creation 매 default. 매 distribution 매 longer "moving atoms" 매 ranking inferences.

매 핵심

매 Era timeline

1900-1948 Wax cylinder, 78rpm shellac.
1948-1980s Vinyl LP / 45rpm, cassette (1963).
1982-2000s CD — digital but physical.
1999-2008 Napster → iTunes Store. Unbundling album → single.
2008-2020s Streaming (Spotify 2008, Apple Music 2015). Per-stream royalty economy.
2022-2024 TikTok-driven discovery. Snippets > full tracks.
2024-2026 AI-generated (Suno v4, Udio, Stable Audio 2). Personalized AI radio.

매 Economic shifts

Album → single → playlist track → 7-second hook.
Royalty: $0.003-0.005/stream (Spotify 2026).
Long tail: 매 100M+ tracks indexed; 매 50% never played.

매 Tech axes

Codec: AAC → Opus → neural codec (Encodec, SoundStream).
Discovery: editorial → collaborative filter → embedding-based → LLM agent.
Rights: ISRC → blockchain experiments → AI-attribution debate.

매 응용

Independent artist — DistroKid → all DSPs.
AI track creator — Suno + 매 manual master + DSP upload.
Personalized AI radio — Spotify DJ AI, Amazon Maestro.

💻 패턴

Music embedding for similarity (2026)

import torch
from transformers import AutoProcessor, ClapModel

processor = AutoProcessor.from_pretrained("laion/clap-htsat-unfused")
model = ClapModel.from_pretrained("laion/clap-htsat-unfused")

def embed_audio(wav_path):
    audio = load_audio(wav_path, sr=48000)
    inputs = processor(audios=audio, return_tensors="pt", sampling_rate=48000)
    with torch.no_grad():
        return model.get_audio_features(**inputs)

Recommendation collaborative filter

import numpy as np
from scipy.sparse.linalg import svds

# user x track plays matrix
U, S, Vt = svds(plays_matrix, k=128)
user_factors = U @ np.diag(S)
track_factors = Vt.T

def recommend(user_id, k=20):
    scores = user_factors[user_id] @ track_factors.T
    return np.argsort(-scores)[:k]

DSP metadata upload

{
  "isrc": "USXYZ2600001",
  "title": "Neon Dawn",
  "artist": "Aria Vox",
  "album": "Synth Bloom",
  "release_date": "2026-06-01",
  "explicit": false,
  "ai_disclosure": {
    "ai_used": true,
    "tools": ["suno-v4"],
    "human_role": ["lyrics", "mastering"]
  },
  "audio_url": "s3://...master.flac"
}

Suno-style generation prompt (2026)

import requests

resp = requests.post(
    "https://api.suno.ai/v4/generate",
    headers={"Authorization": f"Bearer {API_KEY}"},
    json={
        "prompt": "lo-fi hip-hop with rainy night atmosphere, 78 BPM",
        "lyrics_mode": "instrumental",
        "duration_sec": 180,
        "model": "suno-v4-pro",
    },
)

Royalty estimator

def estimate_revenue(streams_by_dsp):
    rates = {"spotify": 0.0035, "apple": 0.008, "youtube": 0.002}
    return sum(s * rates.get(dsp, 0.003) for dsp, s in streams_by_dsp.items())

매 결정 기준

상황	Channel
매 indie release	DistroKid / TuneCore → all DSPs
매 AI track	Disclosure flag + DSP의 AI policy 의 check
매 fan funding	Bandcamp + Patreon
매 viral hook	TikTok + Reels first
매 catalog track	Spotify editorial pitch + algorithmic playlist

기본값: 매 multi-DSP via aggregator + 매 TikTok seeding + 매 AI-disclosure transparent.

🔗 Graph

🤖 LLM 활용

언제: 매 catalog metadata cleaning, lyric generation, playlist description, 매 AI track의 disclosure draft. 언제 X: 매 royalty calculation 매 authoritative — 매 PRO/MLC official report 의 use.

❌ 안티패턴

No AI disclosure: 매 DSP TOS violation, takedown risk.
Album-only release in 2026: 매 algorithmic discovery 매 single 의 reward.
Ignoring TikTok hook: 매 discovery channel 의 miss.
Over-uploading filler AI tracks: 매 stream-farming flagged → ban.

🧪 검증 / 중복

Verified (RIAA 2025 mid-year, IFPI Global Music Report 2026; Spotify For Artists docs).
신뢰도 A.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — distribution timeline + AI-generation 2026 + embedding patterns

4.9 KiB Raw Blame History