feat: add multi-stem audio separation modes

experiments/bs-roformer-vocal-separation/Dockerfile

@@ -19,4 +19,4 @@ ENV MODEL_FILENAME=model_bs_roformer_ep_317_sdr_12.9755.ckpt \
     JOB_DIR=/data/jobs
 
 EXPOSE 8080
-CMD ["sh", "-c", "audio-separator --model_filename \"$MODEL_FILENAME\" --model_file_dir \"$MODEL_DIR\" --download_model_only && exec uvicorn app:app --host 0.0.0.0 --port 8080 --workers 1"]
+CMD ["sh", "-c", "for model in \"$MODEL_FILENAME\" htdemucs_ft.yaml htdemucs_6s.yaml; do audio-separator --model_filename \"$model\" --model_file_dir \"$MODEL_DIR\" --download_model_only || exit 1; done; exec uvicorn app:app --host 0.0.0.0 --port 8080 --workers 1"]

experiments/bs-roformer-vocal-separation/README.md

@@ -1,16 +1,23 @@
-# Athena Vocal Separator
+# Athena Stem Separator
 
-Exklusiver dritter Athena-Betriebsmodus für lokale Zwei-Spur-Trennung in
-`vocals.flac` und `instrumental.flac`.
+Exklusiver dritter Athena-Betriebsmodus für lokale Zwei- bis Sechs-Spur-Trennung.
 
 - Engine: `audio-separator` 0.47.0 (MIT)
-- Modell: BS-RoFormer Viperx 1297,
-  `model_bs_roformer_ep_317_sdr_12.9755.ckpt`
-- Modellbewertung im audio-separator-Katalog: Vocal SDR 12,9,
-  Instrumental SDR 17,0
+- **Gesang / Instrumental:** BS-RoFormer Viperx 1297,
+  `model_bs_roformer_ep_317_sdr_12.9755.ckpt`; Vocal SDR 12,9,
+  Instrumental SDR 17,0.
+- **4 Spuren:** `htdemucs_ft.yaml` mit Gesang, Schlagzeug, Bass und Rest.
+- **6 Spuren (experimentell):** `htdemucs_6s.yaml` ergänzt gemeinsame Gitarre
+  und Piano. Die Instrumentqualität liegt unter der spezialisierten
+  Zwei-Spur-Trennung.
 - GPU: RTX 5080; LLM, Bildmodelle, TTS und ACE-Step sind dabei verriegelt.
 - Privat erreichbar: `http://192.168.1.212:8007/`
 
-Das Modell wird beim ersten Start nach `/data/models/audio-separator`
+Die Modelle werden beim ersten Start nach `/data/models/audio-separator`
 heruntergeladen. Temporäre Jobs liegen unter `/data/audio/separation` und
-werden nach dem ZIP-Download entfernt.
+werden nach dem ZIP-Download entfernt. Eigene Spuren für E-/Akustikgitarre,
+Synthesizer und Streicher sind bewusst noch nicht angeboten: Dafür braucht es
+weitere Zielmodelle; sie aus `other` umzubenennen wäre fachlich falsch.
+
+Die API erwartet `multipart/form-data` mit `file` und optional `mode`:
+`vocals` (Standard), `four_stem` oder `six_stem`.

experiments/bs-roformer-vocal-separation/app.py

@@ -9,7 +9,7 @@ import time
 import zipfile
 from pathlib import Path
 
-from fastapi import FastAPI, File, HTTPException, UploadFile
+from fastapi import FastAPI, File, Form, HTTPException, UploadFile
 from fastapi.responses import FileResponse, HTMLResponse
 from starlette.background import BackgroundTask
 
@@ -22,7 +22,13 @@ ALLOWED = {".wav", ".flac", ".mp3", ".m4a", ".aac", ".ogg", ".opus", ".wma"}
 SEPARATION_LOCK = asyncio.Lock()
 STARTED = time.time()
 
-app = FastAPI(title="Athena Vocal Separator", version="1.0")
+MODES = {
+    "vocals": {"model": MODEL, "stems": ("vocals", "instrumental"), "archive": "athena-vocals-instrumental.zip", "engine": "mdxc"},
+    "four_stem": {"model": "htdemucs_ft.yaml", "stems": ("vocals", "drums", "bass", "other"), "archive": "athena-4-stems.zip", "engine": "demucs"},
+    "six_stem": {"model": "htdemucs_6s.yaml", "stems": ("vocals", "drums", "bass", "guitar", "piano", "other"), "archive": "athena-6-stems-experimental.zip", "engine": "demucs"},
+}
+
+app = FastAPI(title="Athena Stem Separator", version="2.0")
 
 
 @app.get("/", response_class=HTMLResponse)
@@ -32,11 +38,11 @@ def index() -> str:
 
 @app.get("/health")
 def health() -> dict:
-    checkpoint = MODEL_DIR / MODEL
+    available = {name: (MODEL_DIR / mode["model"]).exists() for name, mode in MODES.items()}
     return {
-        "status": "ok" if checkpoint.exists() else "starting",
-        "model": MODEL,
-        "model_ready": checkpoint.exists(),
+        "status": "ok" if all(available.values()) else "starting",
+        "models": {name: mode["model"] for name, mode in MODES.items()},
+        "models_ready": available,
         "busy": SEPARATION_LOCK.locked(),
         "uptime_seconds": round(time.time() - STARTED, 1),
     }
@@ -46,28 +52,40 @@ def _cleanup(path: Path) -> None:
     shutil.rmtree(path, ignore_errors=True)
 
 
-def _run_separator(input_path: Path, output_dir: Path) -> None:
+def _run_separator(input_path: Path, output_dir: Path, mode: dict) -> None:
     args = [
         "audio-separator", str(input_path),
-        "--model_filename", MODEL,
+        "--model_filename", mode["model"],
         "--model_file_dir", str(MODEL_DIR),
         "--output_dir", str(output_dir),
         "--output_format", "FLAC",
         "--sample_rate", "44100",
         "--use_soundfile",
         "--use_autocast",
-        "--mdxc_segment_size", "256",
-        "--mdxc_overlap", "8",
-        "--mdxc_batch_size", "1",
     ]
+    if mode["engine"] == "mdxc":
+        args.extend(["--mdxc_segment_size", "256", "--mdxc_overlap", "8", "--mdxc_batch_size", "1"])
+    else:
+        args.extend(["--demucs_segment_size", "40", "--demucs_shifts", "2", "--demucs_overlap", "0.25"])
     completed = subprocess.run(args, capture_output=True, text=True, timeout=7200)
     if completed.returncode:
         detail = (completed.stderr or completed.stdout or "unknown error")[-4000:]
         raise RuntimeError(detail)
 
 
+def _stem_name(path: Path, expected: tuple[str, ...]) -> str | None:
+    lower = path.stem.lower()
+    for stem in sorted(expected, key=len, reverse=True):
+        if stem in lower:
+            return stem
+    return None
+
+
 @app.post("/v1/separate")
-async def separate(file: UploadFile = File(...)) -> FileResponse:
+async def separate(file: UploadFile = File(...), mode: str = Form("vocals")) -> FileResponse:
+    selected_mode = MODES.get(mode)
+    if selected_mode is None:
+        raise HTTPException(422, f"Unbekannter Trennmodus: {mode}")
     suffix = Path(file.filename or "upload.wav").suffix.lower()
     if suffix not in ALLOWED:
         raise HTTPException(415, "Dieses Audioformat wird nicht unterstützt.")
@@ -87,21 +105,24 @@ async def separate(file: UploadFile = File(...)) -> FileResponse:
                     raise HTTPException(413, "Datei ist größer als 1 GiB.")
                 handle.write(chunk)
         async with SEPARATION_LOCK:
-            await asyncio.to_thread(_run_separator, input_path, output_dir)
+            await asyncio.to_thread(_run_separator, input_path, output_dir, selected_mode)
 
         stems = sorted(output_dir.glob("*.flac"))
-        if len(stems) != 2:
-            raise RuntimeError(f"Erwartet wurden zwei FLAC-Dateien, gefunden: {len(stems)}")
-        archive = job / "athena-vocals-instrumental.zip"
+        expected = selected_mode["stems"]
+        recognized = {_stem_name(stem, expected): stem for stem in stems}
+        recognized.pop(None, None)
+        missing = [stem for stem in expected if stem not in recognized]
+        if missing:
+            found = ", ".join(stem.name for stem in stems) or "keine"
+            raise RuntimeError(f"Fehlende Spuren: {', '.join(missing)}; gefunden: {found}")
+        archive = job / selected_mode["archive"]
         with zipfile.ZipFile(archive, "w", compression=zipfile.ZIP_STORED) as bundle:
-            for stem in stems:
-                lower = stem.name.lower()
-                target = "vocals.flac" if "vocal" in lower else "instrumental.flac"
-                bundle.write(stem, target)
+            for stem in expected:
+                bundle.write(recognized[stem], f"{stem}.flac")
         return FileResponse(
             archive,
             media_type="application/zip",
-            filename="athena-vocals-instrumental.zip",
+            filename=selected_mode["archive"],
             background=BackgroundTask(_cleanup, job),
         )
     except HTTPException:

experiments/bs-roformer-vocal-separation/index.html

@@ -1,4 +1,9 @@
-<!doctype html><html lang="de"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><title>Athena · Stimmen trennen</title><style>
-:root{color-scheme:dark;--bg:#07111c;--card:#101d2b;--line:#26384b;--cyan:#48d7f5;--mint:#63e6be;--text:#ecf5ff;--muted:#91a4b7}*{box-sizing:border-box}body{margin:0;background:radial-gradient(circle at 20% 0,#142a42 0,#07111c 42%);font:16px system-ui,sans-serif;color:var(--text);min-height:100vh;display:grid;place-items:center;padding:24px}.card{width:min(760px,100%);padding:32px;border:1px solid var(--line);border-radius:22px;background:rgba(16,29,43,.96);box-shadow:0 25px 70px #0008}.eyebrow{color:var(--cyan);font-weight:800;letter-spacing:.14em;text-transform:uppercase;font-size:12px}h1{font-size:clamp(30px,5vw,52px);margin:.3em 0 .15em}p{color:var(--muted);line-height:1.6}.drop{display:block;margin:28px 0;padding:44px 24px;border:2px dashed #3f5a72;border-radius:18px;text-align:center;cursor:pointer;transition:.2s}.drop:hover,.drop.drag{border-color:var(--cyan);background:#48d7f50b}input{display:none}.file{color:var(--mint);font-weight:700;margin-top:8px}button{width:100%;border:0;border-radius:13px;padding:15px;font-weight:800;font-size:16px;background:linear-gradient(90deg,var(--cyan),var(--mint));color:#05202a;cursor:pointer}button:disabled{opacity:.45;cursor:not-allowed}.status{min-height:28px;margin-top:18px;color:var(--muted)}.bar{height:7px;background:#07111c;border-radius:9px;overflow:hidden;margin-top:12px}.fill{height:100%;width:0;background:linear-gradient(90deg,var(--cyan),var(--mint));transition:.4s}.run .fill{width:85%;animation:pulse 1.5s infinite alternate}@keyframes pulse{to{opacity:.45}}small{display:block;color:#71879a;margin-top:20px}</style></head><body><main class="card"><div class="eyebrow">Athena Audio Lab</div><h1>Stimmen sauber trennen</h1><p>BS‑RoFormer Viperx 1297 zerlegt deinen Titel in zwei verlustfreie FLAC-Spuren: Gesang und Instrumental.</p><label class="drop" id="drop">Audio auswählen oder hier ablegen<input id="file" type="file" accept="audio/*"><div class="file" id="name">Noch keine Datei gewählt</div></label><button id="start" disabled>Vocals und Instrumental erzeugen</button><div class="status" id="status">Bereit.</div><div class="bar" id="bar"><div class="fill"></div></div><small>Die Verarbeitung läuft lokal auf Athena. Nichts wird in eine Cloud hochgeladen.</small></main><script>
-const file=document.querySelector('#file'),drop=document.querySelector('#drop'),name=document.querySelector('#name'),start=document.querySelector('#start'),status=document.querySelector('#status'),bar=document.querySelector('#bar');let selected;function choose(f){selected=f;name.textContent=f?`${f.name} · ${(f.size/1048576).toFixed(1)} MiB`:'Noch keine Datei gewählt';start.disabled=!f}file.onchange=()=>choose(file.files[0]);drop.ondragover=e=>{e.preventDefault();drop.classList.add('drag')};drop.ondragleave=()=>drop.classList.remove('drag');drop.ondrop=e=>{e.preventDefault();drop.classList.remove('drag');choose(e.dataTransfer.files[0])};start.onclick=async()=>{start.disabled=true;bar.classList.add('run');status.textContent='Modell trennt den Titel – das kann einige Minuten dauern …';let body=new FormData();body.append('file',selected);try{let r=await fetch('/v1/separate',{method:'POST',body});if(!r.ok)throw Error((await r.json()).detail||`HTTP ${r.status}`);let blob=await r.blob(),a=document.createElement('a');a.href=URL.createObjectURL(blob);a.download='athena-vocals-instrumental.zip';a.click();setTimeout(()=>URL.revokeObjectURL(a.href),5000);status.textContent='Fertig – ZIP mit vocals.flac und instrumental.flac wurde geladen.'}catch(e){status.textContent=`Fehler: ${e.message}`}finally{bar.classList.remove('run');start.disabled=false}};
+<!doctype html>
+<html lang="de"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
+<title>Athena · Spuren trennen</title><style>
+:root{color-scheme:dark;--bg:#07111c;--card:#101d2b;--line:#26384b;--cyan:#48d7f5;--mint:#63e6be;--text:#ecf5ff;--muted:#91a4b7}*{box-sizing:border-box}body{margin:0;background:radial-gradient(circle at 20% 0,#142a42 0,#07111c 42%);font:16px system-ui,sans-serif;color:var(--text);min-height:100vh;display:grid;place-items:center;padding:24px}.card{width:min(820px,100%);padding:32px;border:1px solid var(--line);border-radius:22px;background:rgba(16,29,43,.96);box-shadow:0 25px 70px #0008}.eyebrow{color:var(--cyan);font-weight:800;letter-spacing:.14em;text-transform:uppercase;font-size:12px}h1{font-size:clamp(30px,5vw,52px);margin:.3em 0 .15em}p{color:var(--muted);line-height:1.6}.modes{display:grid;grid-template-columns:repeat(3,1fr);gap:10px;margin:24px 0}.mode{display:block;border:1px solid var(--line);border-radius:14px;padding:14px;cursor:pointer}.mode:has(input:checked){border-color:var(--cyan);background:#48d7f510}.mode input{display:none}.mode b,.mode span{display:block}.mode span{color:var(--muted);font-size:13px;margin-top:5px;line-height:1.35}.drop{display:block;margin:20px 0;padding:36px 24px;border:2px dashed #3f5a72;border-radius:18px;text-align:center;cursor:pointer;transition:.2s}.drop:hover,.drop.drag{border-color:var(--cyan);background:#48d7f50b}.drop input{display:none}.file{color:var(--mint);font-weight:700;margin-top:8px}button{width:100%;border:0;border-radius:13px;padding:15px;font-weight:800;font-size:16px;background:linear-gradient(90deg,var(--cyan),var(--mint));color:#05202a;cursor:pointer}button:disabled{opacity:.45;cursor:not-allowed}.status{min-height:28px;margin-top:18px;color:var(--muted)}.bar{height:7px;background:#07111c;border-radius:9px;overflow:hidden;margin-top:12px}.fill{height:100%;width:0;background:linear-gradient(90deg,var(--cyan),var(--mint));transition:.4s}.run .fill{width:85%;animation:pulse 1.5s infinite alternate}@keyframes pulse{to{opacity:.45}}small{display:block;color:#71879a;margin-top:20px}@media(max-width:680px){.modes{grid-template-columns:1fr}}
+</style></head><body><main class="card"><div class="eyebrow">Athena Audio Lab</div><h1>Instrumente und Stimmen trennen</h1><p>Wähle zwischen der besonders sauberen Gesangstrennung und zusammengehörigen Mehrspur-Modellen.</p>
+<div class="modes"><label class="mode"><input type="radio" name="mode" value="vocals" checked><b>Gesang · beste Qualität</b><span>Vocals + Instrumental<br>BS‑RoFormer</span></label><label class="mode"><input type="radio" name="mode" value="four_stem"><b>4 Spuren</b><span>Gesang, Schlagzeug, Bass, Rest<br>HTDemucs FT</span></label><label class="mode"><input type="radio" name="mode" value="six_stem"><b>6 Spuren · experimentell</b><span>Zusätzlich Gitarre und Piano<br>HTDemucs 6s</span></label></div>
+<label class="drop" id="drop">Audio auswählen oder hier ablegen<input id="file" type="file" accept="audio/*"><div class="file" id="name">Noch keine Datei gewählt</div></label><button id="start" disabled>Spuren erzeugen</button><div class="status" id="status">Bereit.</div><div class="bar" id="bar"><div class="fill"></div></div><small>Alles läuft lokal auf Athena. Synthesizer, Streicher sowie elektrische und akustische Gitarre separat benötigen zusätzliche Spezialmodelle und sind noch nicht freigeschaltet.</small></main><script>
+const file=document.querySelector('#file'),drop=document.querySelector('#drop'),name=document.querySelector('#name'),start=document.querySelector('#start'),status=document.querySelector('#status'),bar=document.querySelector('#bar');let selected;const info={vocals:['athena-vocals-instrumental.zip','vocals.flac und instrumental.flac'],four_stem:['athena-4-stems.zip','vier Spuren'],six_stem:['athena-6-stems-experimental.zip','sechs Spuren']};function choose(f){selected=f;name.textContent=f?`${f.name} · ${(f.size/1048576).toFixed(1)} MiB`:'Noch keine Datei gewählt';start.disabled=!f}file.onchange=()=>choose(file.files[0]);drop.ondragover=e=>{e.preventDefault();drop.classList.add('drag')};drop.ondragleave=()=>drop.classList.remove('drag');drop.ondrop=e=>{e.preventDefault();drop.classList.remove('drag');choose(e.dataTransfer.files[0])};start.onclick=async()=>{const mode=document.querySelector('input[name=mode]:checked').value;start.disabled=true;bar.classList.add('run');status.textContent='Modell trennt den Titel – das kann einige Minuten dauern …';let body=new FormData();body.append('file',selected);body.append('mode',mode);try{let r=await fetch('/v1/separate',{method:'POST',body});if(!r.ok)throw Error((await r.json()).detail||`HTTP ${r.status}`);let blob=await r.blob(),a=document.createElement('a');a.href=URL.createObjectURL(blob);a.download=info[mode][0];a.click();setTimeout(()=>URL.revokeObjectURL(a.href),5000);status.textContent=`Fertig – ZIP mit ${info[mode][1]} wurde geladen.`}catch(e){status.textContent=`Fehler: ${e.message}`}finally{bar.classList.remove('run');start.disabled=false}};
 </script></body></html>