diff --git a/backend/lib/ytdlp.js b/backend/lib/ytdlp.js
index 98f99c4..50b817a 100644
--- a/backend/lib/ytdlp.js
+++ b/backend/lib/ytdlp.js
@@ -100,7 +100,9 @@ async function downloadTrack({ jobId, artist, albumArtist, title, album, year, t
   });
 
   const YTM_SONGS_PARAM = 'EgWKAQIIAWoKEAkQBRAKEAMQBA%3D%3D';
-  const ytmUrl = `https://music.youtube.com/search?q=${encodeURIComponent(query)}&params=${YTM_SONGS_PARAM}`;
+  // Use original artist name (with special chars) for YTM URL — better search ranking than stripped version
+  const ytmQuery = `${mainArtist} - ${title}`;
+  const ytmUrl = `https://music.youtube.com/search?q=${encodeURIComponent(ytmQuery)}&params=${YTM_SONGS_PARAM}`;
   const expectedAudio = path.join(dir, `${trackPrefix}${safeTitle}.${CONFIG.audioFmt}`);
   const hasAudio = () => fs.existsSync(expectedAudio);
   const tryYtdlp = async (args) => { await spawnYtdlp(args); };
@@ -155,10 +157,18 @@ async function downloadTrack({ jobId, artist, albumArtist, title, album, year, t
   const variantPart   = variantFilter ? ` & ${variantFilter}` : '';
   const titleKwPart   = titleKws.length ? titleKws.map(k => ` & title ~= (?i)${k}`).join('') : '';
 
-  const mf = (dur, variant) => {
+  // Soft channel check: first word of original artist name used in early cascade steps to avoid
+  // downloading a same-title song by a different artist (e.g. wrong "Moonlight" for N!GHT).
+  // Dropped in later fallback steps so genuine mismatches (label channels etc.) still get found.
+  const escRe = s => s.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+  const artistFirstWord = mainArtist.split(/[\s,/]+/)[0];
+  const artistChannelPart = artistFirstWord.length >= 3
+    ? ` & channel ~= (?i)${escRe(artistFirstWord)}` : '';
+
+  const mf = (dur, variant, strict = false) => {
     const durPart = dur ? `duration > ${dur.min} & duration < ${dur.max} & ` : '';
     return ['--match-filter',
-      `${durPart}!is_live${titleKwPart}${variant ? variantPart : ''}`];
+      `${durPart}!is_live${titleKwPart}${strict ? artistChannelPart : ''}${variant ? variantPart : ''}`];
   };
 
   const PI_FILTERED    = ['--playlist-items', '1-5',  '--max-downloads', '1'];
@@ -171,20 +181,20 @@ async function downloadTrack({ jobId, artist, albumArtist, title, album, year, t
     const dur = { min: Math.max(30, durationSec - margin), max: durationSec + margin };
 
     log(jobId, `🔍 Suche YT Music (~${durationSec}s ±${margin}s): "${query}"`);
-    // 1. YTM top-5 — pre-filtered URLs (no channel expansion), duration + title + variant filter
-    await tryYtmSearch(5, mf(dur, true));
+    // 1. YTM top-5 — strict: duration + title + variant + channel check
+    await tryYtmSearch(5, mf(dur, true, true));
     if (!hasAudio()) {
-      // 2. YT top-5 — ytsearch always returns videos (no channels), safe to use directly
+      // 2. YT top-5 — strict: same filters including channel check
       log(jobId, '⚠️  Kein YTM-Treffer — Fallback zu YouTube…');
-      await tryYtdlp([`ytsearch5:${query}`, ...PI_FILTERED, ...mf(dur, true), ...baseArgs]);
+      await tryYtdlp([`ytsearch5:${query}`, ...PI_FILTERED, ...mf(dur, true, true), ...baseArgs]);
     }
     if (!hasAudio()) {
-      // 3. YTM top-10 — wider net
+      // 3. YTM top-10 — no channel check (label channels, etc.)
       log(jobId, '⚠️  Erweitere Suche auf top 10 mit Dauerbeschränkung…');
       await tryYtmSearch(10, mf(dur, true));
     }
     if (!hasAudio()) {
-      // 4. YT top-10
+      // 4. YT top-10 — no channel check
       await tryYtdlp([`ytsearch10:${query}`, ...PI_FILTERED_10, ...mf(dur, true), ...baseArgs]);
     }
     if (!hasAudio()) {
@@ -216,28 +226,36 @@ async function downloadTrack({ jobId, artist, albumArtist, title, album, year, t
     }
   } else {
     log(jobId, `🔍 Suche YT Music: "${query}"`);
-    // 1. YTM top-5 — pre-filtered, title + variant
-    await tryYtmSearch(5, mf(null, true));
+    // 1. YTM top-5 — strict: title + variant + channel check
+    await tryYtmSearch(5, mf(null, true, true));
     if (!hasAudio()) {
-      // 2. YT top-5
+      // 2. YT top-5 — strict
+      await tryYtdlp([`ytsearch5:${query}`, ...PI_FILTERED, ...mf(null, true, true), ...baseArgs]);
+    }
+    if (!hasAudio()) {
+      // 3. YTM top-5 — no channel check
+      await tryYtmSearch(5, mf(null, true));
+    }
+    if (!hasAudio()) {
+      // 4. YT top-5 — no channel check
       await tryYtdlp([`ytsearch5:${query}`, ...PI_FILTERED, ...mf(null, true), ...baseArgs]);
     }
     if (!hasAudio()) {
-      // 3. YTM top-5 — title only
+      // 5. YTM top-5 — title only
       log(jobId, '⚠️  Letzter gefilterter Versuch…');
       await tryYtmSearch(5, mf(null, false));
     }
     if (!hasAudio()) {
-      // 4. YT top-5 — title only
+      // 6. YT top-5 — title only
       await tryYtdlp([`ytsearch5:${query}`, ...PI_FILTERED, ...mf(null, false), ...baseArgs]);
     }
     if (!hasAudio()) {
-      // 5. YTM — first video result, no filter
+      // 7. YTM — first video result, no filter
       log(jobId, '⚠️  Letzter Versuch: keine Filter…');
       await tryYtmSearch(3, []);
     }
     if (!hasAudio()) {
-      // 6. YT — no filter
+      // 8. YT — no filter
       await tryYtdlp([`ytsearch1:${query}`, ...PI_UNFILTERED, ...baseArgs]);
     }
   }