<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:media="http://search.yahoo.com/mrss/">
<channel>
  <title>Vocaler Pulse</title>
  <link>https://pulse.vocaler.ai/</link>
  <atom:link href="https://pulse.vocaler.ai/feed.xml" rel="self" type="application/rss+xml"/>
  <description>Everything changing in AI music, in one place.</description>
  <language>en</language>
  <image><url>https://pulse.vocaler.ai/og/home.png</url><title>Vocaler Pulse</title><link>https://pulse.vocaler.ai/</link></image>
  <lastBuildDate>Fri, 20 Nov 2026 08:00:00 +0000</lastBuildDate>
  <item>
    <title>Eminem&#x27;s publisher expands Suno suit as label litigation wave grows</title>
    <link>https://pulse.vocaler.ai/news/eight-mile-style-suno-amended-complaint/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/news/eight-mile-style-suno-amended-complaint/</guid>
    <pubDate>Fri, 20 Nov 2026 08:00:00 +0000</pubDate>
    <category>News</category><category>suno</category><category>lawsuits</category><category>copyright</category><category>labels</category>
    <description>Eight Mile Style&#x27;s amended complaint adds 108 alleged infringing outputs and NIL-style voice claims as more publishers pile on Suno.</description>
    <content:encoded><![CDATA[<h2 id="the-filing">The filing</h2>
<p>On November 19, 2026 Eight Mile Style, Eminem's publisher, filed an <strong>amended complaint</strong> against Suno in U.S. District Court (case no. 3:26-cv-13399). According to Digital Music News:</p>
<ul><li>The suit covers <strong>269 Eminem compositions</strong>. It claims copyright infringement, violations of the Digital Millennium Copyright Act (DMCA), and unauthorized imitation of the artist's voice.</li><li>The amendment adds <strong>108 more examples of allegedly infringing outputs</strong> across 26 of the works. They include prompts framed as a "DJ's take," a "heavy alternative rock version" and "jazz" versions, each said to use protected lyrics. The examples are collected in a public Vimeo folder.</li><li>Total damages sought are reported at <strong>$47.8 million</strong> across the infringement and DMCA claims.</li></ul>
<h2 id="a-widening-front">A widening front</h2>
<p>Eight Mile is not alone. Similar suits against Suno have come from <strong>Elvis Presley Enterprises, Reservoir, Gerencia 360 and Round Hill Music</strong>, all represented by attorney Richard S. Busch. Bright Data, first named as a defendant for supplying scraped training data, has since been dropped from these suits.</p>
<p>Separately, in the long-running major-label case, <strong>Universal Music Group and Sony Music withdrew their motion to force DistroKid to comply with a subpoena</strong>. They did so after the distributor handed over documents about its AI-detection tool, as Music Business Worldwide reported on November 17.</p>
<h2 id="context">Context</h2>
<p>Warner Music Group settled with Suno in late 2025, and Suno said it would move to licensed models. These newer suits come from publishers and estates outside that deal. The focus has shifted from training data alone to what users can make the model output: lyrics, styles and voices that resemble real artists.</p>
<h2 id="why-it-matters-for-creators">Why it matters for creators</h2>
<ul><li><strong>Don't prompt with real artists or their lyrics.</strong> The complaints centre on outputs that reproduce protected lyrics or a recognisable voice. Tracks like that are the most exposed to takedowns.</li><li><strong>Detection data is now evidence.</strong> DistroKid's AI-detection documents are part of the litigation record. Expect distributors to screen AI uploads more closely, not less.</li><li><strong>Owning a song isn't immunity.</strong> A paid Suno plan gives you ownership under Suno's terms. It does not protect a track that copies someone else's work.</li><li><strong>Keep your receipts.</strong> Save your prompts, lyrics drafts and project files. They show your own creative input if a song is ever challenged.</li></ul>
<p><em>This is a summary of court filings and press reports, not legal advice.</em></p>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/news-eight-mile-style-suno-amended-complaint.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/news-eight-mile-style-suno-amended-complaint.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Suno launches Visuals: 15-second lyric videos, clips and album art</title>
    <link>https://pulse.vocaler.ai/news/suno-visuals-launch-music-video-generation/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/news/suno-visuals-launch-music-video-generation/</guid>
    <pubDate>Tue, 17 Nov 2026 08:00:00 +0000</pubDate>
    <category>News</category><category>suno</category><category>video</category><category>product launch</category><category>pricing</category>
    <description>Suno&#x27;s new Visuals tool makes lyric videos, album art and lip-synced clips up to 15 seconds from any Suno song, on its own credit system.</description>
    <content:encoded><![CDATA[<h2 id="what-launched">What launched</h2>
<p>On November 17, 2026 Suno released <strong>Visuals</strong>. Suno's release notes describe it as a way to "turn any Suno song into a video clip, a lyrics video, or create a whole world for your music." Reported features:</p>
<ul><li><strong>Lyric videos, album art and short video clips</strong> generated from a song you already made.</li><li><strong>Clips up to 15 seconds</strong> that follow the song's structure and can lip-sync to the vocal.</li><li><strong>A starting point:</strong> Suno's built-in character library or your own uploaded image. You then edit or extend clips with text prompts.</li><li><strong>Separate credits.</strong> Visuals credits are bought and spent apart from song credits.</li><li><strong>Hidden watermarks</strong> go on generated videos before you share them.</li></ul>
<h2 id="under-the-hood">Under the hood</h2>
<p>Digital Music News reports that Visuals runs on a model built on MiniMax's H3 technology, which Suno fine-tuned on its own data for music-video work. MiniMax is being sued by Disney, Universal and Warner Bros. Discovery, which is why some commentators expect studio pushback.</p>
<h2 id="guardrails">Guardrails</h2>
<p>Suno added text and image filters that block prompts trying to show specific artists, public figures or copyrighted characters. The company warns that "prompts that work on other video platforms may not work on Suno, and that's by design."</p>
<h2 id="new-top-tier">New top tier</h2>
<p>Alongside Visuals, Suno added a <strong>Platinum</strong> plan. It costs $120 a month, or $100 a month billed annually, and is reported to include 20,000 visual credits plus 15,000 song credits. That is four times the $30 Premier plan.</p>
<h2 id="availability">Availability</h2>
<p>At launch, Visuals is on the web. Suno's release notes list both web and mobile, while press coverage at launch said the iOS and Android versions would follow in the coming weeks.</p>
<h2 id="why-it-matters-for-creators">Why it matters for creators</h2>
<ul><li><strong>One app for release assets.</strong> Short clips for Shorts, Reels and TikTok, lyric videos and cover art can now come from the same place as the song.</li><li><strong>Watch the credit maths.</strong> Visuals uses its own credits, so a full rollout campaign may push you toward Platinum. Price it before you commit.</li><li><strong>15 seconds is a teaser length.</strong> Full-length music videos still need a dedicated video tool or editing several clips together.</li><li><strong>No real faces.</strong> The likeness filters mean you can't put a real artist in a clip. Plan to use original characters or your own uploaded art.</li><li><strong>Watermarks travel with the file.</strong> Assume platforms can identify Visuals output as AI-generated, and disclose it where a platform asks (for example YouTube's altered or synthetic content label).</li></ul>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/news-suno-visuals-launch-music-video-generation.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/news-suno-visuals-launch-music-video-generation.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Suno Visuals vs ElevenLabs Instant Soundtracks: song-to-video vs video-to-score</title>
    <link>https://pulse.vocaler.ai/compare/suno-visuals-vs-elevenlabs-instant-soundtracks/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/compare/suno-visuals-vs-elevenlabs-instant-soundtracks/</guid>
    <pubDate>Tue, 17 Nov 2026 08:00:00 +0000</pubDate>
    <category>Compare</category><category>suno</category><category>elevenlabs</category><category>video</category><category>comparison</category>
    <description>Suno now turns songs into clips; ElevenLabs turns clips into scores. Here&#x27;s how the two new audiovisual tools compare for creators.</description>
    <content:encoded><![CDATA[<h2 id="opposite-directions">Opposite directions</h2>
<p>Two of the biggest AI audio companies shipped audiovisual tools within five weeks of each other, and they work in opposite directions:</p>
<ul><li><strong>Suno Visuals</strong> (November 17, 2026) starts from a <strong>song</strong> and generates <strong>video</strong>: lyric videos, album art and short clips.</li><li><strong>ElevenLabs Instant Soundtracks</strong> (October 15, 2026) starts from a <strong>video</strong> and generates <strong>music and sound effects</strong>.</li></ul>
<h2 id="side-by-side">Side by side</h2>
<div class="tw"><table><thead><tr><th>Criterion</th><th>Suno Visuals</th><th>ElevenLabs Instant Soundtracks</th></tr></thead><tbody><tr><td>Input</td><td>A Suno song plus character library or your image</td><td>Your uploaded video (and optionally its audio)</td></tr><tr><td>Output</td><td>Video clips, lyric videos, album art</td><td>Original score plus timed SFX</td></tr><tr><td>Length limit</td><td>Clips up to 15 s</td><td>Up to 10 files / 200 MB per API request; reported 600 s combined footage</td></tr><tr><td>Editing</td><td>Edit or extend clips with text prompts</td><td>SFX on separate layers; music and SFX exportable separately</td></tr><tr><td>Lip-sync / timing</td><td>Lip-syncs and follows song structure</td><td>Places SFX at detected timestamps</td></tr><tr><td>Surfaces</td><td>Web (mobile listed / rolling out)</td><td>ElevenCreative, ElevenMusic, Video to Music API</td></tr><tr><td>API</td><td>Not announced</td><td>Yes (<code>/v1/music/video-to-music</code>)</td></tr><tr><td>Billing</td><td>Separate Visuals credits; Platinum plan $120/mo</td><td>Plan credits; API needs a paid plan</td></tr><tr><td>Guardrails</td><td>Blocks real artists, public figures, copyrighted characters; hidden watermark</td><td>No artist names or lyrics in prompts; Enterprise needed for film/TV/radio</td></tr><tr><td>Underlying model</td><td>Built on MiniMax H3 tech, fine-tuned by Suno</td><td>ElevenLabs Music models (version used: unknown)</td></tr></tbody></table></div>
<h2 id="who-should-use-which">Who should use which</h2>
<ul><li><strong>Musicians promoting a release:</strong> Suno Visuals. If the song already lives in Suno, you can make Shorts and Reels teasers, lyric videos and cover art without leaving the app. The 15-second cap suits social media teasers, not full music videos.</li><li><strong>Video editors, YouTubers and indie game trailer makers:</strong> Instant Soundtracks. Score a locked cut, then fix the effects layer by layer. API flags such as <code>use_audio_as_reference</code> and <code>separate_tracks</code> make it suitable for a pipeline.</li></ul>
<h2 id="watch-outs">Watch-outs</h2>
<ul><li><strong>Rights differ.</strong> ElevenLabs' self-serve plans exclude film, TV, radio and studio-game use. Suno Visuals runs on a model whose maker, MiniMax, is being sued by Hollywood studios.</li><li><strong>Disclosure.</strong> Both produce AI media. Use YouTube's altered or synthetic content label, or a platform's equivalent, where it applies.</li><li><strong>Cost stacking.</strong> Suno video credits are separate from song credits. Budget for both if you're making a whole campaign.</li></ul>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/compare-suno-visuals-vs-elevenlabs-instant-soundtracks.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/compare-suno-visuals-vs-elevenlabs-instant-soundtracks.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Suno Bridge in Ableton Live 12: generate stems that fit your session</title>
    <link>https://pulse.vocaler.ai/guides/suno-bridge-ableton-live-12-workflow/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/guides/suno-bridge-ableton-live-12-workflow/</guid>
    <pubDate>Fri, 13 Nov 2026 08:00:00 +0000</pubDate>
    <category>Guides</category><category>suno bridge</category><category>ableton live</category><category>stems</category><category>workflow</category>
    <description>A step-by-step workflow for Suno&#x27;s beta Bridge app: syncing an Ableton set, setting loop windows, prompting parts and protecting your session.</description>
    <content:encoded><![CDATA[<h2 id="what-you-need">What you need</h2>
<ul><li><strong>A Mac</strong> running <strong>Ableton Live 12</strong>. These are the only supported setups in the beta, which went live on November 13, 2026.</li><li><strong>A Suno Premier subscription.</strong> Bridge is included with Premier, which is listed at $30 a month with 10,000 credits.</li><li><strong>Credits.</strong> Each Bridge generation is reported to cost <strong>10 Suno credits</strong>.</li></ul>
<h2 id="setup">Setup</h2>
<ol><li>Download the Suno Bridge app from suno.com/bridge and sign in with your Premier account.</li><li>Open <strong>both</strong> Bridge and Ableton Live 12. Bridge runs alongside your DAW, not as a plugin.</li><li><strong>Sync your session.</strong> Bridge uploads your Ableton project and its audio files to your Suno account so it can read the tempo, key, harmony and groove.</li></ol>
<h2 id="the-core-loop">The core loop</h2>
<ol><li><strong>Set your prompting window</strong> in Arrangement View using Ableton's loop start and end markers. Suno's own example is a loop from bar 9 to bar 16, which gives an 8-bar window starting on the first beat of bar 9.</li><li><strong>Describe the part</strong> you want in Bridge, for example <em>"warm Rhodes chords, laid-back, sparse voicings"</em> or <em>"airy female backing vocal, oohs, no lyrics."</em></li><li><strong>Generate and audition.</strong> Bridge places instrument or vocal stems on your timeline, matched to the session's tempo, key, style and feel.</li><li><strong>Keep, regenerate or edit.</strong> Treat each result as a take and comp it with Live's normal clip and warp tools.</li></ol>
<h2 id="tips-for-cleaner-results">Tips for cleaner results</h2>
<ul><li><strong>Prompt one part at a time</strong> (bass, then pads, then backing vocals), so each stem has one job.</li><li><strong>Keep windows short</strong> (4 to 8 bars) while exploring. Short windows are cheaper to iterate on, and you can extend a part once it works.</li><li><strong>Name the role, not just the sound,</strong> for example <em>"counter-melody answering the vocal in bars 13–16."</em></li></ul>
<h2 id="protect-your-session">Protect your session</h2>
<ul><li><strong>Your project leaves your laptop.</strong> Bridge uploads the project you open in it. Don't open unreleased client work, co-writes or anything you don't fully control without checking Suno's Terms of Service and Privacy Notice.</li><li><strong>Third-party samples may be filtered.</strong> Suno says uploads pass copyright filters, and flagged files are not ingested. A session full of commercial samples may not sync fully.</li><li><strong>It's a beta.</strong> Save versions of your Live Set before each Bridge session.</li></ul>
<h2 id="what-s-next">What's next</h2>
<p>Suno says more operating systems and DAWs will follow, but hasn't named which.</p>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/guides-suno-bridge-ableton-live-12-workflow.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/guides-suno-bridge-ableton-live-12-workflow.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Lab: How well does Suno Bridge lock to an Ableton session?</title>
    <link>https://pulse.vocaler.ai/lab/lab-suno-bridge-tempo-key-lock-test/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/lab/lab-suno-bridge-tempo-key-lock-test/</guid>
    <pubDate>Fri, 13 Nov 2026 08:00:00 +0000</pubDate>
    <category>Lab</category><category>suno bridge</category><category>ableton live</category><category>experiment</category>
    <description>A planned stress test of Bridge&#x27;s tempo, key and groove matching across odd meters, swing and key changes. The results shown are mock.</description>
    <content:encoded><![CDATA[<h2 id="hypothesis">Hypothesis</h2>
<p>Suno says Bridge output will "fit the tempo, key, style, harmonic content and groove" of a session. We expect near-perfect results on straight 4/4 pop, but drift on <strong>swing, odd meters and mid-loop key changes</strong>.</p>
<h2 id="setup">Setup</h2>
<ul><li><strong>Rig:</strong> Mac, Ableton Live 12, Suno Premier, Bridge beta. Each generation costs about 10 credits, so the test needs a budget of roughly 400 credits.</li><li><strong>Five test sessions</strong>, with an 8-bar loop window each:</li><li>120 BPM, 4/4, C minor house groove (control)</li><li>92 BPM hip-hop with 62% swing</li><li>7/8 math-rock riff at 140 BPM</li><li>6/8 folk ballad at 66 BPM</li><li>4/4 pop with a key change from G to A at bar 5 of the loop</li><li><strong>Prompt (identical for all sessions):</strong> <code>&quot;melodic bassline that locks to the kick, simple, supportive&quot;</code>, then a second pass with <code>&quot;airy backing vocal pad, oohs, no lyrics&quot;</code>.</li><li><strong>Eight takes per prompt per session</strong> (80 generations in total).</li><li><strong>Measures:</strong> onset offset from the grid (ms), key correctness (checked with a key detector and by ear), and a blind 1–5 "sits in the mix" score.</li></ul>
<h2 id="mock-findings-illustrative-only-not-real-results">Mock findings (illustrative only, not real results)</h2>
<div class="tw"><table><thead><tr><th>Session</th><th>Grid offset (median)</th><th>Key correct</th><th>Fit score</th></tr></thead><tbody><tr><td>House 4/4</td><td>6 ms</td><td>8/8</td><td>4.5</td></tr><tr><td>Swung hip-hop</td><td>18 ms</td><td>8/8</td><td>3.8</td></tr><tr><td>7/8 riff</td><td>35 ms</td><td>7/8</td><td>2.9</td></tr><tr><td>6/8 ballad</td><td>12 ms</td><td>8/8</td><td>4.0</td></tr><tr><td>Key change</td><td>9 ms</td><td>4/8</td><td>3.1</td></tr></tbody></table></div>
<p><em>Mock interpretation:</em> key and tempo hold up, but odd meters and modulations are the weak spots. Shorter loop windows that sit inside one key would be the practical fix.</p>
<h2 id="notes">Notes</h2>
<p>Use only sessions Vocaler fully owns, because Bridge uploads the project to Suno.</p>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/lab-lab-suno-bridge-tempo-key-lock-test.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/lab-lab-suno-bridge-tempo-key-lock-test.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Spotify signs Believe and TuneCore for its AI covers and remix tool</title>
    <link>https://pulse.vocaler.ai/news/spotify-believe-tunecore-ai-remix-licensing/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/news/spotify-believe-tunecore-ai-remix-licensing/</guid>
    <pubDate>Tue, 20 Oct 2026 08:00:00 +0000</pubDate>
    <category>News</category><category>spotify</category><category>licensing</category><category>remixes</category><category>distribution</category>
    <description>Believe and TuneCore artists can opt in to Spotify&#x27;s planned AI covers and remixes add-on, with royalties flowing to the original rights holders.</description>
    <content:encoded><![CDATA[<h2 id="the-deal">The deal</h2>
<p>On October 20, 2026 Believe signed a licensing agreement with Spotify for Spotify's upcoming <strong>AI-powered covers and remixes tool</strong>. The deal also covers artists and labels on Believe's DIY distributor <strong>TuneCore</strong>.</p>
<p>Key points reported by Music Business Worldwide and others:</p>
<ul><li><strong>Remixes and covers only.</strong> The tool is "designed specifically for AI-powered remixes and covers, not entirely new works." Each remix links back to the original track and artist.</li><li><strong>Royalties to the original rights holders.</strong> Artists and rights holders behind the source tracks get paid when the remixes are streamed. Spotify presents this as income on top of existing royalties.</li><li><strong>A paid add-on.</strong> The feature is planned as an extra for Spotify Premium subscribers. No price or launch date has been announced.</li><li><strong>Two-step consent.</strong> Believe and TuneCore first ask artists whether they want to be considered for AI programmes at all. Those who opt in then get specific offers, such as this one, and can opt out of each individually.</li></ul>
<h2 id="where-it-fits">Where it fits</h2>
<p>This is Spotify's latest licence for the product. It follows <strong>Universal Music Group (May 2026)</strong>, <strong>Merlin (August 4, 2026)</strong> and <strong>Kobalt (August 13, 2026)</strong>. Sony Music and Warner Music, which agreed responsible-AI principles with Spotify in October 2025, have not yet announced their own licences. On Spotify's Q2 2026 earnings call, co-CEO Gustav Söderström said the next step would be a research preview for a subset of users.</p>
<h2 id="why-it-matters-for-creators">Why it matters for creators</h2>
<ul><li><strong>If you distribute through TuneCore or Believe, look for the opt-in.</strong> Choosing to take part decides whether fans can legally remix your catalogue inside Spotify, and whether you earn from it.</li><li><strong>It's a licensed alternative to unofficial AI covers.</strong> Remixes made in the tool credit and pay the original artist. That may reduce the demand for unlicensed AI covers that get taken down.</li><li><strong>It won't host your Suno originals.</strong> The tool is built around existing licensed tracks, not uploads of brand-new AI-generated songs. Those still go through distributors, with Spotify's voluntary AI credits disclosure.</li><li><strong>Independent artists aren't locked out.</strong> With Merlin, Kobalt and now Believe/TuneCore on board, independent catalogues have a route into the product, not just major-label artists.</li></ul>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/news-spotify-believe-tunecore-ai-remix-licensing.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/news-spotify-believe-tunecore-ai-remix-licensing.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Lab: Can ElevenLabs match a video&#x27;s temp track with use_audio_as_reference?</title>
    <link>https://pulse.vocaler.ai/lab/lab-elevenlabs-video-to-music-audio-reference/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/lab/lab-elevenlabs-video-to-music-audio-reference/</guid>
    <pubDate>Mon, 19 Oct 2026 08:00:00 +0000</pubDate>
    <category>Lab</category><category>elevenlabs</category><category>video</category><category>api</category><category>experiment</category>
    <description>A planned A/B test of ElevenLabs Video to Music with and without the audio-reference flag, checking BPM, genre and SFX timing. The results are mock.</description>
    <content:encoded><![CDATA[<h2 id="hypothesis">Hypothesis</h2>
<p>The <code>use_audio_as_reference</code> flag is meant to pull genre, BPM, instruments and mood from a video's own audio. With it on, we expect the generated score to land <strong>within ±5 BPM</strong> of the temp track and in the same genre far more often than with a text description alone.</p>
<h2 id="setup">Setup</h2>
<ul><li><strong>Clips:</strong> six 45-second edits, each with a royalty-free temp track Vocaler owns: lo-fi hip-hop (85 BPM), synthwave (110), orchestral trailer (70), drum &amp; bass (174), acoustic folk (100), afrobeats (105).</li><li><strong>Endpoint:</strong> <code>POST /v1/music/video-to-music</code>.</li><li><strong>Arm A:</strong> <code>description: &quot;score that matches the energy of the edit&quot;</code>, with <code>use_audio_as_reference: false</code>.</li><li><strong>Arm B:</strong> the same description, with <code>use_audio_as_reference: true</code>.</li><li><strong>Both arms:</strong> <code>add_sfx: true</code>, <code>separate_tracks: true</code>, and 3 generations per clip per arm (36 in total).</li><li><strong>Measures:</strong> BPM difference from the temp track, a blind genre-match vote from 3 listeners, SFX hits within ±100 ms of a visual event, and the number of SFX that had to be deleted.</li></ul>
<h2 id="mock-findings-illustrative-only-not-real-results">Mock findings (illustrative only, not real results)</h2>
<div class="tw"><table><thead><tr><th>Metric</th><th>Arm A (text only)</th><th>Arm B (audio reference)</th></tr></thead><tbody><tr><td>BPM within ±5</td><td>6/18</td><td>15/18</td></tr><tr><td>Genre match (majority vote)</td><td>9/18</td><td>16/18</td></tr><tr><td>SFX hits within ±100 ms</td><td>71%</td><td>73%</td></tr><tr><td>SFX deleted per clip (mean)</td><td>2.4</td><td>2.2</td></tr></tbody></table></div>
<p><em>Mock interpretation:</em> the reference flag mainly improves musical match. SFX timing doesn't change, which is expected, since effects are placed from the visuals.</p>
<h2 id="follow-ups">Follow-ups</h2>
<ul><li>Test drum &amp; bass on its own, because very high tempos may get halved.</li><li>Compare <code>separate_tracks</code> stems in a DAW mix against the combined output.</li></ul>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/lab-lab-elevenlabs-video-to-music-audio-reference.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/lab-lab-elevenlabs-video-to-music-audio-reference.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Scoring video with ElevenLabs Instant Soundtracks and the Video to Music API</title>
    <link>https://pulse.vocaler.ai/guides/elevenlabs-instant-soundtracks-video-to-music-guide/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/guides/elevenlabs-instant-soundtracks-video-to-music-guide/</guid>
    <pubDate>Thu, 15 Oct 2026 08:00:00 +0000</pubDate>
    <category>Guides</category><category>elevenlabs</category><category>video</category><category>sound effects</category><category>api</category>
    <description>Use ElevenLabs&#x27; Instant Soundtracks to score a video with music plus synced SFX on separate layers, in the app or via the API.</description>
    <content:encoded><![CDATA[<h2 id="what-it-does">What it does</h2>
<p><strong>Instant Soundtracks</strong> launched on October 15, 2026 in ElevenCreative's <strong>Music Video Editor</strong>, on the <strong>ElevenMusic</strong> platform and through the <strong>Video to Music API</strong>. You upload a video and in one pass it:</p>
<ul><li>writes an <strong>original score</strong> timed to the footage,</li><li>finds moments that need <strong>sound effects</strong> and generates them at the right timestamps,</li><li>puts the SFX on <strong>separate layers</strong>, so you can move, swap or delete them without regenerating the music.</li></ul>
<h2 id="in-app-workflow-elevencreative">In-app workflow (ElevenCreative)</h2>
<ol><li>Open the Music Video Editor and <strong>upload your cut</strong>. Lock the edit first, because the score follows the timing.</li><li>Add a short <strong>description</strong> of mood and genre, and optionally style tags.</li><li>If your video already has temp music or ambience, turn on <strong>use the existing audio as a reference</strong>. This makes the score follow its genre, BPM, instrumentation and mood.</li><li>Generate, then <strong>check the SFX layers one by one</strong>. Delete clutter and nudge hits onto the cut points.</li><li>Export the music and effects separately if you plan to mix in a DAW.</li></ol>
<h2 id="api-workflow">API workflow</h2>
<p><code>POST /v1/music/video-to-music</code> (documented in the October 19, 2026 changelog):</p>
<ul><li><strong>Inputs:</strong> up to <strong>10 video files, 200 MB combined</strong>.</li><li><strong><code>description</code></strong>: free-text steering.</li><li><strong><code>tags</code></strong>: up to 10 style tags.</li><li><strong><code>use_audio_as_reference</code></strong> (default <code>false</code>): pulls genre, BPM, instruments and mood from the video's own audio.</li><li><strong><code>add_sfx</code></strong> (default <code>false</code>): adds matching sound effects.</li><li><strong><code>separate_tracks</code></strong> (default <code>false</code>): returns music and SFX as separate files in a ZIP.</li><li><strong><code>enable_logging</code></strong> (default <code>true</code>): set it to <code>false</code> for Zero Retention Mode. This is available to enterprise customers only.</li></ul>
<p>Since October 12, responses also include a <strong><code>song-id</code> header</strong>, so you can track or re-edit the generated song later. API access requires a paid plan.</p>
<h2 id="licensing-check-before-you-publish">Licensing check before you publish</h2>
<p>ElevenLabs' Music Terms (updated October 9, 2026) say:</p>
<ul><li><strong>Free and Starter:</strong> streaming distribution of music is <strong>prohibited</strong>.</li><li><strong>Creator plan and above:</strong> streaming distribution is permitted.</li><li><strong>All self-serve plans:</strong> film, TV, radio and "Studio Games" need an <strong>Enterprise</strong> agreement. Indie games under $500,000 in lifetime revenue are excluded from that restriction.</li><li><strong>Prompts:</strong> you can't use artist names, song titles or distinctive lyrics.</li></ul>
<h2 id="tips">Tips</h2>
<ul><li>Score <strong>one scene per request</strong> if the mood changes sharply.</li><li>Ask for <code>separate_tracks</code> every time. It costs nothing extra in workflow terms and saves remixing later.</li></ul>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/guides-elevenlabs-instant-soundtracks-video-to-music-guide.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/guides-elevenlabs-instant-soundtracks-video-to-music-guide.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Which AI can write a melody over an instrumental I already have?</title>
    <link>https://pulse.vocaler.ai/answers/ai-melody-over-existing-instrumental/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/answers/ai-melody-over-existing-instrumental/</guid>
    <pubDate>Fri, 25 Sep 2026 08:00:00 +0000</pubDate>
    <category>Answers</category><category>suno</category><category>vocals</category><category>workflow</category>
    <description>A KVR producer wanted AI to add a fitting melody to a finished instrumental without regenerating the whole track. Here are today&#x27;s options.</description>
    <content:encoded><![CDATA[<h2 id="answer">Answer</h2>
<p>You can now add a melody without giving up your mix. The key is to treat the AI output as a <strong>separate stem</strong>, not a new master.</p>
<ul><li><strong>Suno Bridge (beta, from November 13, 2026).</strong> If you work in Ableton Live 12 on a Mac with Suno Premier, Bridge reads your session's tempo, key and harmony. It then writes new vocal or instrument stems onto your timeline, and your original audio stays untouched. Loop markers set the bars it writes for, and each generation costs about 10 credits.</li><li><strong>Suno Add Vocals (Pro and Premier).</strong> Upload the instrumental, give it lyrics, and use the Audio Strength slider to control how closely the result sticks to your track. Suno still renders a new full mix, so split the stems and keep only the vocal, laid over your original file.</li><li><strong>Other options mentioned in the thread:</strong> Tunee, Veena Studio, EasyMusic.AI's upload-and-add-vocals workflow, and local ComfyUI setups for people who don't want to upload anything.</li></ul>
<p>Whichever you choose, check the key and tempo first and work in 8 to 16 bar sections.</p>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/answers-ai-melody-over-existing-instrumental.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/answers-ai-melody-over-existing-instrumental.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Suno v6 vs ElevenLabs Music v2.5 vs Google Lyria 3.5</title>
    <link>https://pulse.vocaler.ai/compare/suno-v6-vs-elevenlabs-music-v2-5-vs-lyria-3-5/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/compare/suno-v6-vs-elevenlabs-music-v2-5-vs-lyria-3-5/</guid>
    <pubDate>Fri, 11 Sep 2026 08:00:00 +0000</pubDate>
    <category>Compare</category><category>suno</category><category>elevenlabs</category><category>lyria</category><category>comparison</category>
    <description>Three flagship song models compared on length, editing, access, price and licensing, based on official docs and launch coverage.</description>
    <content:encoded><![CDATA[<h2 id="the-contenders">The contenders</h2>
<ul><li><strong>Suno v6</strong> (September 9, 2026): a family of three, v6, v6-wild and v6-mini.</li><li><strong>ElevenLabs Music v2.5</strong> (reported September 11, 2026): which ElevenLabs calls "our most advanced music model."</li><li><strong>Google Lyria 3.5</strong> (Flow Music July 29, 2026; Gemini app and API September 4, 2026): Google's flagship song model.</li></ul>
<h2 id="spec-sheet">Spec sheet</h2>
<div class="tw"><table><thead><tr><th>Criterion</th><th>Suno v6</th><th>ElevenLabs Music v2.5</th><th>Google Lyria 3.5</th></tr></thead><tbody><tr><td>Max song length</td><td>Up to 8 min</td><td>Reported up to 5–10 min (varies by access route; unconfirmed in docs)</td><td>30 s to ~3 min</td></tr><tr><td>Vocals</td><td>Yes</td><td>Yes, multilingual</td><td>Yes</td></tr><tr><td>Inputs</td><td>Text, lyrics, audio uploads</td><td>Text, composition plans, audio reference</td><td>Text and images</td></tr><tr><td>Section editing</td><td>Extend, Cover, Replace Section</td><td>Edit sections, inpainting</td><td>Unknown</td></tr><tr><td>Stems</td><td>Up to 12 in Studio (Premier)</td><td>2/4/6-stem separation (reported)</td><td>Vocals, drums and bass adjustable (Flow Music)</td></tr><tr><td>Official API</td><td>Suno MCP beta; no public music API documented</td><td>Yes (Music API)</td><td>Yes (Gemini API, preview)</td></tr><tr><td>Entry price</td><td>v6-mini free; v6 needs Pro/Premier ($30/mo Premier)</td><td>Free tier; streaming rights from Creator plan</td><td>About $0.08 per song via API; included in consumer apps</td></tr><tr><td>Ownership / rights</td><td>Paid plans own outputs; free plan is non-commercial</td><td>Streaming allowed from Creator plan; film/TV/radio need Enterprise</td><td>Unknown for distribution; see Google terms</td></tr><tr><td>Watermark</td><td>Not documented for audio</td><td>Not documented</td><td>SynthID (Lyria family)</td></tr></tbody></table></div>
<h2 id="how-they-differ-in-practice">How they differ in practice</h2>
<ul><li><strong>Suno v6</strong> is the most complete <em>songwriting environment</em>: long songs, three variants for precision versus exploration, Custom Models, and a DAW-style Studio. The trade-off is that its rights picture is shaped by ongoing lawsuits.</li><li><strong>ElevenLabs Music v2.5</strong> leans toward <em>licensed, controllable</em> output for media. It offers composition plans, inpainting and clear commercial tiers. ElevenLabs has a multi-year licensing agreement with Universal Music Group. Restrictions for broadcast and studio-game use are spelled out.</li><li><strong>Lyria 3.5</strong> is the <em>developer-friendly, per-song-priced</em> option. It produces 44.1 kHz stereo, accepts image prompts and plugs straight into Gemini. It is shorter, and Google documents less about editing and rights for distribution.</li></ul>
<h2 id="picking-one">Picking one</h2>
<ul><li><strong>Album-length songwriting and arranging:</strong> Suno v6.</li><li><strong>Background music for client video, games or ads, with clear licences:</strong> ElevenLabs Music.</li><li><strong>Building an app, or generating at scale through an API:</strong> Lyria 3.5.</li></ul>
<p><em>Unknowns are marked in the table. Test vocals in your own genre before you commit.</em></p>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/compare-suno-v6-vs-elevenlabs-music-v2-5-vs-lyria-3-5.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/compare-suno-v6-vs-elevenlabs-music-v2-5-vs-lyria-3-5.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Prompting Suno v6: choosing v6, v6-wild or v6-mini and taming Variety</title>
    <link>https://pulse.vocaler.ai/guides/suno-v6-prompting-model-variants-variety/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/guides/suno-v6-prompting-model-variants-variety/</guid>
    <pubDate>Wed, 09 Sep 2026 08:00:00 +0000</pubDate>
    <category>Guides</category><category>suno v6</category><category>prompting</category><category>style tags</category>
    <description>How to pick the right Suno v6 variant, why your style prompt gets rewritten, and a repeatable brief that keeps v6 on target.</description>
    <content:encoded><![CDATA[<h2 id="the-v6-lineup">The v6 lineup</h2>
<p>Suno released the <strong>v6 model family on September 9, 2026</strong>. All earlier models were retired at that point, though songs you already made stay in your library. Suno's help center describes three variants:</p>
<div class="tw"><table><thead><tr><th>Variant</th><th>Best for</th><th>Plans</th></tr></thead><tbody><tr><td><strong>v6</strong></td><td>Balanced, expressive everyday creation</td><td>Pro, Premier</td></tr><tr><td><strong>v6-wild</strong></td><td>Unexpected choices and genre-blending; results drift further from the prompt</td><td>Pro, Premier</td></tr><tr><td><strong>v6-mini</strong></td><td>Fast, high-volume idea generation</td><td>All users</td></tr></tbody></table></div>
<p>All three support <strong>up to 8 minutes per generation</strong>. You pick the model from the dropdown at the top right of the Create form. v6 also powers Custom Models, Suno Sounds and the Remix and Edit tools (Extend, Cover, Replace Section).</p>
<h2 id="rule-of-thumb">Rule of thumb</h2>
<p><strong>If you know the sound you want, use v6. If you're still searching for it, use v6-wild.</strong> Treat v6-wild as a scouting tool. Keep the happy accidents, then rewrite the idea as a precise v6 brief.</p>
<h2 id="why-your-prompt-keeps-changing-variety">Why your prompt keeps changing: Variety</h2>
<p>Under <strong>More Options</strong> there is a <strong>Variety</strong> control. Suno's FAQ says it introduces variety "by adjusting and updating your style prompts." Setting it to <strong>Off/0</strong> lets you "retain full control of your style tags." In HookGenius's test of 54 takes, the Normal setting rewrote the style prompt on nearly every take, while Off kept the exact text in 52 of 54.</p>
<p>Other More Options controls to check:</p>
<ul><li><strong>Style Influence</strong> (default 50%): how much weight your style prompt carries. Raise it if your prompt seems to be ignored.</li><li><strong>Weirdness</strong>, <strong>Max Mode</strong>, <strong>vocal gender</strong> and <strong>duration</strong>.</li><li><strong>Exclude Styles</strong>: put things you don't want here, rather than writing "no drums" in the style box.</li></ul>
<h2 id="a-repeatable-v6-brief">A repeatable v6 brief</h2>
<p>Field limits are unchanged from v5.5: <strong>1,000 characters for Styles</strong> and <strong>5,000 for Lyrics</strong>.</p>
<div class="code"><button class="cp" type="button" data-copy>Copy</button><pre><code>Styles: 2020s alt-R&amp;B, 88 BPM, minor key, warm Rhodes, sub bass,
crisp trap hats, intimate breathy male vocal, close-mic, sparse verses,
wide stacked chorus harmonies
Exclude: EDM drops, autotune</code></pre></div>
<p>Use section metatags in the lyrics (for example <code>[Verse]</code>, <code>[Chorus]</code>, <code>[Bridge]</code>). Third-party testers report that v5.5-era metatag syntax still works in v6.</p>
<h2 id="workflow">Workflow</h2>
<ol><li>Turn <strong>Variety off</strong> and lock your brief.</li><li>Generate a few takes on <strong>v6</strong>.</li><li>If they feel generic, run the same brief on <strong>v6-wild</strong>, then copy what worked back into the v6 brief.</li><li>Use <strong>v6-mini</strong> for cheap lyric and structure drafts before you spend credits on v6.</li></ol>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/guides-suno-v6-prompting-model-variants-variety.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/guides-suno-v6-prompting-model-variants-variety.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Lab: Does v6-wild really drift? A genre-adherence test across Suno v6 models</title>
    <link>https://pulse.vocaler.ai/lab/lab-suno-v6-vs-wild-vs-mini-genre-adherence/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/lab/lab-suno-v6-vs-wild-vs-mini-genre-adherence/</guid>
    <pubDate>Wed, 09 Sep 2026 08:00:00 +0000</pubDate>
    <category>Lab</category><category>suno v6</category><category>experiment</category><category>prompting</category>
    <description>We plan to run one fixed brief through v6, v6-wild and v6-mini with Variety off to measure genre and tempo adherence. The findings below are mock.</description>
    <content:encoded><![CDATA[<h2 id="hypothesis">Hypothesis</h2>
<p>Suno describes <strong>v6-wild</strong> as pushing "further from the prompt." With <strong>Variety set to Off</strong>, we expect v6 to hit the requested genre and BPM most often, v6-wild least often, and v6-mini in between with thinner arrangements.</p>
<h2 id="setup">Setup</h2>
<ul><li><strong>Models:</strong> v6, v6-wild, v6-mini. <strong>10 generations each</strong> (30 total).</li><li><strong>Controls:</strong> Variety Off, Style Influence 50% (default), Weirdness default, instrumental off, same lyrics.</li><li><strong>Styles prompt:</strong></li></ul>
<div class="code"><button class="cp" type="button" data-copy>Copy</button><pre><code>1970s Ethiopian jazz, 96 BPM, minor pentatonic horn riffs, vibraphone,
fuzzy organ, tight funk drums, male baritone vocal in English, warm tape saturation
Exclude: EDM, trap hi-hats, autotune</code></pre></div>
<ul><li><strong>Lyrics:</strong> 2 verses and a chorus with <code>[Verse]</code> / <code>[Chorus]</code> tags.</li><li><strong>Scoring:</strong> BPM measured with a tempo detector (±3 BPM counts as a hit), blind genre-match votes from 3 listeners (1–5 scale), and whether any excluded element appears.</li></ul>
<h2 id="mock-findings-illustrative-only-not-real-results">Mock findings (illustrative only, not real results)</h2>
<div class="tw"><table><thead><tr><th>Metric</th><th>v6</th><th>v6-wild</th><th>v6-mini</th></tr></thead><tbody><tr><td>BPM within ±3</td><td>8/10</td><td>5/10</td><td>7/10</td></tr><tr><td>Mean genre-match score</td><td>4.1</td><td>3.2</td><td>3.6</td></tr><tr><td>Excluded element present</td><td>1/10</td><td>3/10</td><td>2/10</td></tr><tr><td>"Most interesting" listener pick</td><td>3</td><td>6</td><td>1</td></tr></tbody></table></div>
<p><em>Mock interpretation:</em> v6-wild loses adherence but wins on surprise, which supports the "scout with wild, finish with v6" workflow.</p>
<h2 id="follow-ups">Follow-ups</h2>
<ul><li>Repeat with Variety at Normal to measure how much prompt rewriting costs in adherence.</li><li>Test whether raising Style Influence closes v6-wild's gap.</li></ul>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/lab-lab-suno-v6-vs-wild-vs-mini-genre-adherence.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/lab-lab-suno-v6-vs-wild-vs-mini-genre-adherence.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Why do my recombined Suno stems sound flatter than the full song?</title>
    <link>https://pulse.vocaler.ai/answers/recreate-suno-mastering-on-stems/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/answers/recreate-suno-mastering-on-stems/</guid>
    <pubDate>Thu, 23 Jul 2026 08:00:00 +0000</pubDate>
    <category>Answers</category><category>suno</category><category>stems</category><category>mixing</category><category>mastering</category>
    <description>A Gearspace user found that re-summed Suno stems lack the polish of Suno&#x27;s stereo mix. Here&#x27;s why, and how to close the gap.</description>
    <content:encoded><![CDATA[<h2 id="answer">Answer</h2>
<p>Suno generates the <strong>finished stereo song first</strong>. Stems are then <strong>separated out of that mix</strong>, so they don't sum back to the polished master, and each one can carry separation artifacts. You're not missing a hidden mastering chain you can switch back on. You are rebuilding the bus processing yourself.</p>
<p>A practical approach:</p>
<ol><li><strong>Use the stereo render as a reference.</strong> Put Suno's full mix on a muted track and level-match it to your stem mix (LUFS meter).</li><li><strong>Glue first, then loudness.</strong> Add gentle mix-bus compression (around 2:1, slow attack, 1–2 dB of gain reduction), then broad EQ moves to match the reference's tonal balance.</li><li><strong>Try reference-matching tools.</strong> Match EQ or reference-based mastering plugins can get you close fast.</li><li><strong>Clean the stems.</strong> Gate or EQ out bleed between stems before the bus. Artifacts get louder once you compress.</li><li><strong>Finish with a limiter</strong> at your delivery target.</li></ol>
<p>If you're on Premier, Suno Studio exports up to 12 WAV stems and includes a six-band EQ. Still, finish the master in your own DAW so you control the result.</p>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/answers-recreate-suno-mastering-on-stems.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/answers-recreate-suno-mastering-on-stems.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Spotify vs Deezer vs YouTube: how each treats AI music in 2026</title>
    <link>https://pulse.vocaler.ai/compare/spotify-vs-deezer-vs-youtube-ai-music-policy-2026/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/compare/spotify-vs-deezer-vs-youtube-ai-music-policy-2026/</guid>
    <pubDate>Tue, 21 Jul 2026 08:00:00 +0000</pubDate>
    <category>Compare</category><category>spotify</category><category>deezer</category><category>youtube</category><category>distribution</category>
    <description>Disclosure, detection, labels and monetization: a side-by-side of the three platforms AI music creators release to most.</description>
    <content:encoded><![CDATA[<h2 id="three-philosophies">Three philosophies</h2>
<ul><li><strong>Spotify: disclose voluntarily, punish spam.</strong> AI-assisted music is allowed and not down-ranked. Impersonation and content farms are banned.</li><li><strong>Deezer: detect and contain.</strong> Deezer runs its own AI detector, tags AI tracks, and keeps them out of recommendations.</li><li><strong>YouTube: label it, and don't mass-produce it.</strong> Disclosure is required for realistic synthetic media, and monetization depends on the "inauthentic content" rules.</li></ul>
<h2 id="side-by-side">Side by side</h2>
<div class="tw"><table><thead><tr><th>Criterion</th><th>Spotify</th><th>Deezer</th><th>YouTube</th></tr></thead><tbody><tr><td>AI music allowed?</td><td>Yes</td><td>Yes</td><td>Yes</td></tr><tr><td>Disclosure mechanism</td><td>Voluntary AI credits via DDEX metadata (beta since April 16, 2026, DistroKid first)</td><td>Platform detection; no creator form needed</td><td>"Altered or synthetic content" label in YouTube Studio</td></tr><tr><td>Platform detection</td><td>No platform-wide AI detector announced; spam filter for mass uploads</td><td>Yes, in-house tool (claims 99.8% accuracy)</td><td>Unknown for music specifically</td></tr><tr><td>Listener-facing label</td><td>AI credit in Song Credits (if disclosed)</td><td>AI tag on content</td><td>Label overlay on Shorts / below player</td></tr><tr><td>Recommendations</td><td>Not penalised for AI use, per Spotify</td><td>Fully AI tracks removed from algorithmic and editorial playlists</td><td>Label alone doesn't affect reach, per policy explainers</td></tr><tr><td>Monetization risk</td><td>Spam, duplicates, ultra-short tracks; impersonation removed</td><td>Fraudulent streams demonetised; zero-stream AI tracks after 6+ months to be removed</td><td>"Inauthentic content" (mass-produced) can lose YPP monetization</td></tr><tr><td>Voice clones</td><td>Unauthorised clones banned</td><td>Fraud-focused</td><td>Likeness/privacy complaints apply</td></tr></tbody></table></div>
<h2 id="numbers-behind-the-policies">Numbers behind the policies</h2>
<p>Deezer says that by June 2026 it was receiving about <strong>90,000 fully AI-generated tracks a day</strong>, more than <strong>50% of daily uploads</strong>. Those tracks still made up only <strong>1–3% of streams</strong>, and up to <strong>85%</strong> of streams on fully AI tracks were found to be fraudulent in 2025.</p>
<h2 id="beyond-streaming-charts">Beyond streaming: charts</h2>
<p>Charts are tightening too. From August 29, 2026, Australia's <strong>ARIA Charts</strong> exclude wholly AI-generated recordings. Tracks must be "substantially human made," and AI in a supporting role is still allowed.</p>
<h2 id="takeaways-for-creators">Takeaways for creators</h2>
<ol><li><strong>Disclose honestly</strong> in your distributor's AI fields. On Spotify and YouTube, disclosing doesn't by itself hurt reach.</li><li><strong>Don't count on Deezer's algorithm</strong> for fully AI tracks. Promote them directly instead.</li><li><strong>Add real human input</strong> (lyrics, performance, arrangement, edits) and document it. It matters for chart eligibility and for YouTube's inauthentic content rules.</li><li><strong>Release fewer, better tracks.</strong> Every platform's enforcement targets volume uploading.</li></ol>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/compare-spotify-vs-deezer-vs-youtube-ai-music-policy-2026.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/compare-spotify-vs-deezer-vs-youtube-ai-music-policy-2026.png" type="image/png" length="0"/>
  </item>
  <item>
    <title>Can Suno give me clean parts a live band can actually read?</title>
    <link>https://pulse.vocaler.ai/answers/suno-transcribable-arrangements-live-band/</link>
    <guid isPermaLink="true">https://pulse.vocaler.ai/answers/suno-transcribable-arrangements-live-band/</guid>
    <pubDate>Sun, 12 Apr 2026 08:00:00 +0000</pubDate>
    <category>Answers</category><category>suno studio</category><category>midi</category><category>stems</category><category>live performance</category>
    <description>A Gearspace user wanted Suno arrangements a band could transcribe. Studio&#x27;s stems and MIDI export now make that much more realistic.</description>
    <content:encoded><![CDATA[<h2 id="answer">Answer</h2>
<p>Yes, with some cleanup. The tooling has improved a lot since this was asked. In that thread, the poster found Suno's built-in stem extractor beat RipX, LALAL.AI and UVR, and used Melodyne for audio-to-MIDI.</p>
<p>Current workflow:</p>
<ol><li><strong>Write a sparse brief.</strong> Name a small band lineup (for example <em>"drums, bass, two guitars, keys"</em>) and ask for clear parts. Dense, heavily produced mixes transcribe badly.</li><li><strong>Split in Suno Studio (Premier).</strong> Studio can separate a track into up to 12 stems, exported as WAV.</li><li><strong>Get MIDI per stem.</strong> In the Stems panel, select a melodic stem and choose <strong>Get MIDI</strong>, which costs 10 credits per stem. Vocals are not offered as MIDI. Suno's September 17, 2026 update made MIDI transcription faster and more accurate.</li><li><strong>Import into notation software</strong> (MuseScore, Dorico, Sibelius) and quantize. Then fix voicings so they suit each player's instrument range.</li><li><strong>Chart the vocal melody by ear,</strong> or run the vocal stem through a pitch-to-MIDI tool.</li></ol>
<p>Expect to edit drums and comping parts by hand. AI MIDI is a draft, not a finished chart.</p>]]></content:encoded>
    <media:content url="https://pulse.vocaler.ai/og/answers-suno-transcribable-arrangements-live-band.png" medium="image" type="image/png" width="1200" height="630"/>
    <enclosure url="https://pulse.vocaler.ai/og/answers-suno-transcribable-arrangements-live-band.png" type="image/png" length="0"/>
  </item>
</channel>
</rss>
