REAL AI-SONG EXAMPLES
AI songs.
Before and after.
Three openly licensed songs, processed in full with our browser mastering engine. Hear a Mureka instrumental, Suno vocals and AI-generated jazz, including a lighter Natural setting.
Mureka · Synthwave · Instrumental
Push to Origin
Universal changes integrated loudness from −14.3 to −14.0 LUFS, a difference of just 0.3 LU. The measured loudness range falls from 4.5 to 3.7 LU. These are changes to compare, not a quality score.
Same 1:00–1:30 excerpt. Each MP3 preview measures −18.0 LUFS after fixed-gain level matching. This preview adjustment is separate from the full-song results below.
Full-song measurements
| Measurement | Original | Universal · 50% |
|---|---|---|
| Integrated loudness | −14.3 LUFS | −14.0 LUFS |
| Loudness range | 4.5 LU | 3.7 LU |
| True peak | −0.3 dBTP | −1.4 dBTP |
Settings: Universal · 50%. EQ 0 / 0 / 0 dB, automatic input gain, no reference track.
Source, license & changes
“Push to Origin” by Neon Husky, from Contribution Run. Original source · Creator’s licensing statement · CC BY 4.0.
The source identifies this music as AI-generated with Mureka. The processed versions were mastered by AI Music Mastering. All previews, including “Original,” have been excerpted, attenuated and transcoded to 192 kbps MP3. The creators have not endorsed this service.
Suno V5.5 · Alternative R&B · AI vocals
Alt R&B
Universal changes integrated loudness from −14.3 to −14.0 LUFS, a difference of just 0.3 LU. The measured loudness range falls from 7.8 to 5.7 LU. These are changes to compare, not a quality score.
Same 1:00–1:30 excerpt. Each MP3 preview measures −18.0 LUFS after fixed-gain level matching. This preview adjustment is separate from the full-song results below.
Full-song measurements
| Measurement | Original | Universal · 50% |
|---|---|---|
| Integrated loudness | −14.3 LUFS | −14.0 LUFS |
| Loudness range | 7.8 LU | 5.7 LU |
| True peak | −0.5 dBTP | −1.5 dBTP |
Settings: Universal · 50%. EQ 0 / 0 / 0 dB, automatic input gain, no reference track.
Source, license & changes
“DS_alt_rnb_modern_x8pii” by Kukedlc, from Suno AI Music Dataset (Multi-Genre Curated). Original source · Creator’s licensing statement · CC BY 4.0.
The source identifies this music as AI-generated with Suno V5.5. The vocals are AI-generated, not a human performance. The processed versions were mastered by AI Music Mastering. All previews, including “Original,” have been excerpted, attenuated and transcoded to 192 kbps MP3. The creators have not endorsed this service.
Suno V5.5 · Modal jazz · AI-generated instruments
Modal jazz
The original already measures −12.4 LUFS. Natural at 30% brings it to −15.8 LUFS, with an LRA of 5.8 LU versus 5.2 LU for Universal at 50%. This is a lighter option with no loudness push. The original still has the widest measured loudness range; keep it if you prefer it.
Same 1:00–1:30 excerpt. Each MP3 preview measures −18.0 LUFS after fixed-gain level matching. This preview adjustment is separate from the full-song results below.
Full-song measurements
| Measurement | Original | Natural · 30% | Universal · 50% |
|---|---|---|---|
| Integrated loudness | −12.4 LUFS | −15.8 LUFS | −14.0 LUFS |
| Loudness range | 6.3 LU | 5.8 LU | 5.2 LU |
| True peak | −0.3 dBTP | −2.7 dBTP | −1.5 dBTP |
Settings: Natural · 30%; Universal · 50%. EQ 0 / 0 / 0 dB, automatic input gain, no reference track. Natural uses gentler base compression. Reducing Intensity to 30% then adjusts tone and output gain; it does not undo that compression.
Source, license & changes
“DS_jazz_spiritual_modal_zhkjv” by Kukedlc, from Suno AI Music Dataset (Multi-Genre Curated). Original source · Creator’s licensing statement · CC BY 4.0.
The source identifies this music as AI-generated with Suno V5.5. The instrumental sound is AI-generated, not an acoustic recording session. The processed versions were mastered by AI Music Mastering. All previews, including “Original,” have been excerpted, attenuated and transcoded to 192 kbps MP3. The creators have not endorsed this service.
HOW THESE EXAMPLES WERE MADE
One processing chain. Traceable results.
- Source files came from the linked creators’ pages under CC BY 4.0. The creators identify the generation tools; these are demonstration materials, not customer testimonials.
- Each complete song was rendered on October 28, 2026 with the same browser engine and live-control export graph used by this site. The engine uses preset-based DSP, not a trained AI model. Settings are listed on each case.
- FFmpeg 7.1’s EBU R128 filter measured the full-song decoded original and final 24-bit WAV. These are independent file measurements, including the 30% Intensity edit, not the Studio’s base-render report. True-peak methods may differ; the figures are not a delivery certification.
- Only after processing did we take the same 1:00–1:30 interval from every version, apply constant attenuation and encode all previews at 192 kbps MP3. We measured the encoded clips again to verify matching. No extra compression or limiting was added to the previews.
Measurements describe level and dynamics; they do not establish that a version sounds better. No listener score or invented listening review is used here. Mastering cannot reliably repair generated-vocal distortion, replace instruments or restore details lost in an MP3 source.
View settings, measurements, source links and file hashes (JSON)
Try the same comparison with your song.
Import and preview for free. Mastered downloads and reference matching require membership. In Studio, Original and Mastered play at their own levels; the matched excerpts on this page are prepared separately.
Try mastering freeMeasured October 28, 2026 · Published by Zitchoo LLC
