Spectral Forensics Benchmark for AI Music Generators

Spectral Forensics analysis compares the Lyria 3 and Suno v4.5 AI music generation models. The dataset contains spectral features extracted from audio outputs for benchmarking purposes. Author and creation date are unknown.

Use Cases

• Compare spectral features like harmonic content or noise profiles between Lyria 3 and Suno v4.5 outputs.
• Train classifiers to identify the source AI model from spectral analysis data.
• Benchmark audio quality and artifacts in AI-generated music using forensic metrics.
• Analyze the evolution of AI music generation by comparing spectral signatures across model versions.

Strengths

• Direct comparison of two prominent AI music models, Lyria 3 and Suno v4.5.
• Focus on spectral forensics, a specific technical evaluation method for audio.

Limitations

• Sample size, row count, and specific features are unknown.
• Lacks metadata such as license, author, and update history.

Provenance

Collection Method

Likely involves generating music with the specified AI models and performing spectral analysis.