The system extracts technical metadata from the uploaded file and identifies the most relevant stream-level properties.

Then it analyzes the signal itself to generate waveform, spectrogram, silence detection, and loudness-related statistics.

If enabled, the system also produces a transcription and compares all signals to generate a final adjudicated verdict.

The outcome is analytical and evidence-based, but it should still be treated as a technical assessment rather than absolute forensic proof.