{
  "model": "gemini-2.5-pro",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "You are an independent formal composition listener. Receive two entire story songs as two continuous100s-or-shorter parts each, actual audio in black video containers. The English song X and Ukrainian song Y tell adapted versions of the same story; no preferred ranking is supplied. Analyze musical form and the relationship between prosody and narrative information, NOT a correctness/acceptance score. Need: whether melody follows speech stress or smooth recurring song cells; what rhythmic/melodic cell actually returns (if not recognizable say unknown); if musical texture/register/phrase release articulates change of speaker or change of emotional state; what becomes the perceived climax and what the commercial ending does to it. Contrast especially rhetorical repetition, late self-worth and call to action. Audio structure must come from hearing the ENTIRE tracks; timestamp estimates are tentative. Return JSON in English <=1300words with mediaAccess, audioAccess, completeCoverageOfBothTracks, limitations, trackXForm, trackYForm, melodyVsSpeechComparison, actualMotifReturnsEach, musicalClimaxEach, commercialEndingComparison, strongestCounterexamplesToAnyClaimOfOriginalSuperiority, concretePriorities[{issue,sourceHeardEvidence,newHeardEvidence,approxSourceRange,approxNewRange,whyItMatters,confidence}], conclusionWithoutNumericScores. Include short actual heard lyric clues, not a fabricated transcript. No exact key/BPM/note sequence or psycho/neuro/sales claims. Distinguish stable accompaniment under varied voice from fully constant music. Do not equate conversational singing with ordinary speech, repeated melody with failure or smooth singing with excellence."
        },
        {
          "type": "text",
          "text": "Attachment 1: TrackX (targ),continuous actual soundtrack global0..100s. Local0=global0. Black video is transport only, no picture analysis. This part plus adjacent part covers the whole song."
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "[inline bytes; see inputs.json]"
          }
        },
        {
          "type": "text",
          "text": "Attachment 2: TrackX (targ),continuous actual soundtrack global100..194.489s. Local0=global100. Black video is transport only, no picture analysis. This part plus adjacent part covers the whole song."
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "[inline bytes; see inputs.json]"
          }
        },
        {
          "type": "text",
          "text": "Attachment 3: TrackY (gopure),continuous actual soundtrack global0..100s. Local0=global0. Black video is transport only, no picture analysis. This part plus adjacent part covers the whole song."
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "[inline bytes; see inputs.json]"
          }
        },
        {
          "type": "text",
          "text": "Attachment 4: TrackY (gopure),continuous actual soundtrack global100..197.726s. Local0=global100. Black video is transport only, no picture analysis. This part plus adjacent part covers the whole song."
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "[inline bytes; see inputs.json]"
          }
        }
      ]
    }
  ],
  "stream": false,
  "include_thoughts": false,
  "reasoning_effort": "high"
}