Find where the sound suddenly gets loud
0:18CC0 · B.G. DeSylva Productions Inc
Proof frames
Tool calls
loudness_timeline
Arguments
{
"source": "$CLIP",
"start_sec": 0,
"end_sec": 4,
"window_sec": 2
}
Result
{
"ok": true,
"data": {
"source_path": "~/hill/gallery/tmp/R-loud/cards/loud-music/clip.mp4",
"start_sec": 0,
"end_sec": 4,
"window_sec": 2,
"bpm": null,
"silence_threshold_dbfs": -60,
"peak_dbfs": -13.017,
"peak_sec": 0.563,
"windows": [
{
"start_sec": 0,
"end_sec": 2,
"rms_dbfs": -18.83,
"peak_dbfs": -13.017,
"band_db": {
"sub": -18.563,
"low_mid": -21.567,
"mid": -36.92,
"high": -58.529,
"air": -70.576
},
"spectral_centroid_hz": 137.941,
"silent": false
},
{
"start_sec": 2,
"end_sec": 4,
"rms_dbfs": -20.092,
"peak_dbfs": -13.272,
"band_db": {
"sub": -24.459,
"low_mid": -23.563,
"mid": -26.859,
"high": -54.381,
"air": -70.567
},
"spectral_centroid_hz": 373.836,
"silent": false
}
],
"outliers": [
{
"start_sec": 0,
"end_sec": 2,
"band": "mid",
"direction": "cut",
"energy_db": -36.92,
"baseline_db": -26.859,
"delta_db": -10.061,
"min_delta_db": -10.061,
"max_delta_db": -10.061,
"energy_start_db": -36.92,
"energy_end_db": -36.92,
"entry_jump_db": null,
"exit_jump_db": -10.061,
"internal_range_db": 0,
"trend_db": 0,
"monotonic_fraction": 0,
"shape": "transient",
"window_count": 1,
"fault_level": false,
"classification": "informational",
"ignored": false,
"ignore_reason": null
},
{
"start_sec": 2,
"end_sec": 4,
"band": "mid",
"direction": "boost",
"energy_db": -26.859,
"baseline_db": -36.92,
"delta_db": 10.061,
"min_delta_db": 10.061,
"max_delta_db": 10.061,
"energy_start_db": -26.859,
"energy_end_db": -26.859,
"entry_jump_db": 10.061,
"exit_jump_db": null,
"internal_range_db": 0,
"trend_db": 0,
"monotonic_fraction": 0,
"shape": "transient",
"window_count": 1,
"fault_level": false,
"classification": "informational",
"ignored": false,
"ignore_reason": null
}
],
"ignore_ranges": [],
"silence_gaps": [],
"total": 2
}
}
loudness_timeline
Arguments
{
"source": "$CLIP",
"start_sec": 4,
"end_sec": 8,
"window_sec": 2
}
Result
{
"ok": true,
"data": {
"source_path": "~/hill/gallery/tmp/R-loud/cards/loud-music/clip.mp4",
"start_sec": 4,
"end_sec": 8,
"window_sec": 2,
"bpm": null,
"silence_threshold_dbfs": -60,
"peak_dbfs": -1.38,
"peak_sec": 6.077,
"windows": [
{
"start_sec": 4,
"end_sec": 6,
"rms_dbfs": -21.115,
"peak_dbfs": -13.329,
"band_db": {
"sub": -27.117,
"low_mid": -22.274,
"mid": -25.859,
"high": -43.433,
"air": -66.997
},
"spectral_centroid_hz": 504.579,
"silent": false
},
{
"start_sec": 6,
"end_sec": 8,
"rms_dbfs": -8.766,
"peak_dbfs": -1.38,
"band_db": {
"sub": -12.43,
"low_mid": -9.861,
"mid": -16.007,
"high": -38.649,
"air": -56.531
},
"spectral_centroid_hz": 314.524,
"silent": false
}
],
"outliers": [
{
"start_sec": 4,
"end_sec": 6,
"band": "air",
"direction": "cut",
"energy_db": -66.997,
"baseline_db": -56.531,
"delta_db": -10.466,
"min_delta_db": -10.466,
"max_delta_db": -10.466,
"energy_start_db": -66.997,
"energy_end_db": -66.997,
"entry_jump_db": null,
"exit_jump_db": -10.466,
"internal_range_db": 0,
"trend_db": 0,
"monotonic_fraction": 0,
"shape": "transient",
"window_count": 1,
"fault_level": false,
"classification": "informational",
"ignored": false,
"ignore_reason": null
},
{
"start_sec": 4,
"end_sec": 6,
"band": "low_mid",
"direction": "cut",
"energy_db": -22.274,
"baseline_db": -9.861,
"delta_db": -12.413,
"min_delta_db": -12.413,
"max_delta_db": -12.413,
"energy_start_db": -22.274,
"energy_end_db": -22.274,
"entry_jump_db": null,
"exit_jump_db": -12.413,
"internal_range_db": 0,
"trend_db": 0,
"monotonic_fraction": 0,
"shape": "transient",
"window_count": 1,
"fault_level": false,
"classification": "informational",
"ignored": false,
"ignore_reason": null
},
{
"start_sec": 4,
"end_sec": 6,
"band": "mid",
"direction": "cut",
"energy_db": -25.859,
"baseline_db": -16.007,
"delta_db": -9.852,
"min_delta_db": -9.852,
"max_delta_db": -9.852,
"energy_start_db": -25.859,
"energy_end_db": -25.859,
"entry_jump_db": null,
"exit_jump_db": -9.852,
"internal_range_db": 0,
"trend_db": 0,
"monotonic_fraction": 0,
"shape": "transient",
"window_count": 1,
"fault_level": false,
"classification": "informational",
"ignored": false,
"ignore_reason": null
},
{
"start_sec": 4,
"end_sec": 6,
"band": "sub",
"direction": "cut",
"energy_db": -27.117,
"baseline_db": -12.43,
"delta_db": -14.687,
"min_delta_db": -14.687,
"max_delta_db": -14.687,
"energy_start_db": -27.117,
"energy_end_db": -27.117,
"entry_jump_db": null,
"exit_jump_db": -14.687,
"internal_range_db": 0,
"trend_db": 0,
"monotonic_fraction": 0,
"shape": "transient",
"window_count": 1,
"fault_level": false,
"classification": "informational",
"ignored": false,
"ignore_reason": null
},
{
"start_sec": 6,
"end_sec": 8,
"band": "air",
"direction": "boost",
"energy_db": -56.531,
"baseline_db": -66.997,
"delta_db": 10.466,
"min_delta_db": 10.466,
"max_delta_db": 10.466,
"energy_start_db": -56.531,
"energy_end_db": -56.531,
"entry_jump_db": 10.466,
"exit_jump_db": null,
"internal_range_db": 0,
"trend_db": 0,
"monotonic_fraction": 0,
"shape": "transient",
"window_count": 1,
"fault_level": false,
"classification": "informational",
"ignored": false,
"ignore_reason": null
},
"…trimmed 3 more"
],
"ignore_ranges": [],
"silence_gaps": [],
"total": 2
}
}
loudness_timeline
Arguments
{
"source": "$CLIP",
"start_sec": 8,
"end_sec": 12,
"window_sec": 2
}
Result
{
"ok": true,
"data": {
"source_path": "~/hill/gallery/tmp/R-loud/cards/loud-music/clip.mp4",
"start_sec": 8,
"end_sec": 12,
"window_sec": 2,
"bpm": null,
"silence_threshold_dbfs": -60,
"peak_dbfs": -7.417,
"peak_sec": 8.009,
"windows": [
{
"start_sec": 8,
"end_sec": 10,
"rms_dbfs": -20.376,
"peak_dbfs": -7.417,
"band_db": {
"sub": -26.869,
"low_mid": -20.638,
"mid": -27.051,
"high": -46.134,
"air": -65.991
},
"spectral_centroid_hz": 361.003,
"silent": false
},
{
"start_sec": 10,
"end_sec": 12,
"rms_dbfs": -20.663,
"peak_dbfs": -13.157,
"band_db": {
"sub": -26.637,
"low_mid": -20.787,
"mid": -27.031,
"high": -47.149,
"air": -67.282
},
"spectral_centroid_hz": 337.82,
"silent": false
}
],
"outliers": [],
"ignore_ranges": [],
"silence_gaps": [],
"total": 2
}
}
loudness_timeline
Arguments
{
"source": "$CLIP",
"start_sec": 12,
"end_sec": 16,
"window_sec": 2
}
Result
{
"ok": true,
"data": {
"source_path": "~/hill/gallery/tmp/R-loud/cards/loud-music/clip.mp4",
"start_sec": 12,
"end_sec": 16,
"window_sec": 2,
"bpm": null,
"silence_threshold_dbfs": -60,
"peak_dbfs": -13.07,
"peak_sec": 14.899,
"windows": [
{
"start_sec": 12,
"end_sec": 14,
"rms_dbfs": -19.973,
"peak_dbfs": -13.136,
"band_db": {
"sub": -25.555,
"low_mid": -19.87,
"mid": -27.679,
"high": -48.938,
"air": -65.179
},
"spectral_centroid_hz": 278.142,
"silent": false
},
{
"start_sec": 14,
"end_sec": 16,
"rms_dbfs": -19.94,
"peak_dbfs": -13.07,
"band_db": {
"sub": -24.188,
"low_mid": -19.89,
"mid": -28.105,
"high": -48.207,
"air": -64.738
},
"spectral_centroid_hz": 275.098,
"silent": false
}
],
"outliers": [],
"ignore_ranges": [],
"silence_gaps": [],
"total": 2
}
}
Question and answer
- Question
- Where does it suddenly get loud?
- Answer
- The loudest 2-second stretch runs from 6 to 8 seconds, at -8.8 dB RMS, the loudest of the 8 windows across the clip.
Check passedIoU 1.00, minimum 0.5
How it was checked
- Truth from
- constructed
- Fixed at
Planted by this recipe before any tool ran: ffmpeg cut the first 18 s of the source (its loudest sample measured at -0.1 dBFS by volumedetect), the whole clip was first lowered 12.9 dB so the raised stretch peaks below full scale, and raised the sound from 6 s to 8 s by 12 dB with a volume filter (the picture is not changed). The truth is that stretch, 6 to 8 s, checked with an IoU of at least 0.5. The calls read 2-second windows that start at 0, 2, 4 and so on, so the planted stretch is the window 6 to 8 s.
A script raised the sound by 12 dB from 6 to 8 seconds, a planted change. To keep the raised stretch below full scale, the whole clip was first lowered 12.9 dB. The answer is the loudest 2-second window, so the card does not test timing finer than 2 seconds. Only the first 16 seconds of the source were analysed.
Try it yourself
Use agentic-video to find where the sound in my video suddenly gets loud. Run loudness_timeline on my file in 4-second ranges that cover the first 16 seconds, with a window of 2 seconds each time. Tell me the start and end of the loudest window across all the runs, and how loud it is compared with the windows around it.
- Get the .mcpb file and open it in Claude Desktop.
- Or:
pip install agentic-video
Clip and licence
- Licence
- CC0
- Author
- B.G. DeSylva Productions Inc
- Source
- commons.wikimedia.org
This file is made available under the Creative Commons CC0 1.0 Universal Public Domain Dedication.