{
 "uuid": "7c542614-cebe-427a-b4e1-edc749162799",
 "handle": "uniba/111565",
 "type": "Publication",
 "name": "A Three-Level Evaluation Protocoll for Acoustic Scene Understanding of Large Language Audio Models",
 "url": "/entities/publication/7c542614-cebe-427a-b4e1-edc749162799/",
 "lastModified": "2026-02-03T02:59:32.117+00:00",
 "metadata": {
  "cris.virtual.author-orcid": [
   {
    "value": null
   },
   {
    "value": "0000-0003-4689-7944"
   }
  ],
  "cris.virtualsource.author-orcid": [
   {
    "value": "32681003-96e3-4522-a7c8-90c9bded5c32"
   },
   {
    "value": "85c5415c-f978-4558-a226-e71e8bc2cb80"
   }
  ],
  "dc.contributor.author": [
   {
    "value": "Dilip, Harish",
    "authority": "32681003-96e3-4522-a7c8-90c9bded5c32"
   },
   {
    "value": "Abeßer, Jakob",
    "authority": "85c5415c-f978-4558-a226-e71e8bc2cb80"
   }
  ],
  "dc.date.accessioned": [
   {
    "value": "2025-11-24T07:08:24Z"
   }
  ],
  "dc.date.available": [
   {
    "value": "2025-11-24T07:08:24Z"
   }
  ],
  "dc.date.issued": [
   {
    "value": "2025"
   }
  ],
  "dc.description.abstract": [
   {
    "value": "Reaching a semantic understanding of complex acoustic scenes requires computational models to capture the temporal-spatial sound source composition as well as individual sound events. This is a great challenge for computational models due to the large variety of everyday sound events and the extensive temporal-spectral overlap in real-life acoustic scenes. In this work, we aim to evaluate the acoustic scene understanding capabilities of two large audio-language models (LALMs). As a challenging scenario, we use the USM dataset, which features synthetic urban soundscapes with 2-6 overlapping sound sources per mixture. Our main contribution is a novel three-layer evaluation protocol, which includes four analysis tasks for low-level sound event understanding (sound event tagging), mid-level understanding and reasoning (sound polyphony estimation, sound source loudness ranking), as well as high-level scene understanding (audio captioning). We apply standardized metrics to assess the model's performances for each task. The proposed mulit-layer protocol allows for a fine-grained analysis of model behavior across soundscapes of various complexity levels. Our results indicate that despite their rmarkable controllability using textual instructions, the ability of state-of-the-art LALMs to understand acoustic scenes is still limited as the performance on individual analysis tasks desgrades with increasing sound polyphony."
   }
  ],
  "dc.identifier.doi": [
   {
    "value": "10.5281/zenodo.17251589"
   }
  ],
  "dc.identifier.isbn": [
   {
    "value": "978-84-09-77652-8"
   }
  ],
  "dc.identifier.uri": [
   {
    "value": "https://fis.uni-bamberg.de/handle/uniba/111565"
   }
  ],
  "dc.language.iso": [
   {
    "value": "eng"
   }
  ],
  "dc.publisher": [
   {
    "value": "zenodo"
   }
  ],
  "dc.relation.authororgunit": [
   {
    "value": null,
    "authority": "#PLACEHOLDER_PARENT_METADATA_VALUE#"
   },
   {
    "value": "Professur für Computational Humanities",
    "authority": "37c83a61-3124-4567-8bee-f8afc96ef659"
   }
  ],
  "dc.subject": [
   {
    "value": "audio captioning"
   },
   {
    "value": "large audio-language models"
   },
   {
    "value": "sound event tagging"
   },
   {
    "value": "sound polyphony estimation"
   },
   {
    "value": "sound source loudness ranking"
   }
  ],
  "dc.title": [
   {
    "value": "A Three-Level Evaluation Protocoll for Acoustic Scene Understanding of Large Language Audio Models"
   }
  ],
  "dc.type": [
   {
    "value": "conferenceobject"
   }
  ],
  "dspace.entity.type": [
   {
    "value": "Publication"
   }
  ],
  "ubg.conference.name": [
   {
    "value": "10th Workshop on Detection and Classification of Acoustic Scenes and Events, DCASE 2025"
   }
  ],
  "ubg.conference.place": [
   {
    "value": "Barcelona"
   }
  ],
  "ubg.editor.articlecollection": [
   {
    "value": "Benetos, Emmanouil"
   },
   {
    "value": "Font, Frederic"
   },
   {
    "value": "Fuentes, Magdalena"
   },
   {
    "value": "Martin Morato, Irene"
   },
   {
    "value": "Rocamora, Martín"
   }
  ],
  "ubg.fulltexticon": [
   {
    "value": "true"
   }
  ],
  "ubg.generated.csl": [
   {
    "value": "<div class=\"csl-bib-body\">\n  <div class=\"csl-entry\">Dilip, Harish; Abeßer, Jakob (2025): A Three-Level Evaluation Protocoll for Acoustic Scene Understanding of Large Language Audio Models, in: Emmanouil Benetos, Frederic Font, Magdalena Fuentes, u. a. (Hrsg.), <i>Proceedings of the 10th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE 2025), October 2025</i>, zenodo, S. 185–189, doi: 10.5281/zenodo.17251589.</div>\n</div>"
   }
  ],
  "ubg.generated.cslContainerTitle": [
   {
    "value": "Proceedings of the 10th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE 2025), October 2025"
   }
  ],
  "ubg.generated.cslDoi": [
   {
    "value": "10.5281/zenodo.17251589"
   }
  ],
  "ubg.generated.cslEditor": [
   {
    "value": "Emmanouil Benetos",
    "authority": "ba8a6830-36ac-401a-959e-32096a13d524"
   },
   {
    "value": "Font, Frederic",
    "authority": "cc4a8b52-0cb2-4a50-8fda-8a497cb0c51a"
   },
   {
    "value": "Fuentes, Magdalena",
    "authority": "cfd702c5-e606-4142-be21-2c72284f9829"
   },
   {
    "value": "Martin Morato, Irene",
    "authority": "5a85e27e-ff13-410d-aee4-033f5bf0c261"
   },
   {
    "value": "Rocamora, Martín",
    "authority": "ce20d025-86d3-404b-b30b-7c02cb43910f"
   },
   {
    "value": "Emmanouil Benetos",
    "authority": "ba8a6830-36ac-401a-959e-32096a13d524"
   },
   {
    "value": "Font, Frederic",
    "authority": "cc4a8b52-0cb2-4a50-8fda-8a497cb0c51a"
   },
   {
    "value": "Fuentes, Magdalena",
    "authority": "cfd702c5-e606-4142-be21-2c72284f9829"
   },
   {
    "value": "Martin Morato, Irene",
    "authority": "5a85e27e-ff13-410d-aee4-033f5bf0c261"
   },
   {
    "value": "Rocamora, Martín",
    "authority": "ce20d025-86d3-404b-b30b-7c02cb43910f"
   },
   {
    "value": "Benetos, Emmanouil",
    "authority": "6242138f-ea8b-44dd-890d-49cc847db4d1"
   },
   {
    "value": "Font, Frederic",
    "authority": "cc4a8b52-0cb2-4a50-8fda-8a497cb0c51a"
   },
   {
    "value": "Fuentes, Magdalena",
    "authority": "cfd702c5-e606-4142-be21-2c72284f9829"
   },
   {
    "value": "Martin Morato, Irene",
    "authority": "5a85e27e-ff13-410d-aee4-033f5bf0c261"
   },
   {
    "value": "Rocamora, Martín",
    "authority": "ce20d025-86d3-404b-b30b-7c02cb43910f"
   },
   {
    "value": "Benetos, Emmanouil"
   },
   {
    "value": "Font, Frederic"
   },
   {
    "value": "Fuentes, Magdalena"
   },
   {
    "value": "Martin Morato, Irene"
   },
   {
    "value": "Rocamora, Martín"
   }
  ],
  "ubg.intern.org": [
   {
    "value": "Professur für Computational Humanities",
    "authority": "37c83a61-3124-4567-8bee-f8afc96ef659"
   }
  ],
  "ubg.intern.vonfeld": [
   {
    "value": "",
    "language": "de_DE"
   }
  ],
  "ubg.intern.vonfeldeditor": [
   {
    "value": "",
    "language": "de_DE"
   }
  ],
  "ubg.internatorientation": [
   {
    "value": "Ja"
   }
  ],
  "ubg.pages.range": [
   {
    "value": "185-189"
   }
  ],
  "ubg.peerreview": [
   {
    "value": "Ja"
   }
  ],
  "ubg.submission.type": [
   {
    "value": "full-text"
   }
  ],
  "ubg.titleparent.articlecollection": [
   {
    "value": "Proceedings of the 10th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE 2025), October 2025"
   }
  ],
  "ubg.version.hasversions": [
   {
    "value": "uniba/112887",
    "authority": "ed9eaec7-f459-4495-a0fa-943f553761cd"
   }
  ]
 },
 "relations": {},
 "files": []
}