--- title: "Voices across Decades — Siegert 2026" description: "Studying how a person's voice and speaking style change over decades requires recordings of the same people across long time spans." authors: ["Ingo Siegert"] year: 2026 venue: "Proceedings of the Fifteenth Language Resources and Evaluation Conference" source: "https://aclanthology.org/2026.lrec-1.498/" doi: "10.63317/3vgihkgnkg75" licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)" licence_url: "https://creativecommons.org/licenses/by-nc/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/papers/siegert-2026-voices-across-decades-multimodal-diachronic/" --- [Elamur Research](https://md.research.elamur.ai/) · [this page in HTML](https://research.elamur.ai/free/papers/siegert-2026-voices-across-decades-multimodal-diachronic/) [home](https://md.research.elamur.ai/) / [papers](https://md.research.elamur.ai/free/papers/) / [2026](https://md.research.elamur.ai/free/papers/2026/page/5/) / Voices across Decades: A Multimodal Diachronic Corpus of German Bundestag Debates (GerParlDia-MM) # Voices across Decades: A Multimodal Diachronic Corpus of German Bundestag Debates (GerParlDia-MM) **Authors:** Ingo Siegert · **Year:** 2026 · **Venue:** Proceedings of the Fifteenth Language Resources and Evaluation Conference · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.lrec-1.498/) · [PDF](https://aclanthology.org/2026.lrec-1.498.pdf) · [DOI 10.63317/3vgihkgnkg75](https://doi.org/10.63317/3vgihkgnkg75) · [OpenAlex](https://openalex.org/W7160038730) **Citations:** 0 citations (openalex-snapshot · 2026-10-14) ## Summary Studying how a person's voice and speaking style change over decades requires recordings of the same people across long time spans. Existing speech corpora usually cover only a few years or a single speaker. German parliamentary collections such as GermaParl and OpenDiscourse provide transcripts but no time-aligned audio or video. GerParlDia-MM addresses this gap. It is a multimodal corpus of Bundestag speeches from 1949 to 2025, built around 75 deputies with unusually long careers, aiming for roughly one speech per person per year. The author chose the speakers using official Bundestag Open Data and obtained the recordings from two sources: the Parliamentary Archives and the Bundestag Mediathek. Transcripts were aligned to the audio with WhisperX. Speech boundaries were detected by combining regular expressions with meta-llama-3.1-8b-instruct, so that announcements by the presiding officer could be excluded. The resulting resource has rich person- and speech-level metadata and links to OpenDiscourse and the plenary protocols. It supports research on voice aging, rhetorical change and the long-term robustness of ASR and speaker recognition. Metadata, aligned transcripts and identifiers are released publicly, but the media files themselves are not redistributed. ## Key points - **Speaker selection:** Candidates came from 5,223 Bundestag members. To compensate for low female representation in early decades, thresholds differed by gender: at least 8 terms for men and 6 for women. This produced 75 speakers (43 men, 32 women), with individual coverage of about 18 to 53 years. - **Speeches:** 2,136 speeches of at least 120 s were retrieved, 87% of the 2,459 theoretically expected. About 67% come from the Parliamentary Archive, 53% include video, and 91% are linked to OpenDiscourse. - **Size:** The corpus holds about 2.5 M tokens and 1.26 TB of audiovisual data. - **Technical variety:** Recording quality spans analog archive audio to full-HD video, which makes the corpus useful for testing robustness to changing recording conditions. - **Party coverage:** SPD (33), CDU (20), CSU (9), Greens (6), FDP (4), DIE LINKE (3), plus one non-affiliated member. - **Availability:** Metadata are versioned on Zenodo with DOIs. Media are accessed through stable URLs because of legal restrictions. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* Fields: [Linguistics & languages](https://md.research.elamur.ai/free/shelves/linguistics/) ## Cite The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{siegert2026voices, title = {{Voices across Decades: A Multimodal Diachronic Corpus of German Bundestag Debates (GerParlDia-MM)}}, author = {Siegert, Ingo}, booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference}, year = {2026}, doi = {10.63317/3vgihkgnkg75}, url = {https://doi.org/10.63317/3vgihkgnkg75} } ``` - [← previous Variation Is the Norm: Embracing Sociolinguistics in NLP](https://md.research.elamur.ai/free/papers/lutgen-2026-variation-is-norm-embracing-sociolinguistics/) - [next → VoxKnesset: A Large-Scale Longitudinal Hebrew Speech Dataset for Aging Speaker Modeling](https://md.research.elamur.ai/free/papers/langstream-2026-marmor-voxknesset-large-scale-longitudinal-hebrew/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/papers/2026/page/5/) ## References 20 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Zhiqi Ai, Meixuan Bao, Zhiyong Chen, Zhi Yang, Xinnuo Li, and Shugong Xu. 2025. VoxAging: Continuously Tracking Speaker Aging with a Large-Scale Longitudinal Dataset in English and Mandarin. In Interspeech 2025, pages 3648– 3652. [find](https://scholar.google.com/scholar?q=Zhiqi+Ai%2C+Meixuan+Bao%2C+Zhiyong+Chen%2C+Zhi+Yang%2C+Xinnuo+Li%2C+and+Shugong+Xu%2E+2025%2E+VoxAging%3A+Continuously+Tracking+Speaker+Aging+with+a+Large%2DScale+Longitudinal+Dataset+in+English+and+Mandarin%2E+In+Inters) 2. Max Bain, Jaesung Huh, Tengda Han, Andrea Vedaldi, and Andrew Zisserman. 2023. Whisperx: Time-accurate speech transcription of long-form audio. arXiv preprint arXiv:2303.00747. [↗](https://arxiv.org/abs/2303.00747) 3. Andreas Blaette and Christoph Leonhardt. 2023. Germaparl corpus of plenary protocols (v2.0.1). Zenodo. [find](https://scholar.google.com/scholar?q=Andreas+Blaette+and+Christoph+Leonhardt%2E+2023%2E+Germaparl+corpus+of+plenary+protocols+%28v2%2E0%2E1%29%2E+Zenodo%2E) 4. Deutscher Bundestag. 2021a. Archivordnung für den deutschen bundestag. [https://www.bundestag.de/resource/blob/190272/archivordnung.pdf](https://www.bundestag.de/resource/blob/190272/archivordnung.pdf). Beschluss des Ältestenrates vom 21 March 2019, veröffentlicht 18 October 2021. [↗](https://www.bundestag.de/resource/blob/190272/archivordnung.pdf) 5. Deutscher Bundestag. 2021b. Nutzungsbedingungen für das audio- und videomaterial des parlamentsfernsehens. [https://www.bundestag.de/resource/blob/296016/](https://www.bundestag.de/resource/blob/296016/) b2b8e3ed04b91bbfb235cfed975f1a69/ nutzungsbedingungen_de-data.pdf. Parlamentsfernsehen, 25 March 2021. [↗](https://www.bundestag.de/resource/blob/296016/) 6. Deutscher Bundestag. 2021c. Nutzungsordnung für das parlamentsarchiv. [https://www.bundestag.de/resource/blob/190274/Nutzungsordnung.pdf](https://www.bundestag.de/resource/blob/190274/Nutzungsordnung.pdf). Anlage gemäß §3(10) der Archivordnung für den Deutschen Bundestag, veröffentlicht 18 October 2021. [↗](https://www.bundestag.de/resource/blob/190274/Nutzungsordnung.pdf) 7. Jonathan Harrington, Sallyanne Palethorpe, and Catherine Watson. 2000. Monophthongal vowel changes in received pronunciation: an acoustic analysis of the queen’s christmas broadcasts. Journal of the International Phonetic Association, 30(1–2):63–78. [↗](https://doi.org/10.1017/s0025100300006666) 8. Finnian Kelly, Andrzej Drygajlo, and Naomi Harte. 2012. Speaker verification with long-term ageing data. In 5th IAPR International Conference on Biometrics (ICB), pages 478–483. IEEE. [find](https://scholar.google.com/scholar?q=Finnian+Kelly%2C+Andrzej+Drygajlo%2C+and+Naomi+Harte%2E+2012%2E+Speaker+verification+with+long%2Dterm+ageing+data%2E+In+5th+IAPR+International+Conference+on+Biometrics+%28ICB%29%2C+pages+478%E2%80%93483%2E+IEEE%2E) 9. Philip M. McCarthy and Scott Jarvis. 2010. Mtld, vocd-d, and hd-d: A validation study of sophisticated approaches to lexical diversity assessment. Behavior Research Methods, 42(2):381–392. [↗](https://doi.org/10.3758/brm.42.2.381) 10. Gabriel Mittag, Babak Naderi, and Sebastian Möller. 2021. Nisqa: A deep cnn-self-attention model for multidimensional speech quality prediction with crowdsourced datasets. In Proceedings of Interspeech, pages 2127–2131. [find](https://scholar.google.com/scholar?q=Gabriel+Mittag%2C+Babak+Naderi%2C+and+Sebastian+M%C3%B6ller%2E+2021%2E+Nisqa%3A+A+deep+cnn%2Dself%2Dattention+model+for+multidimensional+speech+quality+prediction+with+crowdsourced+datasets%2E+In+Proceedings+of+Interspeec) 11. Arsha Nagrani, Joon Son Chung, Weidi Xie, and Andrew Zisserman. 2020. Voxceleb: Large-scale speaker verification in the wild. Computer Speech & Language, 60:101027. [↗](https://doi.org/10.1016/j.csl.2019.101027) 12. Florian Richter, Philipp Koch, Oliver Franke, Jakob Kraus, Fabrizio Kuruc, Anja Thiem, Judith Högerl, Stella Heine, and Konstantin Schöps. 2020. Open discourse. Harvard Dataverse. [find](https://scholar.google.com/scholar?q=Florian+Richter%2C+Philipp+Koch%2C+Oliver+Franke%2C+Jakob+Kraus%2C+Fabrizio+Kuruc%2C+Anja+Thiem%2C+Judith+H%C3%B6gerl%2C+Stella+Heine%2C+and+Konstantin+Sch%C3%B6ps%2E+2020%2E+Open+discourse%2E+Harvard+Dataverse%2E) 13. Debjoy Saha, Shravan Nayak, and Timo Baumann. 2022. Merkel podcast corpus: A multimodal dataset compiled from 16 years of angela merkel’s weekly video podcasts. arXiv preprint arXiv:2205.12194. [↗](https://arxiv.org/abs/2205.12194) 14. Ingo Siegert, Jannik Nietzold, Ralph Heinemann, and Andreas Wendemuth. 2019. The restaurant booking corpus – content-identical comparative human-human and human-computer simulated telephone conversations. In Elektronische Sprachsignalverarbeitung 2019. Tagungsband der 30. Konferenz, volume 93 of Studientexte zur Sprachkommunikation, pages 126–133, Dresden, Germany. TUDpress. [find](https://scholar.google.com/scholar?q=Ingo+Siegert%2C+Jannik+Nietzold%2C+Ralph+Heinemann%2C+and+Andreas+Wendemuth%2E+2019%2E+The+restaurant+booking+corpus+%E2%80%93+content%2Didentical+comparative+human%2Dhuman+and+human%2Dcomputer+simulated+telephone+conversati) 15. Ingo Siegert, Yamini Sinha, Oliver Jokisch, and Andreas Wendemuth. 2020. Recognition performance of selected speech recognition apis – a longitudinal study. In Speech and Computer, pages 520–529, Cham. Springer International Publishing. [find](https://scholar.google.com/scholar?q=Ingo+Siegert%2C+Yamini+Sinha%2C+Oliver+Jokisch%2C+and+Andreas+Wendemuth%2E+2020%2E+Recognition+performance+of+selected+speech+recognition+apis+%E2%80%93+a+longitudinal+study%2E+In+Speech+and+Computer%2C+pages+520%E2%80%93529%2C+Cham) 16. Vishwanath Pratap Singh, Md Sahidullah, and Tomi Kinnunen. 2023. Speaker verification across ages: Investigating deep speaker embedding sensitivity to age mismatch in enrollment and test speech. In Interspeech 2023, pages 1948– 1952. [find](https://scholar.google.com/scholar?q=Vishwanath+Pratap+Singh%2C+Md+Sahidullah%2C+and+Tomi+Kinnunen%2E+2023%2E+Speaker+verification+across+ages%3A+Investigating+deep+speaker+embedding+sensitivity+to+age+mismatch+in+enrollment+and+test+speech%2E+In+In) 17. Michael Tornow, Martin Krippl, Svea Bade, Angelina Thiers, Ingo Siegert, Sebastian Han-drich, Julia Krüger, and Andreas Wendemuth Lutz Schega. 2016. Integrated health and fitness (igf)-corpus - ten-modal highly synchronized subject-dispositional and emotional human machine interactions. In Multimodal Corpora: Computer vision and language processing (MMC 2016). Workshop Programme, Portorož, Slovenia. [find](https://scholar.google.com/scholar?q=Michael+Tornow%2C+Martin+Krippl%2C+Svea+Bade%2C+Angelina+Thiers%2C+Ingo+Siegert%2C+Sebastian+Han%2Ddrich%2C+Julia+Kr%C3%BCger%2C+and+Andreas+Wendemuth+Lutz+Schega%2E+2016%2E+Integrated+health+and+fitness+%28igf%29%2Dcorpus+%2D+ten%2Dmo) 18. Ravi Chander Vipperla, Steve Renals, and Joe Frankel. 2010. Ageing voices: The effect of changes in voice parameters on asr performance. EURASIP Journal on Audio, Speech, and Music Processing. [↗](https://doi.org/10.1155/2010/525783) 19. Linlin Wang, Jun Wang, Lantian Li, Thomas Fang Zheng, and Frank K. Soong. 2016. Improving speaker verification performance against long-term speaker variability. Speech Communication, 79:14–29. [↗](https://doi.org/10.1016/j.specom.2016.02.004) 20. Haoning Wu, Erli Zhang, Liang Liao, Chaofeng Chen, Jingwen Hou Hou, Annan Wang, Wenxiu Sun Sun, Qiong Yan, and Weisi Lin. 2023. Exploring video quality assessment on user generated contents from aesthetic and technical perspectives. In International Conference on Computer Vision (ICCV). [find](https://scholar.google.com/scholar?q=Haoning+Wu%2C+Erli+Zhang%2C+Liang+Liao%2C+Chaofeng+Chen%2C+Jingwen+Hou+Hou%2C+Annan+Wang%2C+Wenxiu+Sun+Sun%2C+Qiong+Yan%2C+and+Weisi+Lin%2E+2023%2E+Exploring+video+quality+assessment+on+user+generated+contents+from+aesth)