audio
dict
transcript
stringlengths
24
14.4k
language_code
stringclasses
1 value
subset
stringclasses
1 value
full_language
stringclasses
1 value
switch_id
stringlengths
36
36
segment_id
stringclasses
1 value
transcript_filename
stringlengths
24
24
audio_len_sec
int64
188k
9.83M
orig_file_start
float64
21.6
1.23k
orig_file_end
float64
344
1.8k
channel
int64
0
1
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAABBlgAfC0gAAgUHCgwOERMWGRsfIiQnKSsuMDM2ODs8P0JER0pNUFJVWFpdX2(...TRUNCATED)
"aha . \naha . \ndenen geht es allen okay .\nha ? \nwelcher ?\nder, der eine mit Problemen, ne ? \nj(...TRUNCATED)
de
CallHome
German
048754f3-d580-4d20-9c3d-5b7a9cf56885
None
CallHome/German/5566.cha
9,669,625
122.022125
726.373688
0
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAAAfJQAMP2AAAwUICg0PEhQXGRweISMmKCstMDI1Nzo8P0FERkpMTlFUVllbXm(...TRUNCATED)
"und wie geht es der anderen ? \nau Jesus@s:eng ! \ndas ist ja fürchterlich xxx . \n&=blows_through(...TRUNCATED)
de
CallHome
German
4e12854a-0c88-4283-b4d8-70f9924709fa
None
CallHome/German/4124.cha
4,591,040
59.78
346.72
0
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAAAg0gAPjxAAAgYHCg4QEhQXGhwfIyUoKiwvMTU3OTw+QURGSUtOUVNWWVtdX2(...TRUNCATED)
"was habe ich Neues erlebt, also diese . \ndiese Woche war ich in Köln auf einer Messe, &=breathes (...TRUNCATED)
de
CallHome
German
053446c2-cd51-4461-9fec-ec20499b84db
None
CallHome/German/6917.cha
4,838,225
122.261063
424.650125
1
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAAAHTwADXQwAAwUHCgwOEhQVGBscHyEjJigqLS8xNTc5PT9BREZIS05QVFZYW1(...TRUNCATED)
"gut . \nja . \nja, schön . \nja . \nja .\n&=noise . \nsoll ich es dir sagen ? \nsoll ich es dir sa(...TRUNCATED)
de
CallHome
German
361e0117-0d6e-4ad8-9e25-f138cff24b6f
None
CallHome/German/6350.cha
1,076,320
733.63
800.9
1
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAABA8wAevsgAAgUHCQsOEBIVFxkcHiAiJScpLC4wMzY5Oz9BREdJTE5RVFZaXF(...TRUNCATED)
"kannst du es jetzt einigermaßen ? \nja dann kannst du es . \nmhm . \nja okay . \n0 . \noh Gott ! \(...TRUNCATED)
de
CallHome
German
e70a38ee-ab09-4cc6-89a4-6288777e4903
None
CallHome/German/5097.cha
9,575,427
131.868563
730.33275
1
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAABBsAAeJ5wAAgUHCg0OERMVGRoeICIlJyksLjA0Njg6PUBCRUdKTU9SVVdaW1(...TRUNCATED)
"&+äh ja ich war zweimal diese Woche da zwei oder drei Malen . \n&=noise ja ? \naha aha aha . \n&=l(...TRUNCATED)
de
CallHome
German
84cf6536-06da-4452-a576-ee88364c4a7d
None
CallHome/German/6352.cha
9,684,368
590.742313
1,196.015313
0
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAAAfnAAQANQAAwYICw0QEhUXGx0gIiUnKiwwMjU3OjxAQ0VIS01PUlRXWVxeYW(...TRUNCATED)
"und dann sind wir &+ähm wieder ein Stückchen nach Norden gefahren, aber im Landesinneren und sind(...TRUNCATED)
de
CallHome
German
5f8d0461-e36c-4e4f-aba6-8f1d3525613e
None
CallHome/German/4659.cha
4,659,271
385.63
676.834438
0
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAABASgAdlVwAAgUHCgsOEBMVFxobHiAjJSgqLC8xMzY4Ojw/QUVISk1PU1VYXF(...TRUNCATED)
"nein, ich weiß nichts . \nich war viel weg in der, ich war zwei Wochenenden weg . \nnatürlich, na(...TRUNCATED)
de
CallHome
German
6832517f-7a4f-48a0-b798-9e3279e8c541
None
CallHome/German/4764.cha
9,478,502
1,036.51
1,628.916375
1
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAABBOQAeO0wAAwUHCQwOERIVGBodHyMlKCstMDM1ODk9QEJFR0tOUFNVWFtdYG(...TRUNCATED)
"&=laugh aha . \n&=breath ja ja klar . \n&=breath und &+äh hast du mal was vom Stefan gehört ? \na(...TRUNCATED)
de
CallHome
German
f6684ff1-4bb3-4dc3-8542-fc59c9518ba6
None
CallHome/German/4553.cha
9,616,320
91.48
692.5
1
{"bytes":"//OIxAAAAAAAAAAAAFhpbmcAAAAPAAA/4AAbpfAAAgUHCg0QEhYYGx0fIiQoKi0vMjU3Ojw/QUNFSEpNUFNVV1pcX2(...TRUNCATED)
"ah okay . \nja . \nja . \nja, der Tante Emmi ihre Größe ist ja viel größer, die ist wie meine G(...TRUNCATED)
de
CallHome
German
96a7a072-2236-4a84-9c22-98d335beca96
None
CallHome/German/4711.cha
9,417,338
75.179125
663.76275
0

Dataset Card

Dataset Description

This dataset is a benchmark based on the TalkBank[1] corpus—a large multilingual repository of conversational speech that captures real-world, unstructured interactions. We use CA-Bank [2], which focuses on phone conversations between adults, which include natural speech phenomena such as laughter, pauses, and interjections. To ensure the dataset is highly accurate and suitable for benchmarking conversational ASR systems, we employ extensive set of pre-processing.

Preprocessing Steps

We apply the following preprocessing steps to ensure the dataset’s quality:

  • Manual filtering of conversations
  • Speaker-channel alignment
  • Timestamp alignment using voice activity detection (VAD)
  • Discarding segments based on Word Error Rate (WER) thresholds

Paper and Code Repository

For a comprehensive explanation of the preprocessing pipeline and dataset details, refer to our paper ASR Benchmarking: The Need for a More Representative Conversational Dataset and explore our GitHub repository for code and additional resources.

Segmentation Types: Speaker Switch vs Annotation

We offer two types of segmentation for this dataset:

  • Annotation-based Segmentation: Segments are derived directly from the annotations provided in the original TalkBank corpus.
  • Speaker Switch Segmentation: We consolidate consecutive segments from the same speaker into a single, larger audio segment, providing an alternative structure for analysis.

Citations

While using this dataset please cite:

@article{maheshwari2024asr,
  title={ASR Benchmarking: Need for a More Representative Conversational Dataset},
  author={Maheshwari, Gaurav and Ivanov, Dmitry and Johannet, Th{\'e}o and Haddad, Kevin El},
  journal={arXiv preprint arXiv:2409.12042},
  year={2024}
}

In addition, please acknowledge the TalkBank dataset::

@article{macwhinney2010transcribing,
  title={Transcribing, searching and data sharing: The CLAN software and the TalkBank data repository},
  author={MacWhinney, Brian and Wagner, Johannes},
  journal={Gesprachsforschung: Online-Zeitschrift zur verbalen Interaktion},
  volume={11},
  pages={154},
  year={2010},
  publisher={NIH Public Access}
}

Licensing Information

This dataset is released under the CC BY-NC-SA 3.0.

References

[1]: MacWhinney, Brian. "TalkBank: Building an open unified multimodal database of communicative interaction." (2004). [2]: MacWhinney, Brian, and Johannes Wagner. "Transcribing, searching and data sharing: The CLAN software and the TalkBank data repository." Gesprachsforschung: Online-Zeitschrift zur verbalen Interaktion 11 (2010): 154.

Downloads last month
204
Edit dataset card