< Go to Catalog
Spontaneous IVR

Arabic Speech Data - Spontaneous IVR

Version Number
Published date
Mar 7, 2022
Audio demo Audio clips from the dataset that you can listen to.

How can I get the dataset?

After clicking the button and filling out the form, we will contact you to discuss the details.

Not what you're looking for?

We can collect a customized dataset according to your precise needs.

Use case(s)

call centre, conversational AI, IVR

Model Applications

Acoustic Modelling, ASR Testing, Benchmarking

Packaging description

A zip file containing metadata files in tsv format and a folder with all the audio files
This dataset contains 68 hours of Arabic Spontaneous IVR data, recorded from speakers in Yemen.

Seller Name


Dataset details


Domain Telecommunication
Total recordings 6588
File size 3.71GB
Hours 68
Word error rate (%) Measurement indicating errors in alignment of text representation (actual vs. perfect) of audio, taking into account words omitted, inserted or wrongly replaced. 14.35%
Average amount of recordings per speaker 158.75


Locale The language(s) and country(s) applicable to the speakers in the dataset. ar-ye
Language Arabic
Country Yemen
Female | Male | Unspecified View on chart 5% | 61% | 34%
Age View on chart 19-54
Accent(s) View on chart Abyan, Aḑ Ḑāli‘, Al Bayḑā’, Al Ḩudaydah, Al Jawf, Al Maḩwīt, Amanat al ‘Asimah, Dhamar, Ḩaḑramawt, Ibb, Other, Raymah, Şanʻā’, Tāʻizz, ‘Adan, ‘Amran

Audio Details

Words 329461
Recording environment noisy, silent
Audio format WAV
Bits per sample 8
Device type mobile
Communication band narrowband
Sample rate 8kHz

Details on charts

Want to try?