## About the Dataset
### 604 hours

This audio dataset contains 604 hours of French Speech Data in various domains, recorded by speakers from France and Canada.

French (Canada) domain distribution per dataset:

- 12.55 hours of Banking
- 12.7 hours of Insurance
- 29.78 hours of Retail
- 45.82 hours of Telecommunication

French (France) domain distribution per dataset:

- 118.27 hours of Banking
- 126.48 hours of Insurance
- 122.25 hours of Retail
- 137.43 hours of Telecommunication

IVR stands for Interactive Voice Response. The Spontaneous IVR data is created by having a human respond to an IVR system. The human (playing a customer) is following simplified real-life scenarios, making a spontaneous query on the given topic. The IVR system will ask the human to repeat his/her query in 2 more different ways. The speech is then transcribed. In rare cases, the IVR system is mimicked by another human. The recording is done via telephony and is saved in 8kHz 16 bit per channel.

The dataset is covered by [Defined.ai's standard license agreement](https://www.defined.ai/dataset/data-license-agreement). The license agreement is perpetual and allows for the commercialization of all models built on the data.

Other characteristics:

- WER: 3,68%
- Audio format: WAV
- Recording environment: noisy, silent
- Communication band: narrowband
- Sample rate: 8kHz
- Device type: mobile

## Metadata Distribution

#### French (Canada)

![IVR_French_CA_Age.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/IVR_French_CA_Age_5c18620100.png)
![IVR_French_FR_Age.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/IVR_French_FR_Age_42327e2d68.png)
![IVR_French_CA_Accents.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/IVR_French_CA_Accents_4e89552fcd.png)

#### French (France)

![IVR_French_FR_Accents.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/IVR_French_FR_Accents_15e8bbac4f.png)
![IVR_French_FR_Age.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/IVR_French_FR_Age_42327e2d68.png)
![IVR_French_FR_Gender.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/IVR_French_FR_Gender_14cbee4b36.png)

## Samples
#### Short 5-minute samples
Check this 5-minute audio sample of French (France) Banking [here](https://defineddata.blob.core.windows.net/samples/AFI_fr-fr_single-spontaneous-IVR_banking_30m_v01_Sample/Audio/2f4860a5-d109-4977-8253-6f9563836bea.wav?se=2024-06-15T15%3A34%3A30Z&sp=r&sv=2020-06-12&ss=b&srt=o&sig=d1nR5TSorDJVG2nXjSHtptrItel2USg42rndISgb9gI%3D). Transcription for the sample is also [available](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/IVR_French_France_Banking_Short_Transcription_f5f401d97a.tsv)

#### 30-minute Sample
[AFI_fr-fr_single-spontaneous-IVR_banking_30m_v01_Sample.zip](https://defineddata.blob.core.windows.net/samples/AFI_fr-fr_single-spontaneous-IVR_banking_30m_v01_Sample/AFI_fr-fr_single-spontaneous-IVR_banking_30m_v01_Sample.zip?se=2024-06-15T15%3A34%3A30Z&sp=r&sv=2020-06-12&ss=b&srt=o&sig=d1nR5TSorDJVG2nXjSHtptrItel2USg42rndISgb9gI%3D)

Download Free 30-minute Sample

Speech

Mandarin Chinese (PRC) Spontaneous Dialogue

French Spontaneous Dialogue

Dutch Scripted Monologue

Bengali Spontaneous Dialogue

You might also be interested in these audio datasets:

French Spontaneous IVR

## About the Dataset
### 809 hours

This audio dataset contains 809 hours of French Speech Data in various domains, recorded by native speakers from France and Canada.

There are **305 hours of French (France)** human-to-human audio with the following domain distribution per dataset:
- 78.8 hours of Banking
- 46.8 hours of Insurance
- 88.58 hours of Retail
- 91.3 hours of Telecommunication

There are **504 hours of French (Canada)** human-to-human audio with the following domain distribution per dataset:
- 122.72 hours of Banking
- 128.13 hours of Insurance
- 138.92 hours of Retail
- 114.42 hours of Telecommunication

Defined.ai creates scenarios for our crowd members to follow, which they study beforehand. They then record a conversation, one speaker playing the agent, the other speaker “playing out” the scenario with spontaneous content. The recording is done via telephony and is saved in 8khz 16 bit per channel. That content is then transcribed.

The dataset is covered by [Defined.ai's standard license agreement](https://www.defined.ai/dataset/data-license-agreement). The license agreement is perpetual and allows for the commercialization of all models built on the data.

Other characteristics:

- Audio format: WAV
- Recording environment: noisy, silent
- Bits per sample: 16
- Communication band: broadband
- Sample rate: 8kHz

## Metadata Distribution

#### French (France)
![Spontaneous_French_FR_Age.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_French_FR_Age_0b862a68b1.png)
![Spontaneous_French_FR_Gender.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_French_FR_Gender_f00373a8b1.png)

#### French (Canada)

![Spontaneous_French_CA_Accents.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_French_CA_Accents_b9e30f2011.png)
![Spontaneous_French_FR_Age.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_French_FR_Age_0b862a68b1.png)
![Spontaneous_French_FR_Gender.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_French_FR_Gender_f00373a8b1.png)

## 5-minute audio samples

- [French (France)](https://defineddata.blob.core.windows.net/samples/AFI_fr-fr_multi-spontaneous_retail_30m_v01_Sample/Audio/3d0cd78d-b470-4abb-a961-8d500fef127e.wav?se=2024-06-15T16%3A54%3A29Z&sp=r&sv=2020-06-12&ss=b&srt=o&sig=vY95n9soIiP4UHsHz5rGolLq8n3ALUvoc2cbUeCCxhk%3D). Transcription for the sample is also [available](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_French_France_Retail_Short_Transcription_bbe92b591c.tsv)
- [French (Canada)](https://defineddata.blob.core.windows.net/samples/AFI_fr-ca_multi-spontaneous_insurance_30m_v02_Sample/Audio/de0d53b7-bcc1-4e5b-833c-8082377d961d.wav?se=2024-07-13T14%3A29%3A20Z&sp=r&sv=2020-06-12&ss=b&srt=o&sig=NFK4/meTFfgzAQ/s0o3IWLL7tyPuckbVK89Rncs0Hlg). Transcription for the sample is also [available](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_French_Canada_Insurance_Short_Transcription_447410fbc8.tsv)

305 hours recorded by speakers from France, 504 hours by speakers from Canada

## About the Dataset
### 2310 hours

This audio dataset contains 1045 hours of Dutch Speech Data recored by native Dutch speakers from Belgium, and 1265 hours recorded by speakers from Netherlands.

There are 1045 hours of Dutch (Belgium) scripted monologue with the following domain distribution per dataset:
- 104.5 hours of Automotive
- 224.02 hours of Banking
- 47.27 hours of Generic domain
- 224.02 hours of Insurance
- 223.93 hours of Retail
- 221.33 hours of Telecommunication

There are 1265 hours of Dutch (Netherlands) scripted monologue with the following domain distribution per dataset:
- 308.98 hours of Banking
- 158.43 hours of Generic domain
- 343.43 hours of Insurance
- 286.65 hours of Retail
- 167.53 hours of Telecommunication

The speakers are presented with a prompt (script) and asked to read it out loud and record. Our clients will receive an audio recording, the prompt and information about the speaker. The audio is recorded on-device, typically in 16kHz 16 bit. We also provide information on which device each record was recorded.

The dataset is covered by [Defined.ai's standard license agreement](https://www.defined.ai/dataset/data-license-agreement). The license agreement is perpetual and allows for the commercialization of all models built on the data.

## Metadata Distribution

### Dutch (Netherlands)

![Scripted_Dutch_NL_Gender.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Scripted_Dutch_BE_Gender_2941b07a6b.png)
![Scripted_Dutch_NL_Age.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Scripted_Dutch_BE_Age_8b312ac140.png)

## Samples
- [5-minutes sample Dutch Netherlands](https://defineddata.blob.core.windows.net/samples/AFI_nl-nl_single-scripted_generic_30m_v01_Sample/Audio/102659837.wav?se=2024-06-15T15%3A33%3A46Z&sp=r&sv=2020-06-12&ss=b&srt=o&sig=IcgS/Xkm3pXO2lLdG1a0kM/f%2BiKaUc8suN0EaSv%2BK7s%3D). Transcription for the sample is also [available](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Scripted_Dutch_Netherland_Generic_Short_Transcription_cd645c348b.tsv)


For samples of Dutch (Belgium) please contact us.

1045 hours of Dutch (Belgium), 1265 hours of Dutch (Netherlands)

## About the Dataset
### 1082 hours

This audio dataset contains 1082 hours of Mandarin Chinese Speech Data in Banking, Telecommunication, Insurance and Retail domains, recorded by native Mandarin Chinese speakers from PRC.

Domain distribution per dataset:

- 220.5 hours of Banking
- 261.88 hours of Insurance
- 281.47 hours of Retail
- 318.67 hours of Telecommunication

Defined.ai creates scenarios for our crowd members to follow, which they study beforehand. They then record a conversation, one speaker playing the agent, the other speaker “playing out” the scenario with spontaneous content. The recording is done via telephony and is saved in 8khz 16 bit per channel. That content is then transcribed.

The dataset is covered by [Defined.ai's standard license agreement](https://www.defined.ai/dataset/data-licence-agreement). The license agreement is perpetual and allows for the commercialization of all models built on the data.

Other characteristics:

- Audio format: WAV
- Recording environment: noisy, silent
- Bits per sample: 16
- Communication band: broadband
- Sample rate: 8Hz

## Samples

Check this 5-minute audio sample [here](https://defineddata.blob.core.windows.net/samples/ADR_zh-cn_multi-spontaneous_telco_B1_v01_30m_Sample/Audio/3854204e-97d2-4a6d-a7b6-d25483a12c7d.wav?se=2025-07-11T14%3A59%3A23Z&sp=r&sv=2020-06-12&ss=b&srt=o&sig=aek/N0AmadNo27aTShHl36FDNmeLYBaXmlm9bG2DTDM%3D). Transcription for the sample is also [available](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_Mandarin_Chinese_Short_Transcription_7812162b6e.tsv?updated_at=2023-06-06T13:45:51.924Z).

## About the Dataset
### 256 hours

This audio dataset contains 256 hours of Bengali Speech Data in Banking, Telecommunication, Insurance and Retail domains, recored by native Bengali speakers from India.

Domain distribution per dataset:

- 102.93 hours of Banking
- 86.98 hours of Insurance
- 29.32 hours of Retail
- 36.93 hours of Telecommunication

Defined.ai creates scenarios for our crowd members to follow, which they study beforehand. They then record a conversation, one speaker playing the agent, the other speaker “playing out” the scenario with spontaneous content. The recording is done via telephony and is saved in 8kHz 16 bit per channel. That content is then transcribed.

The dataset is covered by [Defined.ai's standard license agreement](https://www.defined.ai/dataset/data-licence-agreement). The license agreement is perpetual and allows for the commercialization of all models built on the data.

Other characteristics:

- Audio format: WAV
- Recording environment: noisy, silent
- Bits per sample: 16
- Communication band: broadband
- Sample rate: 8kHz

## Metadata Distribution
![Spontaneous_Bengali_Accents.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_Bengali_Accents_419d27c9bf.png)
![Spontaneous_Bengali_Gender.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_Bengali_Gender_f58aebc312.png)
![Spontaneous_Bengali_Age.png](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_Bengali_Age_1355f95eec.png)

## Samples

Check this 5-minute audio sample [here](https://defineddata.blob.core.windows.net/samples/AFI_bn-in_multi-spontaneous_banking_30m_v01_Sample/Audio/ffac8a94-58cc-41ea-88a4-49ef48b182d4.wav?se=2024-11-30T12%3A51%3A34Z&sp=r&sv=2020-06-12&ss=b&srt=o&sig=FvbczXwxp46wbqUOYCTLjvfkAvf3xA9flOW78RYlv0s%3D). Transcription for the sample is also [available](https://prdstrapimediastorage.blob.core.windows.net/prdstrapimediastorage/assets/Spontaneous_Bengali_Short_Transcription_4261ca8bab.tsv?updated_at=2023-06-05T16:32:18.221Z).

100 hours of French Canadian and 504 hours of French (France)