Dataset Viewer
Auto-converted to Parquet
audio
audioduration (s)
0.9
20.2
text
stringlengths
3
205
He also conducts the team's fantasy camps and occasionally commentates on Giants' games.
After Doc Rivers was ejected, the Hawks were down to only four players.
The party operates a weekly newspaper, "unsere Zeit".
Its sloping lawns from a natural amphitheatre.
The triquetral is one of the eight carpal bones of the hand.
Highfield House was another historical building.
Roy Hazelwood was born in Pocatello, Idaho.
At the foot of Gunnerside Ghyll is the Kings Head public house.
Daren Powell's father was a cricketer.
He died from a cardiac failure.
They settled in the small rural community of Reach, Ontario.
Monty pays Simon's fine to get his thesis back.
Other live music, including piano, was also a staple in the old Shakey's parlors.
They are light gray on top with a white rump and a white underside.
He and his siblings were raised as Catholics.
It was not true client-server because processing occurred in the clients.
The goose was brought straight from the old market.
This strong arm shall shield your honor.
Many scientists work their entire careers on this, and you say you have solved this in a single afternoon?
The company elected to expand globally, and acquired other entities.
Some of Balada's works have been recorded by Naxos Records.
Paris hosts what it claims as the "World's Biggest Fish Fry".
That is a weird phrase.
Students may also earn a Certificate in Dispute Resolution.
The Mirror Stage has also a significant symbolic dimension.
Also, it meets requirements of both cargo and passenger transport.
Local solutions should be preferred to cloud solutions in many cases.
The shortest, simplest course is Watkins Glen.
During World War One, the Red Cross operated a canteen from a boxcar.
Firstly, those presented in a yellow-and-silver themed dust jacket relate to 'non-motor-propelled' ships.
Taylor was born in Torrance, California.
Much of the town was in flames.
The first teen winner was Ashley Anglin.
The company was sold on the exchange.
The Chersky Range is part of the South Siberian System.
The song was produced by Christoph Oswald.
A blond girl poses for a picture at Kerry Park in Seattle.
A woman in a blue shirt and headscarf poses for a picture sitting at a picnic table.
Two dogs on the beach.
Eleven other editions of the song were published with different lyrics.
It has a separate night transmitter site.
The town was named for Frances Brooks, the daughter of a railroad official.
Leave to apply for mandamus is hereby refused.
The joke originated on Usenet, and has appeared in several online parodies.
However, the music from the former anthem was retained.
Didn't Charles have any friends?
They're not my children.
There are several small lakes in the northwest and southwest sections.
Three knappers used the same direction all the time.
Thus the expected reinforcements did not come from the north and the uprising failed.
The procedure is usually only performed when other methods have failed.
This friendship would guide all three careers.
France and Wales were automatic qualifiers.
I am an Electrical Engineer.
She loved him dearly; he knew it, and I knew it.
That I cannot tell you.
During the Turkish occupation of Hungary, most of the Turks lived in this neighbourhood.
It was recorded at The Power Plant in London.
There was also the potential for a naval contest.
Nonetheless, all three are considered important contributors to mathematics, despite the variance in method.
Basil of Annonay, France.
Dogs will normally bark just as much as before the procedure.
This fee covers accident insurance and other administrative costs.
Tickell contributed to "The Spectator" and "The Guardian".
It then located one of Franklin's two ships, preserved in good condition.
A few people watch a woman in a black dress dance in front of a red backdrop.
Two women are looking at a computer monitor.
She assists Solid Snake in Eastern Europe.
Topical songs are often, but not always, protest songs.
He assisted his wife with the founding of the school.
The library's greatest strength, however, lies in its French holdings.
While traveling, they meet Princess Nastoya, the leader of the Scrow, a Vinkus tribe.
This resulted in another condominium, this time between Hesse and the Palatinate.
Besides being a talented and tactical competitor Zamba was incredibly popular during her reign.
Coach Gene Shue's eight-man rotation sank more field goals than any other team.
The township borders Woolwich Township, Harrison Township, Elk Township, and Salem County.
Users can install the plugins via Plugin Manager.
Many artists have covered the song.
As Ringo attempts to intervene, Holliday shoots him in the arm.
He went on to lose the general election to Hefner.
This has since become the train.
This region is where gold was first discovered in the United States.
His parents were the comedy double act Callan and Emery.
The Goethe-Institut is active in Australia, there are branches in Melbourne and Sydney.
Bernard Lander called the Lander College for Men the "flagship" of Touro College.
There is also access for bird watching and walking.
Many popular singers have recorded and performed "Some Enchanted Evening".
Mauritanian refugees slowly trickled back into the country during the following years.
Is the same work amplified later with the title of "Europa portuguesa".
The chemical is also marketed for use as a gaseous fire suppressant.
I do as I please.
James Hinton, Junior praised her "meltingly lovely middle voice".
It was initially used for a survey of a thousand stars.
There are only two main populations of the breed.
This effect remains until the taste buds return to normal.
Homes, roads, electricity, telephone and sewer services were all suddenly in short supply.
He began to ply Luca with questions.
The winner is announced in the fall of each year.
Mouth begins a relationship with his boss Alice, to get ahead in his job.
He contributes regularly to La Agenda magazine to Radio Metro.
End of preview. Expand in Data Studio

Common Voice 22 - English Teens Subset

This dataset is a filtered subset of Mozilla's Common Voice Corpus 22.0, containing only recordings from teenage speakers in English.

Dataset Structure

Data Instances

Each instance contains:

  • audio: Audio file with embedded bytes, sampling rate, and path
  • text: Transcription of the spoken text

Data Splits

This dataset provides a single unified split containing all 77,500 samples. Users can create their own train/validation/test splits based on their specific needs.

Split Samples
train 77,500

Speaker Demographics

  • Unique Speakers: 2,166 teenage contributors
  • Average Recordings per Speaker: ~35.9

Gender Distribution:

  • Male/Masculine: 41,539 (53.4%)
  • Female/Feminine: 25,282 (32.5%)
  • Unspecified: ~10,974 (14.1%)
  • Other: 12 (0.01%)

Audio Specifications

  • Sample Rate: 16,000 Hz (16 kHz)
  • Channels: Mono
  • Format: MP3
  • Audio Storage: Embedded directly in the dataset (no external files needed)

Text Transcriptions

  • Total Transcriptions: 77,500
  • Average Length: ~62 characters
  • Range: 3 to 17,096 characters

Usage

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("cryptolock/common-voice-22-en-teens")

# Access the data
train_data = dataset['train']

# Example: Iterate through samples
for sample in train_data:
    print(f"Text: {sample['text']}")
    print(f"Audio sampling rate: {sample['audio']['sampling_rate']}")
    # Access audio array: sample['audio']['array']

Creating Custom Splits

from datasets import load_dataset

# Load dataset
dataset = load_dataset("cryptolock/common-voice-22-en-teens")

# Create train/val/test splits (e.g., 80/10/10)
train_test = dataset['train'].train_test_split(test_size=0.2, seed=42)
test_val = train_test['test'].train_test_split(test_size=0.5, seed=42)

splits = {
    'train': train_test['train'],      # 80%
    'validation': test_val['train'],   # 10%
    'test': test_val['test']           # 10%
}

print(f"Train: {len(splits['train'])} samples")
print(f"Validation: {len(splits['validation'])} samples")
print(f"Test: {len(splits['test'])} samples")

Fine-tuning Whisper Example

from datasets import load_dataset
from transformers import WhisperProcessor, WhisperForConditionalGeneration

# Load dataset and model
dataset = load_dataset("cryptolock/common-voice-22-en-teens")
processor = WhisperProcessor.from_pretrained("openai/whisper-tiny")
model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-tiny")

# Prepare data
def prepare_dataset(batch):
    audio = batch["audio"]
    batch["input_features"] = processor(
        audio["array"],
        sampling_rate=audio["sampling_rate"],
        return_tensors="pt"
    ).input_features[0]
    batch["labels"] = processor.tokenizer(batch["text"]).input_ids
    return batch

# Process dataset
dataset = dataset.map(prepare_dataset, remove_columns=dataset.column_names)

# Continue with training...

Use Cases

This dataset is particularly useful for:

  • Training and fine-tuning speech recognition models on teenage voices
  • Studying acoustic characteristics of adolescent speech
  • Building age-specific ASR systems optimized for youth demographics
  • Voice activity detection for teenage speakers
  • Educational and research purposes in speech processing
  • Benchmarking ASR models on teen voice data

Dataset Creation

Source Data

This dataset was derived from the Mozilla Common Voice Corpus 22.0 released on June 20, 2025.

Data Collection Process

  1. Downloaded the full Common Voice 22.0 English dataset
  2. Filtered all recordings where the speaker's self-reported age group was "teens"
  3. Extracted corresponding audio clips and metadata
  4. Embedded audio bytes into Arrow format for self-contained distribution
  5. Combined into a single unified dataset (no pre-defined splits)

Annotations

The transcriptions were crowd-sourced and validated by Mozilla Common Voice contributors. Each recording includes:

  • Human-verified text transcription
  • Speaker metadata (age group, gender)
  • Quality metrics (upvotes/downvotes from validators)

Quality

  • All samples verified as teen age group
  • Audio validated and embedded
  • Self-contained dataset (no external file dependencies)
  • Ready for immediate use

Licensing

This dataset inherits the CC0 1.0 Universal (Public Domain) license from Mozilla Common Voice.

You are free to:

  • ✅ Use the dataset for any purpose (commercial or non-commercial)
  • ✅ Redistribute and modify without restrictions
  • ✅ Use without attribution (though attribution is appreciated)

No rights reserved. The dataset is dedicated to the public domain.

Citation

If you use this dataset, please consider citing the original Mozilla Common Voice project:

@inproceedings{commonvoice:2020,
  author = {Ardila, R. and Branson, M. and Davis, K. and Henretty, M. and Kohler, M. and Meyer, J. and Morais, R. and Saunders, L. and Tyers, F. M. and Weber, G.},
  title = {Common Voice: A Massively-Multilingual Speech Corpus},
  booktitle = {Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020)},
  pages = {4211--4215},
  year = {2020}
}

And optionally mention this filtered subset:

@dataset{common_voice_22_teens,
  author = {cryptolock},
  title = {Common Voice 22 - English Teens Subset},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/cryptolock/common-voice-22-en-teens}
}

Additional Information

Dataset Curators

This subset was created and curated by cryptolock to make age-specific speech data more accessible for research and development.

Supported Tasks

  • Automatic Speech Recognition (ASR)
  • Speech-to-Text
  • Voice Activity Detection
  • Speaker Verification
  • Acoustic Analysis

Languages

  • English (en)

Source

Known Limitations

  • Dataset contains 77,500 samples out of the original 77,807 teen samples from CV22 (307 samples or 0.4% missing due to processing)
  • Age groups are self-reported by contributors
  • Audio quality varies based on recording equipment used by contributors
  • Some transcriptions may contain minor errors or variations

Updates

  • v1.0 (November 2025): Initial release with 77,500 samples

Contact

For questions about this dataset subset, please open a discussion on the dataset page.

For questions about the original Common Voice dataset, visit https://commonvoice.mozilla.org/


Acknowledgments: This dataset would not be possible without the thousands of Mozilla Common Voice contributors who donated their voices and time to create this public domain resource.

Downloads last month
69