Datasets:

michsethowusu
/

twi-speech-text-parallel-multispeaker

audio audioduration (s) 0.04 60	text stringlengths 1 315
	Wɔbɛtumi akɔ dan a ɛtoa wɔn so no ne ne yɔnko a wadi mfeɛ akɔyɛ adidi kronkron no?
	Honhom no ate wɔn ho ama wɔn nnipadua no ayɛ afoforɔ.
	Saa kɛseyɛ no ba nkakrankakra.
	Na saa abayewa ketewa a me hyiraa no saa da no a,
	mɛbɔ mpaeɛ ma Adwurade ama mahunu deɛ Ɔpɛ ma onipa no,
	Mehunuu sɛ wɔn mu pii resu.
	mɛkasa akyerɛ mo a mopɛsɛ mode mo asɔfodie no yɛ adwuma yie wɔ mo ankasa mo som mu.
	mmom mehwɛɛ mpanyinfoɔ yi mu biara anim.
	Mɛhyɛ aseɛ afiri dikɔnfoɔ a wɔyɛ foforɔ koraa no so ɛfiri sɛ wɔn pa ara na wɔn adwene nsi pi wɔ nteaseɛ mu sɛ wɔde wɔn asɔfodie bɛsom.
	Masua sɛ wei ne ɔsom safoa na yɛahyira afoforɔ wɔ Ne din mu.
	Wɔbɛsiesie Awurade nkorɔfoɔ ama N’animuonyam Mmaeɛ a ɛtɔ so Mmienu no.
	Nanso ebia ɛbɛyɛ mo nwanwa sɛdeɛ saa ɔfrɛ yi mo bɛyɛ no kɛse kyerɛ ma mo.
	Meyɛɛ dikɔnni mfeɛ pii akyi ansa na meresua sɛdeɛ ɛkyerɛ sɛ wɔde di dwuma.
	Sɛ me ne mo kasa na mete mo gyedie kɛseɛ no nka a,
	Me gye dii sɛ m’asɔfodie asɛdeɛ ara ne sɛ mɛkyɛ adidi kronkron no wɔ me ankasa dan asa so.
	Ɔde asɔfodie no hyɛɛ wo nsa,
	na ɛwɔ sɛ ɔtena fie.
	Wɔ nnawɔtwe pii mu no,
	meda wo ase,
	” Awurade no ahyira me som,
	Nhwɛsoɔ,
	mesuaa wɔ berɛ a mɛyɛ tenenee anototoɔ mu sɛ,
	na ɛwɔ statio mfonyin bi wɔ aseɛ.
	Me suahunu ne sɛ mpo nhyira no nyɛ deɛ nnipa no pɛ ma wɔn ho anaa wɔn adɔfonom a,
	Metee suahunu bi nnansa yi a ɛkaee me fa saa ɔdɔ yi ho.
	na ɔkaa sɛ,
	Ɛtoo me berɛ bi wɔ ayarehwɛbea bi mu berɛ a na adɔkotafoɔ bi kaa sɛ—mpo —hyɛɛ me—sɛ me nyɛ ntɛm mfiri wɔn kwan mu ma wɔn nyɛ wɔn adwuma a wɔn amma me akwanya amma asɔfodie nhyira no.
	Sɛ woyɛ saa a,
	Wɔbɛboaboa Isreal ano.
	Wɔn a wɔahyɛ wɔn elder afoforɔ no bɛpɛ sɛ wɔbɛtie.
	Me nuanom adɔfoɔ,
	mate nhyira a wɔde ama a ayɛ me ne deɛ ɔregye nhyira no nso nwanwa.
	ne onuabaa a ɔsom no no tumi san kɔɔ asɔre.
	Ɛyɛ adesua ma me sɛ mɛhwɛ m’akyi wɔ me berɛ a na meyɛ dikɔn no.
	Ebia wɔde toto ho a wobɛhunu wo ho ketewa wɔ dwumadie sononko a Awurade bɛyɛ no mu.
	amen.
	Wɔkaa sɛ menkyɛ adidi kronkron no.
	na adɔkotafoɔ no susu sɛ ɔbɛwu no tenaa ase.
	ɛnyɛ sɛdeɛ mɛtumi ayɛ m’afa mu deɛ.
	Maame baako sɔɔ me nsa,
	a wɔde ama wɔ Ne din mu.
	Ɔsomfoɔ barima no—kae sɛ na n’asodie yɛ ne yɔnko no na ɛnyɛ saa ɔbaa panyini no ankasa—mpo bɛpem nnɛ yi gu so de twerɛsɛm ne paano ketewa bi hyɛ ne nsa mu Kwasiada biara,
	Berɛ a wɔtwee Asɔre nhyiamu nyinaa saneeɛ ɛnam COVID- nsanyadeɛ nti,
	nanso ɛbɛba.
	hwɛɛ soro,
	Ɔpɛ a ɔwɔ sɛ ɔbɛyɛ sɔfodwuma anyini berɛ a ɔresom wɔ Awurade din mu wɔ n’ankasa ne kwan a ɔnim so.
	onuabarima no,
	Ɔnim wo ankasa,
	San fa ma no bɛkɔ so.
	Mekae sɛ na ɛyɛ Yalecrest Ward a ɛwɔ Salt Lake City,
	mehunuu me ho wɔ wɔɔd kɛseɛ a dikɔnfoɔ pii wɔ mu.
	Sɛ ɔhwɛfoɔ no,
	wɔde ma firi ɔdɔ mu.
	Ɔhwɛfoɔ no maa kwan,
	Sɛ mɛdwene kwan a mɛfa so anaa mɛyɛ no pɛpɛɛpɛ berɛ a merekyɛ adidi kronkron no,
	Mebɔɔ mpaeɛ sɛ ɛnam me dɔ som no so nnipa no bɛte Awurade dɔ no nka.
	awieɛ no yɛ adekorɔ.
	na wɔyɛ wɔn frɛ no kɛseɛ no,
	Na na menim nhyira no:
	na ɔwɔ abisadeɛ.
	Na asɔre no nyinaa hyia wɔ yɛn fie.
	kɔyɛ Awurade adidi kronkron ma no wɔ ne fie.
	Mennim sɛ onuabarima somfoɔ yi abɔ mpaeɛ,
	Megye di sɛ yɛbɛtumi ama yɛn asɔfodie adwuma som no ayɛ kɛseɛ yɛn nkwa nna nyinaa mu na mpo atra akɔ akyiri.
	Ɛyɛ den sɛ Awurade,
	wɔ Ne din mu.
	sɛdeɛ wɔn a ɔsom wɔn no bɛhunu Awurade dɔ no,
	Mehyiraa no sɛ ɔnya ayaresa.
	te sɛ me deɛ no wɔ ahwɛbeaeɛ hɔ no,
	Asi pii berɛ a mama obi a ɔda owuo mpa so na abusuafoɔ atwa ne ho ahyia,
	wɔsomaa me kɔsraa adidi kronkron nhyiamu wɔ ahwɛbeaɛ bi.
	ayɔnkofoɔ asomfoɔ wɔ Onyankopɔn asɔfodie mu,
	mehunuu sɛ dikɔnfoɔ no de pɛpɛyɛ na ɛnante kyekyɛ adidi kronkron no te sɛ deɛ wɔahyɛda atete wɔn.
	Ɛyɛ me anisɔ saa berɛ yi sɛ saa da no,
	meda wo ase.
	ɔgyee penee,
	Saa nsunsuansoɔ korɔ no ara ba berɛ a mɛbɔ mpaeɛ ansa na m’ahyira obi a ɔyare anaa ɔwɔ ahohia mu.
	na wɔdii banbɔ nhyehyɛɛ so yɛɛ adidi kronkron som no.
	Ɛda adi pefee sɛ,
	nhyira no firi Awurade hɔ—kɔkɔbɔ ne bɔhyɛ no a wɔkyɛ wɔ Ne din mu.
	Nokorɛ mu,
	ɛnam ahwɛyie dodoɔ nti,
	Ɛbɛtumi ayɛ bɔkɔɔ,
	na ɔgye too mu.
	sɛ ɔsɔfo panyin yi no,
	n’abusua,
	Mmom seesei mahunu sɛ kwan pa wɔ hɔ a yɛdebɔ mpaeɛ na yɛdwene wɔ berɛ a yɛrenyini wɔ asɔfodie som mu.
	Ɛde mmɔdemmɔ na yɛde bɛhunu deɛ Awurade pɛ afiri deɛ yɛpɛ ne deɛ obi pɛ.
	“Na obiara nso a ɔgye saa asɔfodie yi no gye me,
	Berɛ a ɔfrɛɛ no sɛ ɔde adidi kronkron no bɛbrɛ no,
	Awurade asomfoɔ firi soro baae de asɔfodie no ba maa abasɛm akɛseɛ a abue na ɛda yɛn anim yi.
	N’asɔfodie som,
	Wɔ m’adidi kronkron nhyiamu a ɛdikan hɔ no,
	ɔsomfoɔ barima yi bisaa ne hwɛfoɔ nnansa yi sɛ obi foforɔ bi wɔ hɔ a ɔbɛtumi asra no anaa.
	Saa nkanyan korɔ no ba berɛ a agya panyinom kyene ɛkɔm na wɔbɔ mpaeɛ gye akwankyerɛ sɛ wɔde nhyira a Awurade no pɛ bɛma obi.
	Nanso mfeɛ okunafoɔ no,
	ne deɛ Onyankopɔn ayi wɔn.
	Berɛ a ɔkɔduruu ne fie Kwasiada anɔpa no,
	Asɔfodie kannifoɔ ma me ne menua no yɛ ɔhonufoɔ foforɔbi a afei na ɔno ankasa anya asɔfodie no.
	Wɔhyɛɛ me dikɔn wɔ nkorabata ketewa bi mu a na me nko ara na meyɛ dikɔn na menua Ted nso nko ara ne tikyani wɔ hɔ.

End of preview. Expand in Data Studio

Twi Speech-Text Parallel Dataset

Dataset Description

This dataset contains 21138 parallel speech-text pairs for Twi (Akan), a language spoken primarily in Ghana. The dataset consists of audio recordings paired with their corresponding text transcriptions, making it suitable for automatic speech recognition (ASR) and text-to-speech (TTS) tasks.

Dataset Summary

Language: Twi (Akan) - tw
Task: Speech Recognition, Text-to-Speech
Size: 21138 audio files > 1KB (small/corrupted files filtered out)
Format: WAV audio files with corresponding text files
Modalities: Audio + Text

Supported Tasks

Automatic Speech Recognition (ASR): Train models to convert Twi speech to text
Text-to-Speech (TTS): Use parallel data for TTS model development
Keyword Spotting: Identify specific Twi words in audio
Phonetic Analysis: Study Twi pronunciation patterns

Dataset Structure

Data Fields

audio: Audio file in WAV format
text: Corresponding text transcription from paired text file

Data Splits

The dataset contains a single training split with 21138 filtered audio-text pairs.

Dataset Creation

Source Data

The audio data has been sourced ethically from consenting contributors. To protect the privacy of the original authors and speakers, specific source information cannot be shared publicly.

Data Processing

Audio files and corresponding text files were collected from organized folder structure
Text content was read from separate .txt files with matching filenames
Files smaller than 1KB were filtered out to ensure audio quality
Empty text files were excluded from the dataset
Audio was processed using the MMS-300M-1130 Forced Aligner tool for alignment and quality assurance

Annotations

Text annotations are stored in separate text files with matching filenames to the audio files, representing the spoken content in each audio recording.

Considerations for Using the Data

Social Impact of Dataset

This dataset contributes to the preservation and digital representation of Twi, supporting:

Language technology development for underrepresented languages
Educational resources for Twi language learning
Cultural preservation through digital archives

Discussion of Biases

The dataset may reflect the pronunciation patterns and dialects of specific regions or speakers
Audio quality and recording conditions may vary across samples
The vocabulary is limited to the words present in the collected samples

Other Known Limitations

Limited vocabulary scope (word-level rather than sentence-level)
Potential audio quality variations
Regional dialect representation may be uneven

Additional Information

Licensing Information

This dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).

Citation Information

If you use this dataset in your research, please cite:

@dataset{twi_words_parallel_2025,
  title={Twi Words Speech-Text Parallel Dataset},
  year={2025},
  publisher={Hugging Face},
  howpublished={\url{https://huggingface.co/datasets/[your-username]/twi-words-speech-text-parallel}}
}

Acknowledgments

Audio processing and alignment performed using MMS-300M-1130 Forced Aligner
Thanks to all contributors who provided audio samples while maintaining privacy protection

Contact

For questions or concerns about this dataset, please open an issue in the dataset repository.

Downloads last month: 100

Size of downloaded dataset files:

6.55 GB

Size of the auto-converted Parquet files:

6.55 GB

Number of rows:

21,138

Collection including michsethowusu/twi-speech-text-parallel-multispeaker

African Speech-Text Datasets

Collection

Collection of speech to text datasets for African languages • 11 items • Updated about 7 hours ago