Hugging Face's logo

Datasets:
BoxlyX
/
English_Natural_Conversation_ASR_STT

Tasks:
Automatic Speech Recognition
Text-to-Speech
Modalities:
Audio
Text
Formats:
soundfolder
Languages:
English
Size:
< 1K
Tags:
audio
conversational-ai
speech-processing
asr
Libraries:
Datasets
License:
Dataset card Data Studio Files
xet
 Community
1
English_Natural_Conversation_ASR_STT / README.md
shadikurbd's picture
shadikurbd
Update README.md
2856fa9 verified
preview code
|
Raw
History Blame Contribute Delete
2.93 kB
metadata 
license: mit
task_categories:
  - automatic-speech-recognition
  - text-to-speech
language:
  - en
tags:
  - audio
  - conversational-ai
  - speech-processing
  - asr
size_categories:
  - 1K<n<10K
pretty_name: BoxlyX_English_Natural_Conversation_ASR_STT

BoxlyX English Natural Conversation Sample Dataset (ASR/STT)

📌 Overview

This repository contains high-fidelity, studio-recorded English natural conversation samples designed for training and benchmarking advanced Automatic Speech Recognition (ASR) and Speech-to-Text (STT) models.

This dataset is a curated public sample provided by BoxlyX AI Solution, showcasing our end-to-end capabilities in premium audio data generation, multi-speaker recording environment orchestration, and meticulous human-in-the-loop verification.

🛠️ Dataset Specifications

  • Total Duration: ~1 Hour (4 files × 15 minutes each)
  • Audio Format: Lossless WAV, 48kHz, 16-bit, Stereo/Mono uncompressed
  • Content Type: Natural, unscripted, spontaneous human conversations covering various engaging prompts
  • Transcription Quality: 100% human-verified verbatim transcripts (including structural filler words, overlapping speech, and natural pauses)

📂 Repository Structure 

English_Natural_Conversation_ASR_STT/
├── README.md         # This dataset profile
├── metadata.csv      # File mappings, metrics, and exact transcriptions
├── Whats_something_you_used_RC-514568.wav
├── Whats_something_youve_accidentally_RC-789752.wav
├── Whats_the_best_accidental_RC-918844.wav
└── Whats_the_funniest_thing_RC-550548.wav

🚀 About BoxlyX AI Solution

Need large-scale, domain-specific speech datasets tailored to your specific AI models?

BoxlyX is a premier AI data engineering partner specializing in constructing end-to-end custom dataset pipelines with enterprise-grade quality controls:

  • 🌍 Global Multi-Lingual Capabilities: We collect, transcribe, and label high-fidelity data in any language, accent, or dialect across the globe, tailored strictly to your target demographic matrix.
  • 🎙️ Custom Audio Sourcing: Multi-speaker conversational data, telephony acoustic profiles, command-and-control utterances, and specialized domain-specific vocabulary.
  • 🔍 Meticulous Annotation Pipelines: Human-in-the-loop expert validation, precise timestamp alignment, speaker diarization, and emotion/intent labeling.
  • 📈 Scalable Data Fulfillment: Rapid scaling capabilities capable of delivering thousands of hours of fully verified, clean dataset deliverables.

📧 Contact Our Sales Engineering Team

Let's build the foundation for your next state-of-the-art model. Reach out to discuss your specific linguistic, volume, and formatting requirements: