Use cases

Voice assistants
Text-to-speech
Speech recognition
Natural utterance collection
Speech emotion recognition
Audio classification
Audio transcription
Multilingual audio file collection
Side-by-side audio comparsion

Classify audio recordings for training and evaluating text-to-speech models and audio classifiers. Tag audio samples by emotion, category, or event. Use for voice assistants and other speech annotation projects.

Examples of tasks

  • Evaluate quality of synthesized speech
  • Classify emotions for speech emotion recognition
  • Identify recorded sounds
  • Detect speech activity

Convert audio to text with manual transcription or check for errors in automated transcriptions. Use the resulting data to improve speech recognition models for voice assistants or other tasks.

Gather any size dataset of custom speech samples. Our crowd can record audio files in over 40 languages to meet your exact specifications. Use for voice assistants and projects in multiple languages.

Examples of tasks

  • Record requests for a voice assistant
  • Read a text aloud with different emotions
  • Listen to a song and hum the melody
  • Record coughing, sneezing or snoring

Compare two audio files and identify which one sounds better or which one is more relevant.

Examples of tasks

  • Identify natural-sounding speech
  • Decide which generated speech sample sounds more authentic
  • Assess how closely voices match each other
  • Compare speech denoising quality

Success stories

10+ year industry experience solving real-world problems
Recording

Native Danish recording for BYD

  • Recorder Quantity: 300
  • Duration: 2 months
  • Accuracy: 99% sentences
Transcription

German Youtube Transcription

  • Data Quantity: 300 hours
  • Duration: 3 months
  • Accuracy: 98% words
Recording

Native French multi-roles recording

  • Recorder Quantity: 600
  • Duration: 4 months
  • Details: situmulate conversation meeting recording with 2 or 3 roles
Designed by engineers for engineers
SUBSCRIBE TO UPDATES