Audio Data Collection and Recording Job for OTS-TTS Wake Words (En-US)

Description: - We are seeking experienced professionals to assist us in collecting and recording audio data for the development of an OTS-TTS. The primary objective of this project is to create a high-quality dataset of wake words in the English (US) language. Requirements: -Professional Studio Recording: The recordings must be conducted in a professional studio setup using high-quality microphones and audio equipment. This will ensure optimal signal-to-noise ratio (SNR) and clear voice recordings. -Sentence Variety: The dataset should include a diverse range of both general and domain-specific sentences. This will enable our system to handle various scenarios and vocabulary more effectively. -High Signal-to-Noise Ratio: All recordings must have a signal-to-noise ratio greater than 35dB. This will guarantee clear voices and minimal background noise interference. -Sampling Rate: Audio should be recorded at a higher sampling rate, and no upsampling from lower rates should be performed. This will preserve the audio quality and prevent artifacts or degradation. -Performance Consistency: Actors should maintain performance consistency throughout the recordings. This includes persona and style continuity, root pitch or tone, rate of speech, volume levels, and proper pause/break placement on punctuations and between sentences. The goal is to achieve a natural and cohesive dataset. -Normalization: Audio normalization should be applied consistently across all datasets. The silent portions at the beginning and end of each recording should be approximately 100ms to maintain consistent audio length. -Male and Female Voice Actors: We require one male and one female voice actor to provide diversity in the dataset. Both actors should meet the above requirements and deliver high-quality recordings.