Datasets
Standard Dataset
iNoise Indian Noise Database
- Citation Author(s):
- Submitted by:
- Sunil Kumar Kop...
- Last updated:
- Mon, 04/06/2020 - 07:40
- DOI:
- 10.21227/w3xm-jn45
- Data Format:
- License:
- Categories:
- Keywords:
Abstract
Speech Processing in noisy condition allows researcher to build solutions that work in real world conditions. Environmental noise in Indian conditions are very different from typical noise seen in most western countries. This dataset is a collection of various noises, both indoor and outdoor ollected over a period of several months. The audio files are of the format RIFF (little-endian) data, WAVE audio, Microsoft PCM, 8 bit, mono 11025 Hz and have been recorded using the Dialogic CTI card. While the speech noise data was collected 40-50 seconds at a time, we have concatenated them to present a single noise file per category. Type of noises [a] Outdoor (Autorickshaw, Bus, Highway, Railway Station, Street) and [b] Indoor (Airport, Cafteria, Home, Train, Workplace)
The audio files are of the format RIFF (little-endian) data, WAVE audio, Microsoft PCM, 8 bit, mono 11025 Hz and have been recorded using the Dialogic CTI card.
Comments
Useful in building noise robust automatic speech recognition (ASR) based solutions, especially during training of acoustic models.
good