Peer-reviewed research
Automatic Recognition of Wideband Telephone Speech with Limited Amount of Matched Training Data
Patrick Bauer et al., EUSIPCO 2014, European Signal Processing Conference, 2014
Read the originalRead on October 16, 2026
- What it is
- A peer-reviewed conference paper on recognizing wideband (16 kHz) telephone speech. Its introduction states that dictation and other large-vocabulary applications use a sampling rate of at least 16 kHz, especially in noisy environments, and reports a 14% relative gain for noisy speech at 16 kHz in the Aurora tests.
- How much weight it carries
- Peer-reviewed. The 14% and the 20% figures are reported from earlier studies that the authors cite, so the paper is a secondary source for them. The main experiments are on telephone speech, not microphones on a desk.
Facts taken from this source
- Bauer et al. state that dictation and other large-vocabulary applications principally use a sampling rate of at least 16 kHz, especially in noisy environments.
“a sampling rate of at least 16 kHz, especially in noisy environments.” (Introduction)
- The Aurora standardization tests, as cited by Bauer et al., found a 14% average relative gain in recognition of noisy speech sampled at 16 kHz instead of 8 kHz.
“ASR performance gain of 14 % averaged over a weak, medium, and high recording mismatch between training and test conditions, when recognizing noisy speech sampled at 16 instead of 8 kHz” (Introduction)
Pages that rest on this source
- Dictation and Voice Typing: How to Talk Instead of TypeTopic
- How Speech Recognition Works, From Sound to TextGuide
- How to Transcribe Audio to TextGuide
- Best Microphone for Dictation: What Actually MattersGuide
- Speech Recognition in Python: Libraries ComparedGuide
- Can you use a voice recorder for dictation?Question
- Do AirPods work for dictation?Question
- Is a wireless headset good for dictation?Question
- What is the best handheld device or recorder for doctor dictation?Question