MFCCs remain the dominant feature representation for speech recognition systems

Evidence

MFCCs have been found to characterize speech sounds efficiently, hence their frequent use in speech recognition algorithms, especially those employing statistical models for the actual identification. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1593

Categorization

• XREF: connects to: PLP coefficients describe speech spectral features compactly and speaker-independently — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1599; Speech phonemes reduce to 18 articulatory-acoustic features in a linear mapping — Dijk, Pim van, Physiology, Psychoacoustics and…, 4480

Broader Connections

• PLP coefficients describe speech spectral features compactly and speaker-independently #qualifies (0.75) — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1599
• Speech phonemes reduce to 18 articulatory-acoustic features in a linear mapping #qualifies (0.72) — Dijk, Pim van, Physiology, Psychoacoustics and…, 4480