Figure 1From: Discriminative likelihood score weighting based on acoustic-phonetic classification for speaker identificationComparison of speaker identification results for noisy speech signals degraded by Aurora car noise with 20-dB SNR. All horizontal axes represent time in 10-ms- long frame. (a) Speech signal waveform (a TIMIT sentence utterance spoken as ‘How much allowance do you get?’). (b) Acoustic-phonetic class ID sequence. (c) Weight contours. (d) Unweighted scores. (e) Weighted scores.Back to article page