Sentence examples for acoustic training from inspiring English sources

Table 6 describes some attributes of the acoustic training corpus.

This resulted in approximately 104.5 h of acoustic training material.

The acoustic training corpus was then recorded in a clean environment to minimize external noise.

A biphonetically balanced (PB) Icelandic text corpus was used to create an acoustic training corpus.

Merging the lexicons from both CNC-S and acoustic training set transcripts and consequently interpolating corresponding language models yielded WER improvement of 2% absolute [16].

The first one was estimated using only the acoustic training set transcripts, and the second was trained from the selection of the Czech National Corpus (CNC).

This penalty was estimated from the data used for acoustic model training as follows: Training data are given to the HTK tool [50] to train HMM monophone models.

We present system design methodologies to address the problem of acoustic-training data sparsity encountered when using longer length units such as syllables.

In general, in acoustic model training, a short pause model is trained to aid in word or syllable boundary determination and thus to improve the recognition accuracy.

Table 6 Results for the evaluation set using an acoustic model trained with extended training data   Proc.

During acoustic model training, HMM states that are observed in the training data are clustered using decision trees to avoid overfitting.