Next Article in Journal
Comprehensive Flood Risk Assessment for Wastewater Treatment Plants under Extreme Storm Events: A Case Study for New York City, United States
Next Article in Special Issue
The Multi-Domain International Search on Speech 2020 ALBAYZIN Evaluation: Overview, Systems, Results, Discussion and Post-Evaluation Analyses
Previous Article in Journal
Energy Expenditure during Acute Weight Training Exercises in Healthy Participants: A Preliminary Study
Article
Peer-Review Record

Automatic Speech Recognition (ASR) Systems Applied to Pronunciation Assessment of L2 Spanish for Japanese Speakers

Appl. Sci. 2021, 11(15), 6695; https://doi.org/10.3390/app11156695
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Appl. Sci. 2021, 11(15), 6695; https://doi.org/10.3390/app11156695
Received: 27 June 2021 / Revised: 15 July 2021 / Accepted: 19 July 2021 / Published: 21 July 2021

Round 1

Reviewer 1 Report

The paper entitled "Automatic Speech Recognition (ASR) Systems Applied to Pronunciation Assessment of L2 Spanish for Japanese Speakers"  compare the performance of
author's automatic speech recognition system with Google one.

The overall merit of the presented article is good, but i have some remarks/questions.

 

  1. The  automatic speech recognition system kASR is not described in the article at all
  2. The novelty of used ASR should be explained more
  3. The comparision with other known ASR should be performed

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors propose an approach for automatic speech recognition (ASR) systems applied to pronunciation assessment of L2 spanish for japanese speakers. 

The paper is well written and structured.

 Here are some comments regarding the paper:-

  1. in line 45, the authors cite the Kaldi ASR Toolkit which provides two speech recognition models, one based on GMM-HMM and the other based on a neural network. my question to the author, why did they prefer to use the GMM-HMM Kaldi model rather than the neural network-based one? .. knowing that the performance of neural network models surpasses the performance of GMM-HMM
  2. Paragraphs from line 364 to 403 of section 3.2. are a confusing cause of inserting a lot of variables / numeric numbers into text. this makes it harder to understand the outcome discussions. I suggest that the authors try to simplify the presentation of the results. maybe they can divide the tables.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Back to TopTop