Objective Evaluation Methods for Chinese Text-To-Speech Systems

Author(s):  
Teng Zhang ◽  
Zhipeng Chen ◽  
Ji Wu ◽  
Sam Lai ◽  
Wenhui Lei ◽  
...  
2005 ◽  
Vol 24 (2) ◽  
pp. 81-95 ◽  
Author(s):  
Jishnu K. Subedi ◽  
Hiroki Yamaguchi ◽  
Yasunao Matsumoto ◽  
Mitsutaka Ishihara

Annoyance of low frequency pure and combined tones was measured in a laboratory experiment. Three low frequency tones at frequencies of 31.5, 50 and 80 Hz at four sound pressure levels, from about 6 dB to 24 dB above average hearing threshold, were selected as pure tones. The combined tones were combinations of two tones: the four levels of 31.5, 50 and 80 Hz tones and a constant level 40 Hz tone. The results showed that the rate of increase in annoyance of pure tones with increase in the sound pressure level was higher at lower frequencies, as reported in previous studies. The results for the combined tones showed that the increase in the annoyance of the combined tone compared to the annoyance of pure tone was dependent on the level difference of the two tones and their frequency separation. These results were compared with the evaluation obtained from different objective methods. The three methods were Moore's loudness model, the low frequency A-weighting and the total energy summation used as objective evaluation methods. Among the methods, the low frequency A-weighting gave the best correlation.


2020 ◽  
Author(s):  
Jungirl Seok ◽  
Youn Mi Ryu ◽  
Seong Ae Jo ◽  
Chang Yoon Lee ◽  
Yuh‐Seog Jung ◽  
...  

2013 ◽  
Vol 22 (01) ◽  
pp. 34-46 ◽  
Author(s):  
E. Ammenwerth ◽  
M.-C. Beuscart-Zephir ◽  
J. Brender ◽  
H. Hyppönen ◽  
S. Melia ◽  
...  

Summary Objectives: To present the importance of Evidence-based Health Informatics (EBHI) and the ethical imperative of this approach; to highlight the work of the IMIA Working Group on Technology Assessment and Quality Improvement and the EFMI Working Group on Assessment of Health Information Systems; and to introduce the further important evaluation and evidence aspects being addressed. Methods: Reviews of IMIA, EFMA and other initiatives, together with literature reviews on evaluation methods and on published systematic reviews. Results: Presentation of the rationale for the health informatics domain to adopt a scientific approach by assessing impact, avoiding harm, and empirically demonstrating benefit and best use; reporting of the origins and rationale of the IMIA- and EQUATOR-endorsed Statement on Reporting of Evaluation Studies in Health Informatics (STARE-HI) and of the IMIA WG's Guideline for Good Evaluation Practice in Health Informatics (GEP-HI); presentation of other initiatives for objective evaluation; and outlining of further work in hand on usability and indicators; together with the case for development of relevant evaluation methods in newer applications such as telemedicine. The focus is on scientific evaluation as a reliable source of evidence, and on structured presentation of results to enable easy retrieval of evidence. Conclusions: EBHI is feasible, necessary for efficiency and safety, and ethically essential. Given the significant impact of health informatics on health systems, care delivery and personal health, it is vital that cultures change to insist on evidence-based policies and investment, and that emergent global moves for this are supported.


1998 ◽  
Vol 12 (4) ◽  
pp. 393-410 ◽  
Author(s):  
F Yvon ◽  
P Boula de Mareüil ◽  
C d»Alessandro ◽  
V Aubergé ◽  
M Bagein ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document