Q2. What are the future works in this paper?
One natural goal is to enlarge the user community by extending the implemented services to as many languages as possible. Such an approach is especially necessary because the authors plan to introduce more signal processing web services in the coming years ( see below ). The authors plan to investigate the best cloud storage possibilities and implement APIs for these. The authors plan to implement a new automatic chunk segmentation web service based on the results reported in Pörner ( 2016 ) that allows the user to first chunk segment very long recordings and then process the chunks in MAUS batch mode.
Q3. What are the two tasks that occurred frequently?
Two tasks occurred frequently: extending the web interface to incorporate new services and changing/adding parameters that control the back end behaviour.
Q4. What is the advantage of web services?
One advantage of web services is the possibility of easily combining services into more complex processing constructs or processing chains.
Q5. What are the uses of G2P converters?
G2P converters are needed for speech synthesis, for aiding manual and automatic transcription of spoken text and for the generation of pronunciation dictionaries based on text collections (to name but a few possible uses).
Q6. What is the cost function for the alignment of two symbolic sequences?
The aligner considers the alignment of two symbolic sequences v and w as a task to transform v into w by a minimum sum of edit costs, which is known as the Levenshtein distance (Levenshtein, 1966).
Q7. What is the common framework that supports the definition of building blocks?
One popular framework that supports the definition of such building blocks is AngularJS ix which follows a modified Modal View Controller (MVC) pattern.
Q8. What are the two metadata descriptions used for the BAS web services?
The two metadata descriptions, WADL and CMD instance, allow for automatic service invocation by other applications and the automatic generation of documentation.
Q9. What is the morphological analysis of the stress learning-by-analogy?
For German and English, the stress learning-by-analogy is preceded by a compound decomposition step based on the morphological analyses (described in Reichel 2012a).
Q10. What are the main techniques used to modify the synthesized speech?
Several standard techniques to modify the synthesized speech, e.g. 'vocal tract scaling', 'fundamental frequency scaling', 'chorus effect' etc., are available.
Q11. What is the metrical tree induced by the stress learning-by-analogy?
From this decomposition, a metrical tree is induced based on the relative coherence of neighbouring compound parts (Reichel 2012b).