Design And Implementation Of The Client Sdk For Open Platform Of Voice Cloud Based On Ios

With the smart phones, tablet computers and other smart mobile terminals being highly popular today, the mobile internet has a rapid development as well. At the same time, higher requirements of text input are needed desperately by mobile terminal application. What’s more, navigation, chat etc. applications are expecting that they can free users’hand for text input through voice recognition technology. With the maturing of Siri platform on iOS devices, lots of Internet companies have launched their own speech recognition systems to public. But for now, most of the Internet companies have strict limits on speech recognition SDK, iOS system fails to provide a common Siri API for developers to invoke.This article mainly studies the available opening speech recognition SDK presenting in the iOS system, contracts the product features of each voice recognition SDK. After analyzing developer demand for speech recognition SDK, this article puts forward a set of solutions to renew the client speech recognition SDK, which is called the client SDK for Open platform of voice cloud based on iOS, voice cloud SDK in short. Voice cloud SDK makes developers easily build applications with speech recognition functions and strong interaction on iOS devices. In the process of development and using, the developers need not maintain the speech recognition engine but can enjoy the speech recognition service.In this article, We gradually implement the voice cloud SDK based on the software engineering method and following the software development process. First of all, we acquaintance the basic process of voice recognition client. Combining with the users’habits to the speech recognition, we put forward the requirements of the voice cloud SDK. Requirement analysis mainly lists the function provided by the voice cloud SDK and the function need to be implemented during the process of interacting of server and voice cloud. After the detailed requirement analysis, this article has carried on the design to the voice cloud SDK. The whole voice cloud SDK is divided into several main modules according to the functions, namely:recording module, voice activity detection module, compression encoding module, network module and results callback module, and then lists the parameters and the method of each module, finally gives the chart of work flow and interactive relation between each module. Then, according to the design, we code to achieve the SDK, the coding process is implemented By voice data workflow. At last, this article carries on a systematic software testing to detect the voice cloud SDK. According to the results of software testing, we improve the availability and security of the entire voice cloud SDK.