Standardized Speech Recognition API via HTTP
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech recognition applications are resource-intensive and costly to develop, especially for smaller devices, and often require high skill to integrate with network-hosted services, which may not work across different platforms or device boundaries.
Innovation Solution
A standard interface for speech recognition operations is provided using a hypertext protocol (HTTP) based system that allows clients to request and receive speech recognition services through a scalable, hosted platform with a simple API, enabling easy integration and compatibility across various devices and platforms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If speech recognition applications are developed locally on devices, then speech recognition functionality is available, but resource consumption and development cost increase significantly
Solution Approach 1:
The patent extracts the speech recognition engine from the client device and relocates it to a network-hosted server. The client application only retains the functionality to capture audio and communicate with the server, while the resource-intensive speech recognition processing is performed remotely. This extraction resolves the contradiction by maintaining speech recognition functionality while dramatically reducing local resource consumption.
Solution Approach 2:
The patent introduces a network-hosted speech recognition service as an intermediary between the client device and the speech recognition engine. The client communicates with this intermediary service, which in turn communicates with the speech recognition engine on the server. This intermediary architecture allows the client to access speech recognition capabilities without hosting the engine locally, reducing resource consumption while maintaining functionality.
2Use of energy by moving object
If network-hosted speech recognition services are used, then resource consumption is reduced, but integration complexity and skill requirements increase
Solution Approach 1:
The patent creates a universal API interface that can be used across different client applications and platforms to access the network-hosted speech recognition service. This standardized interface allows multiple applications to interact with the speech recognition engine in a consistent manner, reducing integration complexity and making the service accessible to developers with varying skill levels.
Solution Approach 2:
The patent modifies the interface parameters and communication protocols to simplify the integration process. By standardizing the API parameters, data formats, and communication methods, the patent reduces the complexity of integrating network-hosted speech recognition services while maintaining reduced resource consumption benefits.
3Reliability
If proprietary speech recognition APIs are used, then speech recognition functionality is achieved, but compatibility across different platforms and devices is limited
Solution Approach 1:
The patent designs a universal API interface that works across multiple platforms and devices, including mobile phones, personal computers, and other computing devices. This universal interface allows the same speech recognition service to be accessed from different platforms without requiring platform-specific implementations, thereby improving adaptability while maintaining functionality.
Solution Approach 2:
The patent segments the speech recognition system into distinct components: the client application layer that handles audio capture and API calls, the network communication layer, and the server-side speech recognition engine. This segmentation allows each component to be optimized independently and facilitates cross-platform compatibility by standardizing the interfaces between components.
Data Source
AI summary
Techniques and systems to provide speech recognition services over a network using a standard interface are described. In an embodiment, a technique includes accepting a speech recognition request that includes at least audio input, via an application program interface (API). The speech recognition request may also include additional parameters. The technique further includes performing speech recognition on the audio according to the request and any specified parameters; and returning a speech recognition result as a hypertext protocol (HTTP) response. Other embodiments are described and claimed.