Computing and Informatics

This paper presents DSP blocks which were developed to be used as basic elements for realization of the DSP algorithms. For this purpose the description DSL (Domain Specific Language) language was used. The goal of this paper is to define and present a high level language that allows description and development of signal processing algorithms. With the usage of a domain specific language, one can create a compact and easy to understand definition of algorithms. In the paper the authors present the advantages granted by DSL for DSP applications. The created definitions are hardware independent and they can be executed and functionally verified. Efficient code can be generated for various targets without porting. The design of the presented DSL allows code generation for multi-core targets in case of computing-intensive algorithms, code generation for multiple streams and threads. To validate the results these blocks were made available for use to students as an easy method for the introduction of the DSP algorithms in sound and image processing. The main purpose was for the students to gain some basic insight into elementary techniques needed for design, implementation and merging of hardware and software components used in testing of the algorithms for digital signal processing in real time. Through the work with the students it was concluded that the developed DSP blocks presented very good assistance in educational process and therefore this paper was elaborated on that idea. Since real hardware systems were used in this case noise was introduced in the system which does not exist in simulation software and therefore this option produced much larger capabilities for development of the robust algorithms.