The computer memory has been revolutionized in the last 25-30 years, in terms of both capacity and speed of execution. Along with this, even the logic controlling the memory has also become more and more complex and difficult to interface. Usually, memory subsystems will be designed to interact with a single system. Whenever we consider a two system is sharing a common memory, there comes the need for an Arbiter. The major difference between a memory arbiter and a processor scheduler is that the memory arbiter works at a much finer level of granularity. The time taken for the task execution may range from micro to milliseconds, while a RAM controller needs to serve the request in a few nanoseconds. Because of this reason the resource arbiters are usually designed and implemented in hardware rather than in software.

Multimedia applications play a very important role in the field of VLSI design and embedded systems. They need a large amount of memory storage with higher bandwidth and higher speed. To overcome this hazard, a memory controller is required. A memory controller is a device that stores the data and gives it back whenever required. Real-time recording of an audio data and finally storing it without losing the data is a difficult task. This paper describes the usage of Double Data Rate Synchronous Dynamic Random Access memory controller for storing the audio data. The design uses finite state machine (FSM) architecture that is developed for testing of this algorithm. Audio codec device is used for the conversion of analog data into digital and vice versa. The tool used to simulate this design is Xilinx ISE design suite. The hardware used to synthesize this design is FPGA Spartan-3 kit.