Lossless and lossy audio codecs for low-performance microcontrollers for use in IoT

Abstract

Internet of Things (IoT) is a very fast developing branch of IT. IoT needs low-power systems, that are often also low-performance. For such devices, there is a problem to use one of the popular audio codecs because of very high-performance demand, for example floating point operations. In this paper, two codecs: lossless and lossy are proposed. Lossless codec makes use of the simple bit operations for reducing the number of bits needed to store one sample and therefore lead to reduce the size of the whole audio file. All needed arithmetic operations are executed very fast and thus don’t need a high performance of the microcontroller. Lossy codec additionally utilizes some principles borrowed from the decision systems, among them from rough-set theory. In this case, the resulting audio is lossy, but the difference is very small and thus could not be heard. Experiments show that either lossless, as well as lossy codec, reduces the size of the audio file, and are very fast (don’t need high-performance microcontrollers), so can be used in IoT systems.