rtfss: a hardware description language for digital audio processing

• Autor(a) principal: Pinho, André da Silva
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/31347

The pursuit of audio synthesis in digital circuitry and computing systems dates back to the time where programming languages such as Lisp, COBOL and Fortran were still being conceptualized. This era (the 1950s) marks the appearance of the MUSIC I programming language. Since then, and as the processing capabilities of computers kept rising, there was an influx of new languages capable of synthesizing audio. Languages such as SuperCollider, Pd (pure data), Max/MSP, ChucK and Faust are among the most popular ones, nowadays. These languages, although having different approaches and designs, have one aspect in common: they are all computer languages. Albeit these solutions grant great flexibility, they present an issue: timing constraints, particularly audio latency. In this document, a new Audio Processing Language, called rtfss, is introduced. This language was designed to be compiled mainly to hardware targets, exploiting their inherent parallelism. To complement the formal description of the language, a compiler that implements a subset of the language's capabilities will also be presented. This compiler is capable of analysing a design description written in rtfss and compile it to the lower level hardware description language VHDL. One of the main highlights of rtfss (and its compiler) is that the hardware architecture it synthesizes is pipelined. This type of digital architecture is used in logic designs to maximize the working frequency and the efficiency of the design. Throughout this document, both the language's formal specification and the compiler's implementation will be thoroughly analysed and explained. Some use-case examples of this language, accompanied by their compilation procedure, will also be illustrated and examined. In this document, it is also described a simpler architecture of another Hardware Audio Processing Language, which served as a stepping stone towards the development of rtfss. As a product of its development, some useful VHDL Audio (and MIDI 1.0) related hardware blocks were implemented.