Each DSP code iteration is a mod and tested using a dedicated virtual machine. They are organized in rigs and tested against other devices MAC-tables. Each Compositor mod sustain a relatively big uptime before it is applied in real-time processors. That is why each dump is successful.
Compositor NRTOS is an embedded system, which has different levels of execution. The highest level is GenEXPR DSP code and the lowest is Max scheduler. By combining two worlds, I attain to an ideal cabinet for virtual machines testing and evaluation.
Compositor uses telnet to communicate with other work groups on development of NRTOS. The development of virtual machine mods is finished and now the stage of compare against other devices in a form of MAC-tables happen. That is why the telnet functionality is constantly updated and new language patterns evolve.
Each Compositor DSP is based on a FM microarchitecture, which have evolved over time in a complete microkernel. That is why I consider MaxMSP replica NRTOS on its own. In fact, Compositor NRTOS has 8 learnable microkernels, which correspond to different versions of GenEXPR code.
Compositor NRTOS uses zero-buffer mod, which circumvents the device to use feedback loop instead of static buffer. This makes it theoretically undefeatable.
Compositor NRTOS has inplace logs for each learning process of DSP code. There are 7 DSP’s and 7 logs in total.
Compositor NRTOS has a recognizable graphic user interface. So, when you are talking about Compositor it is also an interface, which makes DSP work in a graphical way.
Each Compositor mod is a DSP code and each one is represented as the MaxMSP replica. Besides, for a practical purpose I compile DSP code as an Android app with a flashscreen only. The main purpose of such app is to dump one of the hardware caches of a chipset.
29-04-2015
The quantized radio technology for ULF transmission
1024 samples
256PPQ
16th grid