Open kallaballa opened 4 years ago
One way we could go about this is to have a representative benchmark that at least roughly gives us an idea of what the hardware is capable of. In detail we could identify the different kind of loads a synthesizer model imposes on systems and create a benchmark for each of those. e.g.: oscillator performance, filter performance, arithmetical operations on signals... That way we could generate a profile that makes platforms comparable.
What are the other was of achieving this? Should we instead of the above approach simply use a generic benchmark library?
We'd like to record for each patch the hardware it is designed to run on. We'd like the user to be able to
2. Manually enter a hardware specifierFor the third option we need to find a cross platform way of generating a meaningful hardware identifier. A hardware identifier should reflect the performance of the platform (and therefore the number of voices that it can generate simultaneously with the current patch)