While I’m less and less allergic to float where performance is concerned, it does and always will have a significant precision problem since float is inherently ‘lossy’. When precision and performance matter with decimal values, there’s still no substitute for fixed point math. Using Fixed-Point Math for Embedded Applications by Steve Hageman has a lot of excellent advice on getting the most out of your MCU for the least amount of work, and contains some generally excellent tips for signal capture and manipulation.Related
No comments yet.
Sorry, the comment form is closed at this time.