imtheKingCat
Newbie
Hello everyone,
I am currently working on a sound source localization project, and my initial plan was to implement the MUSIC (Multiple Signal Classification) algorithm on the STM32H747i Disco development board with a 4-microphone linear/circular array. However, after some research, I’ve found that many practitioners tend to prefer algorithms like beamforming or GCC-PHAT, as they are computationally less expensive and more suitable for real-time applications on microcontrollers.
On the other hand, MUSIC is computationally intensive, and it’s unclear whether it can meet real-time constraints on a microcontroller like the STM32H747. Since I’m still quite inexperienced in this field, I would like to ask for advice from those with more experience:
I am currently working on a sound source localization project, and my initial plan was to implement the MUSIC (Multiple Signal Classification) algorithm on the STM32H747i Disco development board with a 4-microphone linear/circular array. However, after some research, I’ve found that many practitioners tend to prefer algorithms like beamforming or GCC-PHAT, as they are computationally less expensive and more suitable for real-time applications on microcontrollers.
On the other hand, MUSIC is computationally intensive, and it’s unclear whether it can meet real-time constraints on a microcontroller like the STM32H747. Since I’m still quite inexperienced in this field, I would like to ask for advice from those with more experience:
- Is MUSIC commonly used in sound source localization applications? If so, in which contexts or industries?
- If I focus my thesis on optimizing MUSIC for embedded systems, specifically the STM32H747, what aspects of the algorithm should I prioritize to make real-time operation feasible? I’m aware that this is a broad and challenging task, and I’m trying to avoid getting overwhelmed.
