Can you link to where it is substantiated that the Nebulae v2 is using the Cortex A53?
If so that is impressive. Almost 11k in DMIPS.
Also while the CPU appears to be fast, it is still limited by its implementation as mentioned above regarding it’s sampling abilities and playback limits.
As the platform is open for custom algorithms if the CPU is for real, it should be able to run far more complicated versions of it, which should mean higher fidelity across those applications. The implementation, and no visual feedback seems to obscure this ability though. At least IMHO.