[MatjazO]

My understanding is that OSPDAF systems started as hybrid implementation. That is, they worked as combination of PDAF and CDAF. PDAF used for initial distance determination and therefore moving focus close to exact position. CDAF then doing final step(s). In that case, color information should be available. And focus should be accurate.

I wonder what has changed in latest implementations: Sony, Canon’s dual pixel as well as other have improved significantly. Do today (at least some) systems use PDAF only? Light sensitivity and accuracy were an issue of OSPDAF, maybe less so nowadays. In that case, wavelenght could be the reason for missfocus.