Is Feynman gauge reduce always physical gauge?
I heard in QCD, Feynman gauge does not always give correct physics.
The lecture says, "Feynman gauge gives physical gauge, if the theory contains only conserved current." Thus in QED Feynman gauge gives correct physics. (It is a physical gauge) But in QCD, If we neglect the ghost term in the Feynman rule, and compute its scattering amplitude via Feynman gauge, the answer may wrong.
(But also Lecture says, if we include the ghost term in the gluon loop, and apply Feynmann rule then it gives the correct answer. )
I always think in QCD and non-abelian gauge theory contains ghost term (due to its path integral measure). Thus I am confused about computing loop integrals without ghost term. If you know the intention of lecture, please give me some detail explanation.
Comments or example of computing QCD loops without ghost term which gives physical meaning are also welcome.