clarification of photon "arrows" in feynmans QED book

BW

Hi! In Feynmans QED, strange theory of light and matter, he spends the
first 60 pages by pedagogically describing the in-flight turning
arrows and the resulting interference phenomena which result after
path summation.

Then on p.60 I think it was, he mentions almost in passing that for
the photon, there is no turning of the amplitude in flight, but
instead the reason we get interference effects is that the source
amplitude in the light-source (a monochromatic source in his example
there) varies with time - so there really is no turning of the arrows
for a photon but you have to consider many photon paths coming from
the same source at different times instead to keep the effect.

Leaving the "real" QED math aside, which obviously works fine, is this
all there is to this issue ? What about for example red-shift due to
space-time expansion in cosmology. If the photons arrows do not turn
in flight, what is the "Feynman" arrow description of such an event ?
How do different photons have different energies at absorption if
there is no oscillation or turning at the fundamental level - am I
missing something for example that the emitted photon needs to be in a
temporal superposition of many amplitudes and times and the collective
superposition leads to the conventional effects of individual photons'
"wavelengths" ?

Sorry if this sounds messy, but I'm interested in Feynmans simple
discussions on QED as a complement to "just calculating". I also have
a hunch that the resolution to this lies in the difference between
virtual and real photons..

/Bjorn