I've got a similar issue visualising faces. Recognition is ok most of the time but it often feels like I'm using other cues more to differentiate between people

I've got a similar issue visualising faces. Recognition is ok most of the time but it often feels like I'm using other cues more to differentiate between people. I can visualise things but like you it's hazy and never fully formed. I've often wondered if this might be affecting my dev work - people describe so much of what we do in terms of how we visualise things.