"Fakeness" may be a judgment, but it most certainly is not based ONLY on the Fi type's own feelings. In regards to FPs, they also use Pe, and that is real world data.
The FP DOES NOT reason in this way: If I cried in this situation it would be fake, so this person's tears must be fake & manipulative.
The FP MAY reason in this way: This person doesn't appear to be crying naturally. The sounds seem forced & I don't actually see tears. From my understanding of people, when they cry naturally they usually display ____, and I don't see that here. I suspect this person is being fake in their emotional display.
An ISFP may conclude someone is fake from very real sensory information, such as facial expression & body language.
An INFP may conclude that someone is fake from intangible vibes they get & patterns about people they note, often which can be traced back to concrete info, although they don't think of it in those terms.
Of course, they may still conclude wrongly, but this is NOT entirely projection. There is a perception of external reality to lead to these conclusions. Note that the conclusion is often suspended somewhat still, with the FP not committing to a sure conclusion (they only suspect). This, IMO, is closer to how an FP thinks.
Can we talk Fe & disbelief in feelings without emotional displays? Anyone want to explain that unfair assumption?