I think the answer has to be founded in research for the particular subject you're building.
There's no generalized, one-sized fits all answer, and rationalizations can only be applied once you narrow the subject down to some particular airplane, AFV, or ship.
Once you have a particular subject in mind, if, after thorough research, you can't determine just "how hard" the edges were on any camouflage pattern present on that subject, then speculation and rationalization about how that particular camouflage might have been applied is necessary. Comparison and contrast to other similar subjects at the same time in the same unit, adjacent units, the same theater of operation can then give you clues and help to find commonalities and trends that might reasonably be applied to your particular subject to fill the gaps in the research.
But without a particular subject in mind, the possibilities are just too broad and potentially varied to come up with valid generalizations.
Using color plates in books of artwork done by others assumes that the artist did the research and got all of his assumptions right. There's a risk with color plates that you might not know how old the original artwork was, and the state of knowledge constantly improves and advances. What were generally acceptable assumptions when the art was done may have been over come by newer information. Also, the quality of the color reproduction and other aspects of the art in whatever source it's been produced in must be assessed. Is the edge of the camouflage soft because that's the way it was actually applied to the prototype, or is the production of the artwork simply "fuzzy"?
I would submit that if you examine the outside edges of the aircraft artwork profiles, those edges are just as soft and fuzzy as the edges of the camouflage pattern. This suggests to me that the pattern appears slightly softer than the model photos because of the way the artwork was reproduced and not because of how the paint was applied to the prototype.
Another example of the risks in using artwork or dated references: We now know that the famous German "ambush" camouflage patterns on late war armor were painted at the factory. Many older references (ex the Panzer Colors series) go to great lengths to describe how this was devised and produced in the field. We know this to be generally wrong (the rarity of the field applied examples being part of that proof). In fact, it's now possible to actually identify variations in the "ambush" schemes that were factory specific (in the case of the Panther tank). The state of knowledge has advanced, but the old artwork and incorrect references are still out there.
Some hard edged camouflage patterns are clearly brush painted with tight, clean, sharp edges. Some patterns are painted with ragged brushed edges. Some patterns are painted using stencils with either tight, sharp edges or slightly softer edges as if the stencils were not tight to the surface. Still other patterns were painted with the spray gun held close to the surface producing soft edges with little over spray and others painted with the spray gun held some distance away producing very soft, translucent edges.
Some camouflage patterns were applied at factories, some applied at well-established and equipped depots, and some applied at field-level operations. Some camouflage was applied by trained and experienced factory workers, some by trained military depot maintenance personnel, and other camouflage was applied by crew members just trying to get the job done in between all the other things that had to worry about.
Only good research of your particular subject can let you figure out how the camouflage pattern should look.
