As a generic rule, for an object in your render, the amount of pixels you see on the render requires at least twice that amount on the object.

So, let's take a head taking 2/3 of a portrait render of 1500x1500. So the diameter of that head equals 2/3 x 1500 = 1000 pixels, hence the circumference of that head equals pi x 1000 = 3500 pixels (rounded) and according to the rule mentioned above this has to be doubled: 7000 pixels of the texture map should cover the head.

In a full body render however, the same head will only take say 1/10 of that render. So the head will require a 2x pi x 1/10 x 1500 =  1000 pixel wide map.

For short: the 3k or even 4k texture maps that come with the characters nowadays are (more than) fine for close ups shown on screen, but will fall short for high quality large scale prints (A4 size requires a 2000x3000 render => say a 10.000 pixel wide texture map for a facial close-up).

In the examples shown above, the head has a 500 pixel diameter and would therefor require a 2 x pi x 500 = say 3500 pixel wide map for good quality. Half of that ignores the rule-of-2 and would result in modest quality, and the 1k wide map actually used is even half of that. The result says it all. But show the character full boy in context, and the 1k map will do fine.

Edit: plus - following vilters: the 1095 should be shrunk to 1024. 

- - - - - 

Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.

visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though