In the absence of a text transformation matrix (CGContextSetTextMatrix
), 1 em will the be the same number of points as the font size. So for a 12 point font with the default identity text matrix, 1 em = 12 pt.
Note that almost nothing in Core Graphics is given in pixels. They're in points (Core Graphics points, not PostScript points). For a retina display, 1 point contains 4 pixels (2x2). This is probably what you mean anyway, but if you interact with images, the point/pixel conversion is very important.
(I kind of feel like I've repeated what you already know; is there more to the question? From a given origin, you would compute the coordinate by multiplying the em unit offset by the font size.)