Does anyone know of a true algorithm which mimics VFP's foxels to pixels conversion?

I found this algorithm online and modified it somewhat to adjust for the font attributes bold, underline, and italics. It ranges in accuracy (depending on the font and attributes) from being perfect, to being within a few pixels of being accurate.

I can't figure out what VFP is doing to convert foxels to pixels as it does. Does anyone know?

**********
* Converts VFP foxels to screen pixels
* tnFoxels        - fractional foxel value to convert
* tlVertical      - Is this a Y coordinate?
* tcFontName      - "Arial" for example
* tnFontSize      - 10
* tlFontBold      - Is the font bold?
* tlFontItalics   - Is the font in italics?
* tlFontUnderline - Is the font underlined?
*****
FUNCTION foxels_to_pixels
LPARAMETERS tnFoxels, tlVertical, tcFontName, tnFontSize, tlFontBold, tlFontItalics, tlFontUnderline
LOCAL lcStyle
    **********
    * Note:     B Bold
    *           I Italic
    *           N Normal
    *           O Outline
    *           Q Opaque
    *           S Shadow
    *           – Strikeout
    *           T Transparent
    *           U Underline
    *****
    lcStyle = IIF(tlFontBold, "B", "") + IIF(tlFontItalics, "I", "") + IIF(tlFontUnderline, "U", "")
    IF PCOUNT() > 2
        RETURN tnFoxels * FONTMETRIC(IIF(tlVertical, 1, 6), tcFontName, tnFontSize, lcStyle)
    ELSE
        RETURN tnFoxels * FONTMETRIC(IIF(tlVertical, 1, 6))
    ENDIF