[Chapter 3] 3.2 FontMetrics

3.2 FontMetrics

The abstract FontMetrics class provides the tools for calculating the actual width and height of text when displayed on the screen. You can use the results to position objects around text or to provide special effects like shadows and underlining.

Like the Graphics class, FontMetrics is abstract. The run-time Java platform provides a concrete implementation of FontMetrics. You don't have to worry about the actual class; it is guaranteed to implement all the methods of FontMetrics. In case you're curious, on a Windows 95 platform, either the class sun.awt.win32.Win32FontMetrics ( JDK1.0) or the class sun.awt.windows.WFontMetrics ( JDK1.1) extends FontMetrics. On a UNIX/Motif platform, the class is sun.awt.motif.X11FontMetrics. With the Macintosh, the class is sun.awt.macos.MacFontMetrics. If you're not using the JDK, the class names may be different, but the principle still applies: you don't have to worry about the concrete class.

The FontMetrics Class

Variables

protected Font font: The font whose metrics are contained in this FontMetrics object; use the getFont() method to get the value.

Constructors

protected FontMetrics (Font font): There is no visible constructor for FontMetrics. Since the class is abstract, you cannot create a FontMetrics object. The way to get the FontMetrics for a font is to ask for it. Through the current graphics context, call the method getGraphics().getFontMetrics() to retrieve the FontMetrics for the current font. If a graphics context isn't available, you can get a FontMetrics object from the default Toolkit by calling the method Toolkit.getDefaultToolkit().getFontMetrics (aFontObject).

Font height

Four variables describe the height of a font: leading (pronounced like the metal), ascent, descent, and height. Leading is the amount of space required between lines of the same font. Ascent is the space above the baseline required by the tallest character in the font. Descent is the space required below the baseline by the lowest descender (the "tail" of a character like "y"). Height is the total of the three: ascent, baseline, and descent. Figure 3.1 shows these values graphically.

Figure 3.1: Font height metrics

[Graphic: Figure 3-1]

If that were the entire story, it would be simple. Unfortunately, it isn't. Some special characters (for example, capitals with umlauts or accents) are taller than the "tallest" character in the font; so Java defines a value called maxAscent to account for these. Similarly, some characters descend below the "greatest" descent, so Java defines a maxDescent to handle these cases.

NOTE:

It seems that on Windows and Macintosh platforms there is no difference between the return values of getMaxAscent() and getAscent(), or between getMaxDescent() and getDescent(). On UNIX platforms, they sometimes differ. For developing truly portable applications, the max methods should be used where necessary.

public int getLeading ()

The getLeading()method retrieves the leading required for the FontMetrics of the font. The units for this measurement are pixels.

public int getAscent ()

The getAscent()method retrieves the space above the baseline required for the tallest character in the font. The units for this measurement are pixels. You cannot get the ascent value for a specific character.

public int getMaxAscent ()

getMaxAscent() retrieves the height above the baseline for the character that's really the tallest character in the font, taking into account accents, umlauts, tildes, and other special marks. The units for this measurement are pixels. If you are using only ordinary ASCII characters below 128 (i.e., the English language character set), getMaxAscent() is not necessary.

If you're using getMaxAscent(), avoid getHeight(); getHeight() is based on getAscent() and doesn't account for extra space.

For some fonts and platforms, getAscent() may include the space for the diacritical marks.

public int getDescent ()

The getDescent() method retrieves the space below the baseline required for the deepest character for the font. The units for this measurement are pixels. You cannot get the descent value for a specific character.

public int getMaxDescent () public int getMaxDecent ()

Some fonts may have special characters that extend farther below the baseline than the value returned by getDescent(). getMaxDescent() returns the real maximum descent for the font, in pixels. In most cases, you can still use the getDescent() method; visually, it is okay for an occasional character to extend into the space between lines. However, if it is absolutely, positively necessary that the descent space does not overlap with the next line's ascent requirements, use getMaxDescent() and avoid getDescent() and getHeight().

An early beta release of the AWT API included the method getMaxDecent(). It is left for compatibility with early beta code. Avoid using it; it is identical to getMaxDescent() in every way except spelling. Unfortunately, it is not flagged as deprecated.

public int getHeight ()

The getHeight() method returns the sum of getDescent(), getAscent(), and getLeading(). In most cases, this will be the distance between successive baselines when you are displaying multiple lines of text. The height of a font in pixels is not necessarily the size of a font in points.

Don't use getHeight() if you are displaying characters with accents, umlauts, and other marks that increase the character's height. In this case, compute the height yourself using the getMaxAscent() method. Likewise, you shouldn't use the method getHeight() if you are using getMaxDescent() instead of getDescent().

Character width

In the horizontal dimension, positioning characters is relatively simple: you don't have to worry about ascenders and descenders, you only have to worry about how far ahead to draw the next character after you have drawn the current one. The "how far" is called the advance width of a character. For most cases, the advance width is the actual width plus the intercharacter space. However, it's not a good idea to think in these terms; in many cases, the intercharacter space is actually negative (i.e., the bounding boxes for two adjacent characters overlap). For example, consider an italic font. The top right corner of one character probably extends beyond the character's advance width, overlapping the next character's bounding box. (To see this, look back at Figure 3.1; in particular, look at the ll in O'Reilly.) If you think purely in terms of the advance width (the amount to move horizontally after drawing a character), you won't run into trouble. Obviously, the advance width depends on the character, unless you're using a fixed width font.

public int charWidth (char character)

This version of the charWidth() method returns the advance width of the given character in pixels.

public int charWidth (int character)

The charWidth() method returns the advance width of the given character in pixels. Note that the argument has type int rather than char. This version is useful when overriding the Component.keyDown() method, which gets the integer value of the character pressed as a parameter. With the KeyEvent class, you should use the previous version with its getKeyChar() method.

public int stringWidth (String string)

The stringWidth() method calculates the advance width of the entire string in pixels. Among other things, you can use the results to underline or center text within an area of the screen. Example 3.1 and Figure 3.2 show an example that centers several text strings (taken from the command-line arguments) in a Frame.

Example 3.1: Centering Text in a Frame

import java.awt.*;
public class Center extends Frame {
    static String text[];
    private Dimension dim;
    static public void main (String args[]) {
        if (args.length == 0) {
            System.err.println ("Usage: java Center <some text>");
            return;
        }
        text = args;
        Center f = new Center();
        f.show();
    }
    public void addNotify() {
        super.addNotify();
        int maxWidth = 0;
        FontMetrics fm = getToolkit().getFontMetrics(getFont());
        for (int i=0;i<text.length;i++) {
            maxWidth = Math.max (maxWidth, fm.stringWidth(text[i]));
        }
        Insets inset = insets();
        dim = new Dimension (maxWidth + inset.left + inset.right,
            text.length*fm.getHeight() + inset.top + inset.bottom);
        resize (dim);
    }
    public void paint (Graphics g) {
        g.translate(insets().left, insets().top);
        FontMetrics fm = g.getFontMetrics();
        for (int i=0;i<text.length;i++) {
            int x,y;
            x = (size().width - fm.stringWidth(text[i]))/2;
            y = (i+1)*fm.getHeight()-1;
            g.drawString (text[i], x, y);
        }
    }
}

Figure 3.2: Centering text in a frame

[Graphic: Figure 3-2]

This application extends the Frame class. It stores its command-line arguments in the String array text[]. The addNotify() method sizes the frame appropriately. It computes the size needed to display the arguments and resizes the Frame accordingly. To compute the width, it takes the longest stringWidth() and adds the left and right insets. To compute the height, it takes the current font's height, multiplies it by the number of lines to display, and adds insets. Then it is up to the paint() method to use stringWidth() and getHeight() to figure out where to put each string.

public int charsWidth (char data[], int offset, int length)

The charsWidth() method allows you to calculate the advance width of the char array data, without first converting data to a String and calling the stringWidth() method. The offset specifies the element of data to start with; length specifies the number of elements to use. The first element of the array has an offset of zero. If offset or length is invalid, charsWidth() throws the run-time exception ArrayIndexOutOfBoundsException.

public int bytesWidth (byte data[], int offset, int length)

The bytesWidth() method allows you to calculate the advance width of the byte array data, without first converting data to a String and calling the stringWidth()method. The offset specifies the element of data to start with; length specifies the number of elements to use. The first element of the array has an offset of zero. If offset or length is invalid, bytesWidth() throws the run-time exception ArrayIndexOutOfBoundsException.

public int[] getWidths ()

The getWidths() method returns an integer array of the advance widths of the first 255 characters in the FontMetrics font. getWidths() is very useful if you are continually looking up the widths of ASCII characters. Obtaining the widths as an array and looking up individual character widths yourself results in less method invocation overhead than making many calls to charWidth().

public int getMaxAdvance ()

The getMaxAdvance() method returns the advance pixel width of the widest character in the font. This allows you to reserve enough space for characters before you know what they are. If you know you are going to display only ASCII characters, you are better off calculating the maximum value returned from getWidths(). When unable to determine the width in advance, the method getMaxAdvance() returns -1.

Miscellaneous methods

public Font getFont ()

The getFont() method returns the specific font for this FontMetrics instance.

public String toString ()

The toString() method of FontMetrics returns a string displaying the current font, ascent, descent, and height. For example:

sun.awt.win32.Win32FontMetrics[font=java.awt.Font[family=TimesRoman,
name=TimesRoman,style=bolditalic,size=20]ascent=17, descent=6, height=24]

Because this is an abstract class, the concrete implementation could return something different.

3.2 FontMetrics

The FontMetrics Class

Figure 3.1: Font height metrics

Example 3.1: Centering Text in a Frame

Figure 3.2: Centering text in a frame

Font Display Example

Example 3.2: Font Display

Figure 3.3: Fonts available with the Netscape Navigator 3.0 and Internet Explorer 3.0