Java Nio中Charset的decode()方法问题

Question

    try {
        ByteBuffer byteBuffer = ByteBuffer.allocate(20);
        System.out.println("capacity = " + byteBuffer.capacity() + ", limit = " + byteBuffer.limit() + ", position = " + byteBuffer.position());

        FileChannel channel = new FileInputStream(PATH).getChannel();
        channel.read(byteBuffer);

        System.out.println("capacity = " + byteBuffer.capacity() + ", limit = " + byteBuffer.limit() + ", position = " + byteBuffer.position());
    byteBuffer.clear();
        System.out.println("capacity = " + byteBuffer.capacity() + ", limit = " + byteBuffer.limit() + ", position = " + byteBuffer.position());

        CharBuffer buffer = Charset.defaultCharset().decode(byteBuffer);
        System.out.println("capacity = " + buffer.capacity() + ", limit = " + buffer.limit() + ", position = " + buffer.position());
    } catch (Exception e) {
        e.printStackTrace();
    }

代码如上，假设有一个文件是系统默认GBK编码，文件内容只有一个数字1，那么上述代码在decode之后得到的CharBuffer的容量为什么是21，比原ByteBuffer都大？而且CharBuffer的容量与极限为什么不同？