A capacitor wouldn't protect it from high input voltages. it's possible that the sound card has protection diodes, however those can still be overloaded if you don't limit the input current. For that reason, you should DEFINITELY use a series resistor on the input, such as the 47K you used.
As far as I know, sound cards always have an AC-coupled input, so you cannot use them for measuring DC signals, thus they would be no good for a voltmeter, unless you mean an AC voltmeter.
Using a 47k series resistor alone would not significantly affect the voltage of the input, you would have to use a resistor divider to drop the input voltage to a range that the sound card could read.
A "real" oscilloscope can often handle voltages of over 100V, and often over 1000V if an attenuated probe is used. Real oscilloscopes also aren't AC-coupled (well, unless you enable it) so they can directly measure DC voltages. And, of course, an oscilloscope can go to MUCH higher frequencies than a sound card, even low-end scopes can handle 10's of MHz. Also, you should know that a sound card is NOT going to provide distortion-free measurements of your signal. I'm not saying it's totally worthless; those sound-card-oscilloscope programs are definitely interesting because they let you experiment with some limited signals, and are great for demonstration and experimentation for people without any real oscilloscope experience, however they are not all that useful when it comes to actual electronics testing, because they can only handle a rather narrow set of signals.
Face it, soundcards are NOT built to be oscilloscopes.