The PC speaker is NOT capable of producing accurate frequencies under ANY circumstances, no matter what anyone claims (there are some people writing software and making claims to that effect). The reason why it cannot is a hardware issue and nothing to do with software.
The PC speaker is connected directly to the output of a timer chip in the PC. Basically, the timer can switch the speaker output high or low within certain time intervals. This will create a square wave by setting the output low for a certain time and then setting it high for the same time, then back to low again etc.
However the timing is limited to the timing resolution of the timer chip itself. The timer in the PC is based on very old technology and for compatibility is run with a master timing reference clock of around 1.8 Mhz (this is incredibly slow by modern standards). What this means in practice is that the timing resolution of the chip cannot possibly be better than (1/1.8Mhz = 0.55 microseconds). This sounds quite good, but is actually very poor. What it actually means is that the output wave can only be timed in integer multiples (whole numbers not fractions) of 0.55 microsecond steps.
Now when you go from one frequency to another each successive frequency will differ from the last by a certain timing fraction - there comes a point as the frequency increases where the step from one frequency to the next is less than the timing resolution of the speaker timer chip - when this happens the frequency becomes inaccurate. In the case of a PC (ALL PC's!) this "break point" occurs at 1341 Hz for 1 Hz steps. In other words the PC speaker output will give a frequency no more than 1Hz out for any frequency below 1341 Hz. Above that point it becomes inaccurate and what you ask for is not what you get.
The output from the speaker driver circuitry is a square wave but if the speaker is left connected it won't stay anywhere near square. If the speaker is removed the wave will be better but it will degrade in any connecting wire. So a good square wave is near impossible from the speaker output without extra circuitry.
Now to address the issue of sound cards. Sound cards can create waves by digital sampling. An analog signal is synthesized by lots of samples. This can create accurate SINE waves up to one half the sampling rate of the signal.
In the case of modern sound cards the highest sampling rate is 44 Khz so it can reproduce sine waves accurately up to 22Khz (in theory). In practice however, some special filtering is needed to avoid what is called aliasing noise and not all sound cards have it.
But a sound card cannot synthesize an accurate SQUARE wave for the same reason as the speaker circuit - the timing resolution of the sampling is the limiting factor - in the case of the speaker the timing resolution was 1/1.8 Mhz=0.55 microseconds). In the case of a sound card the timing resolution is (1/44Khz=22.7 microseconds) which is much WORSE than even the speaker! In the case of a sound card creating square waves the limiting frequency for 1Hz accuracy is approx 209 Hz! So a sound card cannot produce accurate square waves (to within 1Hz that is) above 209Hz!
However, the SINE wave output is perfect up to around 20 Khz IF a very good low pass filter (with sharp cut off above 20Khz) is applied through external circuitry. An external circuit to make a square wave from this sine will produce accurate square waves in the audio range.
Whether perfect square waves are necessary or not for Rife work is a separate issue. As I have explained on my web site the original Rife machines and the Hoyland variants used pure sine waves for modulation, NOT square waves and they worked just fine - and Rife's original principle was that a single pure frequency could kill pathogens. So it's not correct to say that ONLY square waves will have beneficial effects. However in practice most people find that square waves seem to work better for pad type devices.
CD's work the same way as sound cards, they cannot produce accurate square waves above 209 Hz and also as Ralph Hartwell has pointed out in the past, the timing of commercial CD players can drift significantly.
As always, I strongly recommend that people research the technicalities of any product before believing advertising claims - many in the Rife game are quite ridiculous if not outright cons.
ALWAYS get someone to check any device on real test equipment like an oscilloscope and frequency counter.
In my own research I've found time and again that certain conditions will cause pathogens to GROW - the signals will stimulate rather than kill them - others will wipe them out. It's not just a factor of frequency in general, but in most modern machines frequency is a critical factor that should not be ignored. There is a very fine line between between stimulation and mortality and even if you know what you're doing, caution is always necessary.
Finally, just as a separate issue. Windows 2000 (and XP) are based on Windows NT. The technology is different than the one used in Windows 98, ME and earlier ones. Basically, the NT core contains mechanisms to protect I/O devices from being accessed directly by user programs (it's possible but complicated). A program that accesses direct I/O (like the PC speaker) on an earlier platform will NOT work on NT, 2000 or XP unless it is designed to specifically access specialised API's in those operating systems or has a full custom device driver compliant to NT core standards. The program will run but may not give any actual output - or may bomb out with protection error messages.
Aubrey Scoon
The waveform coming out of a soundcard won't be very square, no matter what you tell the soundcard to generate (check this yourself with an oscilloscope). This is a limitation of the hardware. You need a waveshaper to get a nice squarewave.
You may get effects because you are applying electricity to your body (as with the various zappers/blood electrifiers) but you probably won't get resonant frequency effects.
And then there is the issue of voltage coming out of your soundcard (not very high) and the current that it can drive.
I know someone that, using a waveshaper, hooked himself up to his PC and used it as a cheap contact Rife device, but he has an engineering background and knew what he was doing, and attached the electrodes to his legs (less likelyhood of enough current reaching the heart and accidently stopping it).
Personally, I wouldn't be hooking myself directly up to a computer, and once I found out about how little current it takes to cause heart problems, I stopped helping people who want to do that using my software.
Fred Walter (Author of one of the many PC software solutions)