Think of using DSD as a signal pathway:
- you need a source signal aka baseband audio (unfiltered) and up till recently that meant the good old discriminator tap to provide the baseband audio for decoding purposes, but nowadays (within the past 2 years) there's a new way of getting such source material and that's by using "cheap USB TV tuner sticks" made specifically for use outside the US (because the rest of the world typically uses DVB-T technology for digital TV broadcasts while here in the US we just gotta do our own thing so we use ATSC technology and they're incompatible with each other)
- you'll then be routing that through some software on a computer (Windows being the most popular operating system overall so Windows apps, but Linux has a strong showing with respect to software defined radio and other aspects related to decoding digital communications modes)
- the decoded signal is then sent to your computer's audio subsystem for translating it into something you can actually hear
So point A to point B to point C, or more less, the idea being the source goes through all that, ends up at DSD's door so to speak, and DSD decodes it as needed and then outputs to your computer's audio subsystem and voila, magical voices from the Ether!!! MUAHAHAHA (ok, a little overboard there, I admit it)
With DSD, it can be run entirely by itself if you have a discriminator tap on a receiver/scanner and you can feed that into the computer's sound card (preferably into the line-in but mic-in can be used with very
very careful attention to the input levels so you can keep the signal useful but not distorted or overloading the input). So with a discriminator tap and DSD you can decode digital comm formats that DSD supports basically with just that program alone.
Now, if you happen to have one of those "cheap USB TV tuners" you'll need software to control/tune the stick (commonly referred to as RTL devices since they tend to use a Realtek control chip - Realtek doesn't make the tuner chip on such hardware, just the RTL2832U chip that controls the USB interface and data pump for connecting to a computer). For most people that I'm aware of, the popular software is known as SDR# (SDR-Sharp) and it's pretty awesome for what it can do and also considering that - like most SDR software these days - it's completely free of charge. How cool is that?
I don't use any other SDR software except SDR# but there are different ones available such as RTL-SDR, GNU Radio, SDR-Radio, and HDSDR to name the ones I'm aware of, so because of that I can't offer anything in terms of instructions for all those other ones. Because SDR# tends to be the most referred to hence most popular (at least as I read it) that's what I'd suggest for starting out aka "getting the damned thing to work."
Before I do a full write-up to help, I need to know what your setup is:
- do you have a discriminator tapped receiver/scanner + DSD on a computer already attached by way of some patch cord, or...
- are you using some RTL device + SDR software + DSD on a computer
Once you describe what hardware/software setup you have, then I/we can answer some more questions or provide more info.
And just as a final note, there is now another digital speech decoder program known as DSD+ or DSDPlus, it's only been around for about 2 weeks (first release) but even so it's proven itself to be quite remarkable in terms of how solid it is, the ability to decode even weak digital signals with relative ease by comparison to DSD (for most people based on what they're reporting), and a lot of options like the ability to actually record a given weak signal (raw recording of the actual data as a WAV file) and then go back with another piece of software called dsdtune (from a different author, for the record) and "tune" the decoding potential. You then feed the decoder tuning parameters back to DSD+ and it will amaze you at just how well it's able to clean up what was previously almost not listenable at all.
Just something to make a note of, try DSD and DSD+ and then figure out which works best for you, sometimes you may discover that one works better for a given system, so use both.