The Analog Devices bit is a plugin for the widely-available math software package Matlab but it appears to be restricted to code that's destined to run on their processors (they have their own language, VisualDSP++, that you use to code for their chips. I believe that the plugin ultimately spits out VisualDSP++ code after you're done doing your design work.). Matlab in and of itself could be used to model the mathematics of what's going on in DSP calculations, along with many other math-related software packages (Mathematica, PARI-GP, Maple, etc). That math is universal regardless of which programming language you choose to implement it.
C++ is an object-oriented language, a "low-level" language would be C (not the same as C++) or Assembly language. For the most part I believe that DSP algorithms are written in C and/or Assembly (if the target processor or processors are known) and the higher-level control code in C++. Some DSP libraries are starting to appear for newer languages as well, like .net or Java. These, however, are generally less efficient and better suited for prototyping than running real-time high performance code (unless you plan to use lots of hardware horsepower to make up the difference).
iPhone development is done using Objective C, which is a variant of C that is object-oriented. You can read the full spec here:
http://developer.apple.com/documentation..._section_1.htmlHowever, you need to be running MacOS v10.5.4 or greater to get into iPhone development. The SDK includes a sort of "virtual" iPhone that you can use to test your code. You can sign up to download a free version of the SDK on dev.apple.com.