We are looking for a programmer to develop a module to be incorporated in our existing system. Basically, we need that this module recognizes a WAV file consisting of spoken alphanumeric codes, and returns a plain ASCII string.
We require this module as an un-managed Win32 DLL with the fewest possible dependencies, highly efficient, thread-safe, no memory leaks, etc. We code in C/C++ so they have preference, but anything callable from these two languages will be also considered. You are free to use any 3rd-party library as long as it also complies with our requirements. Some details: Each WAV file will include a single (and short) alphanumeric code.
Each code is spelled out, character by character, by a human. There will be ambient noise in the audio. Precision of 85% or higher is required**. A response time under 1 second will be enough. **About precision: just to be clear, we mean 85% of successfully recognized codes after hundreds, maybe thousands.
We request that if you are using some sort of 3rd party liberary, that it doesnt interfere with the overall scope of the project. The details of this project are very specific, and if they cant be met, time is wasted on both parties sides. We are not willing to compromise on the scope.