i'd like to know if there is still a way to pick the microphone signal from sound card, just using a "standard" win32 api.

the goal would be first to do a simple "clap event".

each time you clap your hands, or just snap fingers, and the procedure will return a true in eax register.
that would be pretty no?

why not later implement a sort of sound mouse?up, down, right, left, stop, clic! clap, rollup, rolldown, etc...

http://msdn.microsoft.com/en-us/library/dd743847%28VS.85%29.aspx

waveInOpen from winmm.dll

yes, wave recording api can handle that.

but the task would fall on matching the input sound with the database sound. You could try using a checksum but might not always work because voice and sound can be the same but different length

typedef:

it will use filtering, and FFT in order to detect clap, clip, clac...
you cannot just compare waveforms. you should at least work in the frequency/amplitude dimension, the sounds spectrum.

the main problem i see is the real time aspect.
i don't know how works the execution of a window program. then, i cannot predict if it will be easy to code or not.

but i am sure it is possible to do. there are a lot of programs that still do FFT, but mainly in order to do sound effects, not to analyse it.

the audio recognission is not an hurry problem apprentlly. not more than a good lotery/poker/casino progam.


what i'd like to do with this kind of code would be very simple, and limited to very basic actions. for example, in a PPS presentation, it can be nice to have an original way to switch pages, clap!

I remember 6 years ago, I installed an IBM voice recognition program.

It had an option to integrate with windows, you could say 'start' and Start menu would popup. You could create shortcuts with voice commands, like in a mobile phone. The thing didn't work well.

Just make a simple CONSOLE app with a message loop and add a timer.