Here is my problem :
I try to record from a mic and put ONLY words or sentences in .wav file.
So I put the flag for looping recording to TRUE.
I now want to analyze my FSOUND_SAMPLE to determine if an user said something or if it is only ambiant noise. And then if a word or a sound higher than ambiant noise has been pronounced, it will start to record in a wav file.
HELP ME PLIZ
- BuGbOy asked 14 years ago
Ok I found how to use those function. But I have no another problem.
I don’t want my recording sound to be played. They should be no sound. But if I change Output or I SetVolume of the current playing channel. It doesn’t work anymore.
I used the record example.
I managed to record to a wav file using a circular buffer modifying the record example (the one that plays back recorded sound in real time). You can set up a dsp unit to monitor current audio and write it to the file only when you decide.
You probably will need a second buffer to hold data coming out from DSP before writing it to disk.
Not the easiest thing in the world (depends on your skills, of course ) but it can be done.
Yes, but this way he would have to poll recording data very fast to determine whether level is below treshold or not… it seems an easier solution but (in my opinion) would not give optimal results.
BuGbOy, if you have no skills it will be a good idea to study the examples (and may be other things) until you build enough confidence to try yourself.
With DSP (or VUlevels) you can determine if sound level is higher than something you specify, if you are VERY clever I suppose you could even tell the difference between spoken words and background noise (or other interferences).
A quick and dirty way could be analyzing the buffer before writing it to disk and writing only what you need. May be a good compromise that uses no dsp and no polling. Of course you have to do this while locking your recording buffer so you must do it quickly but is should work.
Please login first to submit.