Sound Recorder work...Advice me.

Hi all!

I am creating a sound recorder with some advanced features. This sound recorder will be released in the beginning of year 2004 and will go to download.com for selling.

What I wanted to ask you guys is: what features would you like it to have?
For now it has:

- Record/Play/Pause/Stop
- Silence detection (+calibration: you say "how silent")
- Scheduled recording
- Save the recorder as WAV,WMA, MP3 and OGG
- Converter (from one format to another) and resampler
- I intend to put some sound effects like reverse and such... but I am not sure what would be useful.
- I am still wondering if I should put a mixer (to mix more two or more files into one).

Please, advice me... what should I add as funcitonallity?
Also, I intend to sell that for 15$-20$... do you think this price is good?

Thank you!

what makes it better than something like lets say goldwave?

It is MUCH cheaper for one!

Also, GoldWave is more like editor than to recorder.
My app will be mostly recorder.
Also, I found no Silence Detetion in GoldWave, did you?

And at last- I want to hear from people what would they like from such a software?

Silence deletion is a lot higher quality if you use a compressor.

ah compressors. sheer genius.

Quote:

---

Originally posted by loobian
It is MUCH cheaper for one!

Also, GoldWave is more like editor than to recorder.
My app will be mostly recorder.
Also, I found no Silence Detetion in GoldWave, did you?

And at last- I want to hear from people what would they like from such a software?

---

i use goldwave for recording all the time.

and it does have silence detection/removal, along with a bazillion other things.

I recommend including a feature that seperates audio from different video formats. I had a problem with that a few weeks ago when I wanted to extract and record audio from a real player video file onto a cd....it became a very hard task. that would be a great thing to have

I doubt that you'll be able to produce anything close to a production quality sound program. Not because you don't have the ability but because you lack the resources. There is a lot of mathematics behind all the filters and such and many of these software companies have sound engineers that come up with the stuff.


But if you really want to do it here is a way to add additive mixing.

1. Take value from each sound sample
2. Add all values
3. Divide by number of values
4. Place result from step 3 into primary sound buffer
5. Play the sound - everything will be mixed.

One problem is that the more sounds you mix, the quieter the result. There are other mixing algorithms out there but additive mixing is prob the most popular.

or voice cancelling. subtract the values of hte left channel from the right channel, since voice (usually) is evenly distributed across both channels, it should be effectivley cancelled out.

though this also has the effect that anything else which is equal in both channels gets cancelled out too.

try developing something like a frequency filter for voice. just an idea

Also you can linear interpolate the sound wave which will take some of the 'squareness' of the wave out. To linear interpolate it use the slice of time you want as the basis for the interpolation.

If you want the sound that should be halfway between certain bytes or words of sound then use .5 as the value.

Code:

---

double LI(double v1,double v2,double f1)
{
  return (v1+f1*(v2-v1));
}

double LI(double v1,double v2,double f1)
{
  double result=0.0;
  asm {
    fld    [v2]
    fsub  [v1]
    fmul  [f1]
    fadd [v1]
    fstp [result]
  }
  return result;
}
 
for (int i=0;i<Buffer_length-1;i++)
{
  BYTE ss1=PrimaryBuffer[i];
  BYTE ss2=PrimaryBuffer[i+1];
  double interpolatedsample=LI((double)ss1,(double)ss2,.5);
  PrimaryBuffer[i]=(BYTE)interpolatedsample;
}

---

Returning values in variables that are popped off of the stack is not a good practice. I only did it here because it was the easiest to represent.