Creating a soundtrack (and saving it)

[Seymour Clufley]

I want to create a WAV file from scratch, specifying other WAV files to be mixed into it at different times. Something like:

Code: Select all

total length = 13seconds
start clip1.wav at 2seconds
start clip2.wav at 7seconds
save mix to my_soundtrack.wav

I've never worked with PB and audio before. Can anyone tell me the simplest way to get this done?

Thanks,
Seymour.

[Guimauve]

Maybe a little look to the CodesArchiv can give you some ideas :

http://www.purearea.net/pb/CodeArchiv/M ... nMemory.pb
http://www.purearea.net/pb/CodeArchiv/M ... aveFile.pb
http://www.purearea.net/pb/CodeArchiv/M ... keSound.pb
http://www.purearea.net/pb/CodeArchiv/M ... ordWave.pb

Best regards
Guimauve

[Seymour Clufley]

Thanks for the links, but they don't give any information about how to mix several sounds. Does anybody have experience of doing this in PB?

[kenmo]

I actually wrote some code that may be exactly what you need... I can post it tonight when I am home.

It never became a complete WAV library like I planned, but it can load basic WAVs, mix them, and save to files (or you can use CatchSound() directly).

[Seymour Clufley]

That sounds really good, Kenmo.

Using the code linked by Guimauve, I did eventually get the job done, but the sounds aren't mixed. Each one stops when the next one starts - result is clicks and pops, but it's better than nothing.

But I would be very interested in the code you mentioned.

[kenmo]

Hmmm, I looked over my old code... and I don't like it. (I could send you it if you'd like to look through it.)

It works, but it's really slow, because I wrote it to be TOO general... every sample you modify (millions) checks for mismatches in bit depth (16 or , number of channels, etc. and compensates accordingly.

Now I remember thinking "this must be why pro audio software just loads every input file into a stereo floating-point buffer..."

I'm tempted to rewrite a nicer, floating-point WAV library (which would simplify all the processing)... but I'm busy the rest of this week ... Maybe in the near future.

[Seymour Clufley]

Kenmo,

How about just making it a rule that all files used must be the same bitdepth/samplerate etc? Then you could get rid of the checking and compensating. Or as you say, you could convert all files upon loading to a uniform bitdepth/samplerate, and again, get rid of the constant checks. Wouldn't that speed it up?

Either way, yes I'd like you to send me it. That'd be great.

Seymour.

[kenmo]

I messaged you 1600 lines of code

[oryaaaaa]

My audio souce code (I needed 4 month 24/7)
http://purebasic.coolverse.jp/_userdata/nr3.7z

Format
CreateFile
<WAV header>
FlushFileBuffers(#File)
<WAV RAW1>
FlushFileBuffers(#File)
<WAV RAW2>
FlushFileBuffers(#File)
<WAV RAW3>
FlushFileBuffers(#File)
CloseFile
I recomend FileBuffersSize(#File, Size). Size is 10MB over.

If you fail creating Raw wave data, then you should use Audacity Import RAW-WAV for Test.
My souce is very crazy. Supported CD-RIP, FLAC Import, FLAC Encode, FLAC Decode,
Up Sampling, Waveout, and CDDA PLAY (and Copy control CD supported).

[waffle]

wow, I have been struggling for 4 days on this, and just found the answer here
Thanks for the help. Basically, I hacking Doom1.WAD for sound information that is in raw format.
The PC format is fun too, but I choose to ignore that, the newer (DOOM2.WAD +++) do not use
PC speakers because sound cards were available then. But thanks again.
Maybe I will update my website with the new doom browser when I get this working...
And, NO, I do not export the sounds to a .wav file, I only want to play them.

Maybe for extra fun, I will try and decifer the MIDI files

[Seymour Clufley]

Just in case anyone else is looking for this kind of thing, it can now be done very easily using PureSoX. There is a dedicated procedure for it, SoX_GenerateSoundtrack(), and demo code can be found in that thread.