c# 4.0 - Mixed wave stream to mp3 using Lame -

....... .......  var mixer = new wavemixerstream32(); mixer.autostop = true;  var messageoffset = background.totaltime; var messageoffsetted = new waveoffsetstream(message,                                           timespan.fromseconds(10),                                           timespan.zero,message.totaltime.subtract(timespan.fromseconds(10)));               var background32 = new wavechannel32(background); background32.padwithzeroes = false; background32.volume = 0.6f;                mixer.addinputstream(background32);  var message32 = new wavechannel32(messageoffsetted); message32.padwithzeroes = false; message32.volume = 0.3f; mixer.addinputstream(message32);  var ws = new wave32to16stream(mixer); 

we trying mix multiple mp3's , wave files create single mp3 finally.

example: 5 sources files (3 mp3's, 2 wave files)

we giving input of each stream wavemixerstream32 , converting using wave32to16stream

we need final stream converted mp3. using lame , passing stream encodemixedstreamasmp3 returning error "invalid file format".

after researching, found riff header missing in mixed wave stream.

how add riff mixed wave stream generated based on multiple sources (mp3 , wave)?

an naudio wavestream stream of sample data specified format, not riff wav file expected lame. convert riff wav file need add riff headers , on. naudio.wave.wavefilewriter class this.

if you're working smallish output files aren't going blow memory, can simple (assuming yeti's lame wrapper or similar):

(code updated 19-aug-2013):

public byte[] encodemp3(iwaveprovider ws, uint bitrate = 128) {     // setup encoder configuration     wavelib.waveformat fmt = new wavelib.waveformat(ws.waveformat.samplerate, ws.waveformat.bitspersample, ws.waveformat.channels);     yeti.lame.be_config beconf = new yeti.lame.be_config(fmt, bitrate);      // encode wav mp3     int blen = ws.waveformat.averagebytespersecond;     byte[] buffer = new byte[blen];     byte[] mp3data = null;      using (memorystream mp3strm = new memorystream())     using (mp3writer mp3wri = new mp3writer(mp3strm, fmt, beconf))     {         int rc;         while ((rc = ws.read(buffer, 0, blen)) > 0)             mp3wri.write(buffer, 0, rc);         mp3data = mp3strm.toarray();     }     return mp3data; } 

-- update: setting mp3 encoding parameters

using yeti lame wrapper, can specify mp3 encoding parameters passing in be_config structure so:

wavelib.waveformat fmt = new wavelib.waveformat(ws.waveformat.samplerate, ws.waveformat.bitspersample, ws.waveformat.channels);  yeti.lame.be_config beconf = new yeti.lame.be_config(fmt, 64); using (memorystream mp3strm = new memorystream()) using (mp3writer mp3wri = new mp3writer(mp3strm, fmt, beconf)) { ... } 

the be_config constructor takes wavelib.waveformat , optional cbr bit rate in kb/s. if don't specify bit rate defaults 128kb/s output. if don't supply configuration mp3writer creates default 1 waveformat.

the code above creates be_config output rate set 64kb/s. if want different bit rate, specify in place of 64 in new yeti.lame.be_config(fmt, 64);

be_config contains lot of other options, of won't ever use. check out (slightly sketchy) documentation here.