I downloaded one album encoded in mp3. The problem is that tracks have silent blocks on the beginning/end of each track, and since the tracks fade one into other i can't get seamless playback if I just decode them into wav and burn with 0 pause between tracks. There is always hearable transition (like one in sync error) between them. I was wondering if is possible to solve this problem and get the copy identical to original album somehow (with correct transitions), without messing with wav files in editor. I tried correcting that in Cool Edit, but the result wasn't perfect. Why do the mp3 files have these strange silent areas (usually at beginning, but they also appear on the end)? I read about encoder delay, is it that maybe?

That's a drawback of the MP3 format - it simply doesn't support gapless playback. I'd try removing the silence with a Wave editor.

You can try mp3trim! It will trim part of the beginning and ending of each file! I do this on my mp3s, even the ones I encoded myself...

You can try mp3trim! It will trim part of the beginning and ending of each file! I do this on my mp3s, even the ones I encoded myself... you save space, get your files checked for errors...

Thanx. I tried to find about this problem on the net, and found few pages. Everything is well explained on mp3dev. Every mp3 encoder has its own delay, and that delay produces silence in output mp3. The decoders also have delay. Lame decoder does it well, it has automatic removing of the encoder/decoder delay at the beginning of decoding. That means that in the output wav file the overall delay will be corrected. The only problem is finding which encoder encoded certain mp3, and knowing their encoder delay. I know Lame's (576 encoder delay + 529 decoder delay = 1105 samples of delay) because I tried, but for other encoders I don't know. The other problem is silence at the end of decoded mp3 (wav file), as a result of padding. Since Lame decoder doesn't have padding silence removal, I was wondering if there is some program which does it?

holkie:



The problem is not silence that is actually *in the file*, it's silence that gets generated when decoding the file. Decoding the trimmed MP3s would not solve the problem.