OK. My first problem is solved. I learned the hard way that LAME 3.93 and 3.93.1 have some pretty bad bugs, especially when compressing whole CD sized wav files (see topic in General MP3).

In discovering I needed to back step to 3.90.3, I discovered that the 3.90.3 APE version supports Cue Sheets. Apparently this allows one to feed in a large wav file (whole CD) with a cue sheet (apparently generated by Exact Audio Copy?) and get out "gap-less" MP3s (which if your player can handle them, and you have the right sound card, will result in seamless play).

I've looked around, but I don't find any step-by-steps that tell me how to start with a CD, EAC, LAME 3.90.3 APE, and end up with a set of "gap-less" mp3s. If anyone knows this, please (please) share.

Compress the large wav file into mp3 using lame. Then open the cue sheet with notepad, and change the location of the wav file to the mp3 file (just change wav to mp3). Then search for mp3cue, (sorry I don't remember where I got it because the guerillasoft.com website is gone) and tell winamp to open the cue file after you install mp3cue. The mp3s won't be gapless, if you want to know why maybe someone with more knowledge will reply. Or just search the forum.

You should do a search - you probably would have found this thread.

It should have what you need

Thank you for the reply. I didn't realize you could do this with WinAmp. I was refering to a method that uses features inside of this version of Lame to create the files on the fly (while 'encoding'). I ran across a document at http://www.geocities.com/nyaochi2000/lame/cuesheet/ that somewhat discusses the technique, but it does not have complete details. I know "gap-less" files are possible, because I managed to create them, but the process took hours (actually days this first time if you consider all the problems I had).

I did discover that it matters where the transition takes place. I had to fiddle with each one to get a clean transition. Moving it a few frames either way may introduce an audiable transition. My theory about what causes this is completely unsubstantiated, but I suspect it has something to do with the "Byte Bucket".

I recently read somewhere that it can take several frames to "start up" mp3 playback because of something to do with the "Byte Bucket"; apparently some sort of buffer. If this is the case, then I surmise that if you break a file in the wrong place, this "startup" makes noise. (anyone feel free to wack me with the reality stick if I'm completely off base on this). I don't know if this "Byte Bucket" is VBR specific or not. If not, I was thinking that a short run of CBR frames across the break might make the transition smoother (again, wild speculation on my part). Now some might say that frames are so small that it doesn't matter, but from experience I can tell you that a single bad sample value can make quite a bit of noise.

Anyway, if I choose the spots carefully, there is no audiable transition on my computer when I play back in WinAmp. That being said, when I play them back on my kid's computer, there is a slight audiable transition. I believe this is caused by the different sound card (all other things being equal).

The other benefit of doing all of this is that if I burn these tracks to a CD, they play back seamlessly when I use Easy CD Creator and write them "Disk At Once".

Thank you! I did do a search (several actually) but perhaps I just didn't use the right key-words...

I'll check it out.

hmm. I didn't know that there are big bugs in 3.93.1.

I know about a big bug in 3.93(.0), and that's the reason for the 3.93.1 bugfix relesase.

just wondering,
amano

Well, I'm no expert, but when I 'encoded' my whole CD wav file (of 509,447 Mb) and found that the resulting mp3 was pretty seriously corrupted I started digging. In my e-travels I was told that I should steer away from 3.93.0. So my first inclination was to go up to 3.93.1. I installed that version and re-'encoded' the wav and got the same corrupted result. When I backed down to 3.90.3, the problem went away. Sounds pretty serious to me . I've saved this wav file in the event that any developer wants to take a look at why this happens, probably something I would have to send them on a data CD (which I would be happy to do).

F-everyone's-I, I've decided not to persue this anymore. I've discovered that Ogg Vorbis doesn't have this problem. So I'll keep my mp3s for most albums, but use Ogg Vorbis for any CDs that have cross track play. My main player (WinAmp) supports the format and plays back flawlessly; using quality -q6 seems as good (at least to my ears) as the standard preset. I figure there's no point killing myself over it. I'll encode a full-CD mp3 version for my portable player, but at home I'll just ogg-v.

The mp3's will be gapless since you have made one large gapless mp3.


The easiest way to playback the mp3+cue is to use Foobar, it has cue support built in.


MPC files playback gaplessly when ripped as seperate tracks.