[dreilly1]

Anyone who has been on these forums for a while knows that I can't stand the way most MP3 players insert gaps between tracks. I appear to be unusual because I listen to whole albums, and not individual tracks. (I don't even have playlists in iTunes, I just browse for the album that I want...) This is particularly noticable on live albums and on albums where one song flows into another. I'm not an audiophile or anything, but I was a bass player in a former life and anything that disturbs the flow or rhythm of a song, even between songs, really gets my goat.

It's also something that some software gets right. I had to go through some pain to encode MP3's to my satisfaction -- I would rip the entire CD at once with cdparanoia, and then encode the entire album with lame and the --nogap option. I could play these tracks back to my satisfaction with the Unix command-line mpg123 player (which apparently is too basic to know about gaps), or that little-known program WinAmp on that other platform.
(For what it's worth, OS X/Unix geeks who use fink can duplicate this: fink doesn't have cdparanoia, but its version of cdda2wav comes with the paranoia libraries as an option...)

One problem with this setup is that lame does the gapless encoding by slightly moving where one track ends and the other begins to fill out an MP3 frame. As a result, if you start playing in the middle of an album, you may get a split-second of the previous track before the desired track starts. That doesn't bother me, but I know it may bother some people.

So when the new iTunes came out, I downloaded it hoping that the Apple engineers figured out how to eliminate the gap, and much to my chagrin, it didn't. I even posted to that effect in the iTunes features thread. Then it hit me: I downloaded the new iTunes, but not the new QuickTime! D'oh! So, I did some experimenting:

- QuickTime 6.5 could not play anything without the gap, either the lame MP3's or MP3's encoded through QuickTime.

- QuickTime 6.5.1 still couldn't play lame MP3's without the gap. But it could play stuff that it encoded itself! Better than that, it could play songs encoded with QT6.5! (That tells me that they tweaked the playback instead of the encoder).

- WinAmp could play the lame MP3's without the gap, but could not play QuickTime's MP3's that way. (Which means that QuickTime is likely doing something other than buffering the next track like WinAmp does).

When I realized that the playback algorithm was tweaked I tried a further experiment, and played an album that I bought through iTunes in its first week of existence. Back then, I noticed a gap (even when writing the album to a CD), and haven't bought a full album on iTunes since. But QuickTime 6.5.1 played this back flawlessly! I tried burning another CD after that, but for some reason, the gap was on the CD, although I had run out of time to tinker by then and it's possible that there was a burning preference that I has set incorrectly.

It bothers me somewhat that gapless iTunes tracks won't necessarily play gapless in WinAmp. But I still have some experimenting to do: this was all with MP3, and I believe WinAmp can play AAC/MP4 now, so some experimenting with that may be in order. Or, I can always install iTunes on my PeeCee at work and use that instead. In any case, it looks like I may want to re-rip all my CD's so iTunes can play them gapless....

Well, that's that. Any thoughts?

[Mr Scruff]

Just double checked - I'm still seeing gaps just as before. Are you doing anything special to play them back, or just crossfade on with period of 0?

[dreilly1]

Nope, nothing special, just the crossfading like you said.

What hardware are you running on? I did all my experiments last night on a G4 466MHz desktop, and I have an 800 MHz G3 ibook which does the same thing. I don't think the gaps are a function of the hardware, but it's possible, I suppose.

I recorded a few snippets using Ambrosia's WireTap to convince myself that the gap was really there. Maybe if I get some time this week, I'll post them.

[Mr Scruff]

iBook G4. The album I was using to test was Thievery Corporation 'DJ Kicks', which is a mix. Tried ripping from CD tracks 1 & 2 and I can clearly hear the track transition (with AAC & mp3).

[saltines17]

Burning from iTunes is really annoying. It only does DAO when you burn from AIFF-sourced files. I know it's iTunes and not the files themselves, since when I burn Toast DAO audio CDs from Mp3s or AACs or what-have-you, the gaps are eliminated (I *think*)

[dreilly1]

More endless pontificating about the Gap:

I tried ripping and encoding into Apple Lossless format and directly to WAV files. Suprisingly, there was a glitch between tracks when playing either of these back, which I didn't expect, since these are not processed the same way as MP3 and AAC. It's not really a gap -- it keeps the same tempo and doesn't add any time to the playback, but it definitely squashes the volume on the first few milliseconds of the track. If something is going on then (like, say, a quick guitar chord at the beginnign of the song), it can be quite noticable.

Then I did something else which really suprised me: I took the .wav files that I captured through iTunes, and ran them thhrough lame with --nogap. And I could play the resulting files gapless in mpg123. So it seems that iTunes/Quicktime rips raw audio data from the CD correctly, at least. cdda2wav doesn't even do this right without using the Paranoia libraries that are included in the version I have through fink, which makes cdda2wav much slower.

So, it seems that there is definitely something wrong with playback through QuickTime, since there's a glitch when playing raw .wav files from the hard disk through QuickTime, but not when you play the CD directly (which I assume sends the raw audio directly to the speakers, and doesn't use quicktime). I think I'll submit this as a bug sometime soon...

Also, does anyone know whether toast can burn audio CD's of purchased iTunes music (decrypting them through QuickTime, of course). If that's the case, then that might be my solution for burning iTunes albums using DAO...

[saltines17]

Originally posted by dreilly1:
Also, does anyone know whether toast can burn audio CD's of purchased iTunes music (decrypting them through QuickTime, of course). If that's the case, then that might be my solution for burning iTunes albums using DAO...

Toast 5.1.4 can indeed automatically convert m4p files for DAO burning.

I don't know about later versions of Toast, since I haven't used them.

[dreilly1]

Originally posted by saltines17:
Toast 5.1.4 can indeed automatically convert m4p files for DAO burning.
I don't know about later versions of Toast, since I haven't used them.

Hmmm... I wonder if it's possible to write a script or an app which would decrypt the m4p files to wav files (through QuickTime, again, although this might be a good use for playfair/hymn) and then burn them with cdrecord (which I believe can do DAO?)

[dreilly1]

Yes, I know nobody cares about this thread, but I do. So there. Maybe somebody else on this forum is obsessed about this gap as I am, but then again, maybe not. At least the fact that I only have the time to revisit this every two to three weeks indicates that I still have some sanity left.

I found the perfect album to listen for the gap on.

Promise not to laugh?

OK, it's Tesla's Five-Man Acoustical Jam album. Not only is it a live album, but it also is the only Tesla album that doesn't make me want to vomit. It could be due to the fact that about half the songs are covers that they didn't write themselves, though.

Their cover of "We can work it out" melds seamlessly into their cover of "Signs", and there is a strong chord right at the beginning of "Signs" which gets swallowed whenever you play music from the hard drive in iTunes, no matter what format you capture it in (MP3, AAC, WAV, whatever...). You can play directly from the CD in iTunes and it plays perfectly, but import the two songs into whatever format you like, compressed or not, and that telltale chord gets swallowed when you play them back to back.

So, now my task was to figure out whether the problem was with the encoding or the playback in QuickTime. Inspired by that last post, I obtained cdrdao through Fink (I'd only ever used cdrecord before, which I couldn't get audio DAO to work with). I took the version of the songs that I ripped directly to WAV and burned them to a CD-RW in DAO mode, and the gap was gone. Not suprising, since I kind of did the same thing last month with encoding and decoding a wav dile ripped with iTunes.

Now, here's the cool part -- I took the versions that I ripped in MP3 and AAC format, used iTunes/QuickTime to convert them to wav, and burned those wav files the same way I burned the first disc. While the MP3-derived wav file still had a small gap in the playback from the burned CD, the AAC-derived file did not! This implies to me that at least the AAC is encoded correctly, and further points to a bug in QuickTime as the cause of my neuroses. (Well, at least this neurosis). And not even a bug in the decoding, since QuickTime was able to decode the AAC properly, but a bug in the way that QuickTime plays back any set of audio files, compressed or not.

(I'm hoping that someone else who will admit to buying this CD will try it out and let me know if this is true for them, too.)

My next step will be to run one of my m4p albums through hymn, convert the resulting m4a file to wav in iTunes, and see if that eliminates the gap on the burned CD. As you might have guessed already, I like listening to live albums and this gap has kept me from buying iTunes music. If I can eliminate it, I will buy lots more music through iTunes. (Hear that, Steve?)

[tj722]

For what it's worth, I'm obsessed about the gap as well. I can't believe it's 2004 and they haven't figured this out yet, either in iTunes or on the iPod as well. I've been forced to use a program like Toast with Jam to burn a CD without gaps and then listen to that CD on a plain old CD player made with 1970's technology.

[AKcrab]

Originally posted by dreilly1:
(I'm hoping that someone else who will admit to buying this CD will try it out and let me know if this is true for them, too.)

I own this cd. Not that I *bought* it, but I own it.
I'll try and reproduce your bug later.

[toobiloo]

I am a fellow gap hater, and would agree that this is a big issue. I also looking for a good gui for to use th paranoia library. I tried missing media burner, but it won't work with my dvd-r drive.