Rip Your CD's to MP3 files!

"Trade in" Your CD's for an iPod!

[Download] the encoder package

[Watch] Flash movies showing, step-by-step, how to do the job superbly!

Do you have a big stack of CD's that you never seem to play anymore? Do you do all of your music listening on MP3's these days? Do you lust after Apple's spiffy iPod (or other similarly-priced hard disk player)?

Well, you've come to the right place, because I was once like you. Over the past year, I've been on a quest to come up with the be-all/end-all of converting all of your CD's to MP3.

The two things that had been holding me back, up to then, were that I still didn't have an MP3 player which could hold all of the files once I had converted them, which means that I still couldn't play my MP3's in all of the places where I could currently play my CD's.

Then I realized that, having 250 CD's, if I could sell my CD's for a mere $2 per CD (on average), I'd get $500 for them... which would pay for a 40GB top-o-the-line iPod! In fact, after selling almost all of my CD's, I averaged closer to about $3.25 for them. This means, that, if you had as few as 100 CD's, you could rip them, sell them, and have about $300+... enough for a 15GB iPod. The 1,000 or so songs from your CD's would take up about 4GB of that, leaving you with room for 11GB of stuff you've downloaded! In fact, if you bought one on eBay, you could still get a 10GB iPod, and those go for around $210, which means that you could have as few as 70 CD's and still break even!

The process is a little involved but, once you learn it, it really goes pretty quickly. In about a week or so, you could have all of your CD's ripped, converted, expertly tagged, normalized, and be on your way to selling your CD's and being part of the modern age!

Although there are tools out there which claim to do a lot of the steps all-in-one, I've steered clear of them for three reasons. First, I wanted to be able to pick the best tool for each stage in the process. For example, ExactAudioCopy (EAC) is believed by many to be the best CD ripper (digital music extractor) that you can get. However, I don't want EAC doing the MP3 encoding for me and I also don't want it doing my normalization and tagging, either (for reasons which I'll get into later).

The second two reasons The second reason is that I feel that it's best to keep the ripping process and the encoding process separate for the sake of the user. You see, ripping goes fairly quickly and requires user interaction (to eject the old CD and insert the new one). Encoding takes a longer time, but does NOT require anything from the user. With the all-in-one tools, you insert a CD and wait. The tool rips a track, encodes it, rips the next track, encodes that, rip, encode, etc. This will take about 10-15 minutes on most machines. Now, on the other hand, if you could just rip a bunch of CD's at once, you could go through a CD in about 3-5 minutes. Also, ripping doesn't bog your machine down, so you can keep using your machine while it's going. Then, you could leave the lengthy (and CPU crunching) encoding process for when you go off to bed, or to work, or to eat or something.

The third reason to split the ripping and encoding is because you might want to make multiple encodings from a single ripped wave file. For example, you might want to encode a wave file to a normal MP3 and also to an OggVorbis file. Or, maybe you might want to encode to an MP3 of nominal quality (like 160k) and then one "archival-quality" one at, say 320k. Whatever the reason, I've written a program which will take a single wave file and encode it into a variety of compressed formats for you.

So, I've chosen to do each step individually. Now, this all works best if you do a bunch of CD's at a time. So, you rip a whole bunch, then encode, etc. If you do them one-at-a-time, it will become qutie cumbersome. So, try to free up as much space off of your hard drive as possible. You'll need about 500 megabytes of space to rip each CD, and then another 50-200MB of space for the encoded files (depending upon which formats you encode to). So, for example, if you can free up about 5 gigabytes, you should be able to do about 7 CD's at a time. Needless to say, this might be a great time to go get that new hard drive you've been thinking about getting. Some 80GB drives (as in... being able to rip 112 CD's in a session, and then encode them) are going for as cheap as $40-50 these days.

Now, the whole process is broken down into five steps: Ripping the raw sound information to wave files, encoding the raw wave files into compressed audio (like MP3's), tagging/organizing the sound files, normalizing the files, and then selling your CD's on eBay.

Ripping the Raw Sound Information to Wave Files

For the ripping program, I'm going to use ExactAudioCopy (EAC). It's very highly regarded and it hasn't failed me yet.

Encoding the Raw Sound Files to Compressed Audio

For encoding the files, I used the best MP3 and Ogg encoders I could get my hands on. The MP3 encoder is the "LAME" encoder, and the Ogg encoder is the "oggenc" encoder. In spite of its name, the LAME encoder is far from lame. It's one of the finest encoders out there.

Now, throughout my CD ripping escapades, I've had a need for a variety of different encodings. For the music CD's that I rip, I like to have an MP3 version, an Ogg version (In listening tests, Ogg usually scores higher than MP3 for the same file size), and a super-high-bitrate archival MP3. The archival one is because I'm selling the CD's. If some better encoding format comes along someday and I want to have all of my albums in that new format, I can't just re-encode my "everyday" MP3's into the new format, because that would defeat the whole point of trying to go to a better-sounding format, since anything made from those everyday MP3's can't ever sound better than the MP3's themselves. So, the "archival" 320kbps MP3's are meant to serve at the electronic embodiment of the CD's themselves, and I keep them on a spare hard disk in my closet. However, I also rip my audiobooks from CD to MP3. Since audiobooks are only spoken-word stuff, I find it sufficient to encode them to a much smaller size.

So, to make all of this easy to do, I wrote a little program that just goes through a folder full of wave files and it runs a variety of encoders (which you can alter) with a variety of options. As it completes each set of encodings for a wave file, that wave file is moved to another directory to indicate that it has been processed.

This program, and the two encoding program (the LAME encoder and the oggenc encoder) are available off of this webpage. Go to the top and click the "Download the encoder package".

Tagging the Files

If you've ever downloaded any MP3's with peer-to-peer software, you probably know how low most people's IQ and standards are. MP3's can contain extra information in them which lists the artist, the album it came from, the track number from that album, etc. Many MP3 players examine this information to let you search for a given artist, album, etc. However, with many MP3's you get off of the net, this information isn't even there. Just as bad are cases where the individual who made the MP3 didn't even know enough about the artist or song to name it properly. For example, it's common to see Bob Seger's last name spelled "Seeger", "Seigar", or "Seegar". George Thorogood's last name on MP3 files is spelled "Thoroughgood" more often than the correct spelling! And, in spite of how often he yells it out in the song, there is no song by the who entitled "Teenage Wasteland" (the real name is "Baba O'Riley"... and it's not "Babba", either). Slightly less annoying is the fact that some people like to capitalize most of the title ("I Put a Spell on You"), while others like to only capitalize the first word, as though it were a sentence ("What is and what should never be").

Now, if our MP3's are going to completely replace our beloved CD's.... this substandard and non-uniform naming just won't suffice. We need something which will tag all of our MP3's with the best information that's obtainable.

Enter MusicBrainz (www.musicbrainz.org). MusicBrainz is going to save the world. Okay... maybe just your world. MusicBrainz reads your MP3's, calculates a audio "fingerprint" of the music therein, and checks their database to find the proper artist name, album, track number, and track name. MusicBrainz has well-documented guidelines on naming style (like capitalization, use of any punctuation in the name, etc) which ensures that all of your MP3's will be consistently and properly named and tagged.

It can also organize all of your MP3's into separate directories based upon artist and/or album.

Normalizing the Files

You may have noticed that some of your MP3's are noticably quieter than others. This isn't so much of a problem these days as it was, but it brought to light the need for something called "Normalization". Normalization is a process where the music file is examined, and the volume is adjusted to make the listening level of the song closer to that of the rest of your songs.

The first normalizers (and most that you can still get today) did this by finding the loudest point in the file, and adjusting all of the information in the file so that this loudest point was now 100% of the loudest that an MP3 file could represent. In other words, it cranked every song up to "10".

There are three problems with this method. First, due to the wierd way humans hear sound, making all of the MP3's numerically equivalent like that doesn't yield songs that sound the same volume to humans. There's a whole field of study around this kinda stuff called "Psycho-acoustics", and that has to be incorporated into any sensible normalization plan.

Second, most normalization programs would alter the actual sound data. In other words, a normalization program would read the MP3, decypher it into the sound data, normalize it, and the re-encode it into an MP3. This is not a good plan. If you decode, alter, and re-encode MP3 (and other compressed audio) files too much, you could start getting noticable defects in the sound. Ideally, we could adjust the volume without having to alter the actual MP3 data at all... leaving it pristine. The solution to this is an idea called ReplayGain. With ReplayGain, a normalization program just makes a little notation at the beginning of the MP3 file, telling the player to adjust the volume (or "gain" as they say) up or down for this file. Since it happens when the file is replayed, they call it ReplayGain. Most players now support this feature. WinAmp5 does. Last I heard, the iPod does not, so we need to keep bugging them. Anyway, we want our normalization scheme to use ReplayGain.

The third problem with most normalizers is that they look at each file individually. I personally encountered the problem with this method when I was listening to an album that had a mixture of loud and soft songs on it. On the CD, I grew accustomed to the ebb and flow of the loud and soft songs on the album. However, when I turned the CD into MP3's and normalized them individually, the soft songs were all just as loud as the loud ones. Not good. So, we need a normalization system which, if possible, can normalize an entire album as a unit, to preserve the "album-ness" of the songs, if you will.

Enter MP3Gain and VorbisGain. MP3Gain is a program which does all of these things. It uses psycho-acoustic models to estimate how loud a human would perceive the song to be, it tags the files with a ReplayGain tag (leaving the encoded audio untouched), and it considers whole albums as a unit. VorbisGain does the same thing for OggVorbis files.

Selling Your CD's on eBay

Coming soon...