A CD-R can store either 650MB or 74 minutes in my experience. Seventy four minutes of music is only about 74 MB though, so how does the means of storing info vary from audio recording to data recording? Why can you store many hours of music on a Data CD that can only hold about 70 minutes of recorded audio music, and why does a 32 bit recording take up as much space as a 192 bit recording on an audio CD even though it takes up 6x less space on a Data CD?
Are you asking about compression schemes by any chance? I believe you are. MP3 is the most famous compression scheme but there are others. It is a lossy compression scheme which means that bits that are inaudible, redundant, or don’t contribute much to the overall music quality are stripped out leaving a much smaller file than the original.
Music stored on a regular CD are uncompresses and therefore the file size per song is very large.
I think that is basically what you are looking for. Let us know if we need to elaborate. There are tons of lengthy threads about music compression in the archives however.
Probably has something to do with the different formats that people use to record their songs in: mp3/wmv vs. wav (or whatever format audio CDs come in). 74 min of music in mp3 format takes up about 74 MB, but a whole lot more in the “regular audio CD” format.
Audio CDs are not compressed at all, and 74 minutes of uncompressed 16-bit 44.1kHz audio does take up about 650 MB.
When you burn an audio CD from MP3 files, the MP3 files need to be uncompressed first. So it doesn’t matter how much it was compressed to begin with.
The other answers are essentially right but the lossy compression that is used means that you have a lower fidelity recording. Not all of the parts that are thrown out are inaudible. An audio cd has about 10MB or so per minute for a stereo recording. .mp3’s are much, much smaller. I don’t know why they don’t use a lossless compression. You would still get a pretty good compressions rate without effecting the fidelity of the music.
I’d like to know what kind of lossless compression you’re contemplating. A quick search shows that most lossless compression schemes give less than 2:1 compression ratio. I’ve got broadband and nearly .5TB of storage, but I’m not gonna download your 50M demo track!
Well, people still use .jpeg when .tiffs and .pngs are available. It’s a standard that hasn’t been supplanted by superior versions yet.
Also, .mp3 is a variable compression system, so it’s your choice whether or not to use 64, 128, 256 or whatever you want. VBR is also another choice. 128 is popular because to the untrained ear (mine included), it’s very difficult to hear the loss and the file size is most managable.
Just wanted to add a quick note.
I found a great article about compression on How Stuff Works a few weeks ago, and I thought i’d share it.
Speak for yourself. When I rip one of my CDs, the tracks are compressed into .oggs, not .mp3s. Here’s a FAQ on the codec.
OK, there’s some truths and half-truths being flung around this thread and I just wanted to break my four month long “SDMB boycott” to set everyone straight:
CD audio recording works by “sampling” audio data 44,100 times a second. When these individual “samples” are played back at the same rate they were recorded, the illusion of music is created on your stereo or CD player - in much that same way that the illusion of movement is created when the individual frames of a movie are played back in the correct order and the correct speed. Because the “standard” for CD audio (also called the “Red Book”) is 16-bit stereo, we can extrapolate 16bits x 2 channels = 172bits per second = 10,320 bytes per minute, or around 10MB per minute of CD audio.
This audio is recorded is PCM format, which stands for “Pulse Control Modulation”. PCM and WAV - the standard Windows sound format - are, for all intents and purposes, the same format. Thus, all of the sounds Windows makes and all of your audio CDs are the same format - although most Windows sounds tend to be mono (and thus require half the data) and have a lower sampling rate - that is, they use sounds typically recorded at 8000 samples per minute (kHz) as opposed to the 44,100 required for CD audio. Using mono and a lower sampling rate results in a smaller file than you would expect if you were using CD audio.
As has been mentioned, MP3 removes a lot of extraneous data from a WAV\PCM file of the same song. On average, MP3 reduces file sizes by 10 or 12 to 1, so you get a much smaller file size at the cost of reduced fidelity.
As others have mentioned, many people can’t tell the difference between an MP3 and the CD audio source. For this to happen, two things come in to play: bitrate and the quality of the encoder.
As you might have noticed, MP3s can be encoded at various bit rates. 128kbps used to be the “default” bitrate, although 192kbps is rapidly becoming the new “broadband standard” (if it’s not there already). In any case, all you need to know is that a file encoded at 64kbps will take up half the disc space as one encoded at 128kbps, although the quality will be half as good as the 128 one. The smallest bitrate I’m aware of (32kbps) is typically reserved for speech and/or “samples” of music like you find on Amazon and other online sites. You can fit (roughly) 40 hours of music at 32kbps on a data CD-ROM, although the quality will sound even worse than a homemade crystal radio. On the other hand, you can minimize your MP3 compression to 320kbps and still get around 5 hours of music on a data CD, but “purists” will say that you’re still forsaking quality for space.
Secondly, the quality of the encoder is quite important. Most all-in-one music apps - like Windows Media Player and MusicMatch - do a decent job of most things, but a crappy job of encoding MP3s. All other things being equal, I’d prefer a 128kbps LAME encoded file (LAME is widely regarded as being the best MP3 encoder, bar-none) over a 192kbps (or higher) MusicMatch or RealJukebox encoded file.
Other things mentioned in this thread that need mentioning:
1 ) Ogg Vorbis sucks. If you REALLY want to save your music in a lossless fomat, try FLAC (google it). FLAC will compress WAV files to 3/4 or 1/2 the size of WAV files but can still be “uncompressed” to their original WAV fomat with no loss of quality.
- Audio CDs are indeed compressed, although this compression is minimal compared to typical audio compression such as MP3.
I’m really not following the math here, though the end result seems to be about right (a 650MB CD could hold 64 minutes if it was 10MB/minute)
Let’s see here. Don’t you multiply the bits by the sampling rate by the number of channels? 44,100 x 2 x 16 would be 1,411,200 bits per second, 176,400 bytes per second, or 10,584,000 bytes per minute.
Do you mean compression in the sense that you don’t get exactly the same waveform out as was put in when the CD was made, or do you mean that some sort of data compression is done on the disc?
FLAC certainly deserves mentioning in the context of this thread. However, you seem to be suggesting that Ogg Vorbis is also a lossless compression format, which is not true. Vorbis is a lossy format, just like MP3. It’s main claim to fame is that it’s an initiative from the open-source community, to create an audio compression format which is unencumbered by patents and can therefore be freely implemented as open-source software.
Could you elaborate on why you feel that Ogg Vorbis sucks? I’m quite happy with it, but admittedly I don’t have “studio ears”. What makes it inferior to MP3 according to you?
I’m quite sure this is not true. 74 minutes of stereo 16-bit sound at 44100Hz is
74 * 60 * 44100 * 16 * 2 = 6265728000 bits = 783216000 bytes = 746.9 MB
which is pretty much the maximum storage capacity of a CD. It’s a little more than the effective capacity of a data CD, but I believe this is because data CDs have some extra error correction.
Where does this minimal compression come in, and what algorithm is used for it?
Trivia:
And of course we are aware that a CD contains 74 minutes of music because Norio Ohga of Sony, having studied opera in Berlin, decreed that a CD should be able to carry all of Beethoven’s Ninth Symphony (generally 70 minutes). The size of the CD itself was dictated by shirt pocket sizes - they were made slightly too big to fit to reduce theft.
Rex, don’t you know that CDs `suck’ and that the only way to listen to music is vinyl?
I mean, come on. We can go over which formats and media `suck’ until the cows come home and record a note-for-note copy of “The Rite of Spring” and it will mean jack shit. Until you come up with something objective, I’m going to go with what works best for me.
Calm down dude. I like my oggs too, but consider how many music download services provide music on mp3 and how many provide oggs. Also, what proportion of portable music players play oggs and not mp3s? Consider this too: my Windows Media player plays mp3s, but not oggs. AFAIK, this is the most common (though Winamp does play oggs).
While imperfect, I don’t think that it’s out of line to call mp3s the standard.
Agreed. Its other main claim to fame is that it is a second-generation audio codec (WMA and MP3 Pro). These codecs try to be more intelligent about what they toss and what they keep so that their quality per bit ratio is better.
Another thing about Ogg Vorbis versus MP3 which may be relevant to this thread:
With MP3, you select a bitrate, and the sound will be encoded at exactly that rate. So if you choose 128kbps, and then encode a sound file containing nothing but silence, it will still take 128kb per second – meaning you get very high quality silence. 
On the other hand, if you have a very complex piece of music and try to encode it at a too-low bitrate, the codec will be forced to throw away information that is audible to the human ear.
With Ogg Vorbis, on the other hand, you don’t choose a specific bitrate, or at least it is not guaranteed that the sound will be encoded at exactly that rate. Instead, you specify a quality level from 0 to 10. While it is possible to roughly predict the bitrate you will get for a given quality level (e.g. level 3 is about 110kbps for a typical music sample), the exact bitrate is not constant and will vary with the complexity of the sound being encoded.
Having a fixed bitrate can be advantageous for some applications, such as streaming audio, or for the audio track of a video CD or DVD. However, Ogg Vorbis proponents claim that for any given (average) bitrate, Vorbis will give better quality than MP3 or MP3 pro.
Since “better quality” is neccesarily subjective, it’s hard to say who’s right – and to make matters worse, some forms of distortion will actually sound “more pleasing to the ear” to some people than an exact representation of the original. However, it’s safe to say that for either MP3 or Vorbis, at 128kbps the vast majority of people will be unable to tell the difference from the originial, uncompressed CD data.
I don’t think anyone here has actually addressed the OPs question–why do audio CDs hold less music than data CDs that use compressed formats?
Part of it goes back to the original design specs of the CD. No compression was used because the decompression hardware would have made CD players too darned expensive. They chose a very simple format, that worked well, and it was quite an improvement over the amount of music you could store on vinyl.
You could fit more on a longer cassette format, but the thinner tape was subject to stretching, and you could fit a lot more CDs in the same physical space (and weight) as a pile of cassettes.
This is true with fixed-bitrate (FBR) MP3 encoding, but you can use variable-bitrate (VBR) MP3 encoding instead and avoid this problem. LAME supports VBR nicely.
One downside to VBR MP3s is that many players will show the MP3 length incorrectly; a five minute song might show up as 48 minutes. This is because the MP3 header does not include the song length, but only the encoding info. Players then use the encoding info to estimate the song length. This works for FBR MP3s, but not for VBR MP3s. Some rippers will mitigate this problem by including the song length ID3 tag and some players will display this info instead.
I stand corrected, thanks.