Bit Rate
CD media
We speak about AudioCD, not about MP3 collections: these are two absolutely different things. Audio CDs offer reference sound quality, because they capture everything without any clipping or compression. What concerns MP3 collections, they are just compressed CD tracks recorded on a CD. Music in AudioCD format has strict technical parameters, which cannot be changed. That's why it's possible to calculate the total length of a disc - 74 or 80 minutes (an album may be shorter, of course). You can copy music from a CD to your computer, but it's not very easy. In order to save your music to a hard drive, CD tracks should be encoded into any audio format. You can do it with various programs - for example, Windows Media Player, which can be found in most computers. The audio quality will depend on encoding options. The higher audio quality, the larger your audio files will be. However, this is already the next point :)
Uncompressed media
Digital data from a CD can be read 1:1 and recorded as a WAV PCM file. As no details are lost, this operation is fully reversible - you can record a CD from your files. It will be almost the same or absolutely identical (if you didn't break tracks into files) as the original CD. The advantage of this format is that it's lossless. However, from the point of view of portable devices, such files are too big: one minute of music in this format requires over 10 MB of disc space. Besides, WAV support is not available in all players: HDD players tend to support this format, while flash players usually don't.
Lossy compression
There exist methods to compress music, when some audio data are lost. For example, MP3, WMA, OGG, AAC, etc. It's these formats that are supported by all modern players. But this support is limited. For example, any player supports MP3. WMA is supported by 3/4 of the modern devices. However, only 30-40% of modern players support OGG and AAC.
Lossless compression
An encoding program can compress original audio from a CD several times. However, the file size is relatively big - a usual CD can be compressed twofold. However, original audio is not clipped. While lossy formats just sound like the original audio CD, lossless formats provide original sound, which is compressed to take less storage space. There are several lossless codecs and formats. The most popular is the format used by Apple - it's supported in all iPods but shuffle. You may also come across such formats as FLAC and APE. These formats are especially promising from the point of view of their top audio quality - they are no worse than WAV PCM, but they take up 50% less storage space and they support tags (we'll touch upon them below).
Regardless of a format, music files differ by compression quality. The main parameters here are bit rate and sampling rate. Bit rates vary much in music files. The higher the bit rate (all other things being equal), the higher the audio quality. The highest audio quality is provided by bit rates starting from 128 Kbps and higher. Lower bit rates are used only when audio quality is not critical, for example for voice recording. Besides, it's agreed that 320 Kbps hardly differs from the original audio track. However, it's a subjective factor, which depends on a given person, ear for music, and sensitivity. By the way, the bit rate of audio CDs is 1411 Kbps. As we can see, there is a great difference between compressed and uncompressed audio. Audio file size grows proportionally to its bit rate. The lower the bit rate, the smaller the audio file, the lower audio quality, and vice versa. Let's describe the notions of bit rate and sampling rate.
Bit rate shows how much data is required to encode one second of an audio file. So, when the bit rate of a file equals 256 Kbps, it means that each second of this audio file takes 256 Kbit of memory. So you can calculate the size of a music file, if you know its length and bit rate. Another definition of this term is a bitstream. It's often used to describe audio microcontrollers. It's easier to understand this term from the point of view of interface throughput: how much bits are processed (go through a given interface) per second. It's impossible to say which bit rate to use to encode audio - it's a subjective matter. One user will be always satisfied with 128 Kbps, the other will grumble about 320 Kbps.
Sampling rate—it resembles video resolution: the higher it is, the more details you will see. What concerns sampling rate, the higher it is, the more distinctly a file sounds. So the higher the better. The optimal quality is usually provided by the sampling rate of 44100 Hz (it's often called 44 kHz). This value is not chosen at random - according to the sampling theorem, a sampling rate of a signal must be at least twice as high as the rate of the source signal (if you know physics well, don't write angry emails to us, we simplified it deliberately ;)). As a human ear distinguishes sounds from 20-30 Hz to 20-22 kHz, we should use 40-44 kHz for digital audio processing. The standard of professional equipment is 48 kHz - with a little safety margin. But when the AudioCD standard was created, its authors did not want to compete with professional equipment, so the sampling rate was cut down to 44 kHz. Computer formats support even higher values. But as most music tracks are copied from AudioCDs, there is nothing surprising about this value being the widest spread sampling rate.
This theorem gives only a necessity criterion, not a sufficiency criterion. That is, we know for sure that if we don't provide the twice as high sampling rate, the audio quality will definitely be bad... But we don't know whether it will be good, if we do. And we don't know the exact sampling rate to ensure good audio quality :) That's why the 48 kHz sampling rate is used in DVDs and professional audio equipment, and HD audio (for example, SACD or DVD Audio) supports sampling rates of 96 kHz or even 192 kHz, because even 48 kHz may be insufficient. However, there is no need to support these excesses in portable players :)
What does it all mean in practice?
Here is a ballpark statistics for a 4.5 minute audio file. We'll show you how many files your player can store depending on the quality of these files and capacity of the player:
Quality and size of an audio file/ player capacity
512 MB
1 GB
2 GB
4 GB
8 GB
16 GB
128 Kbps - 4.3 MB
120
240
475
950
1900
3800
192 Kbps - 6.48 MB
80
160
315
630
1260
2520
256 Kbps - 8.64 MB
60
120
235
470
940
1880
320 Kbps - 10.8 MB
45
95
190
380
760
1520
The speed at which a player can save data into its memory is not a key factor, but it's still important. The higher the speed, the faster you can copy files to memory. The usual speed is 2-5.5 MB/s for flash players and up to 15-20 MB/s for HDD players.
Prices for flash memory are constantly decreasing now. When you choose between identical players with different memory capacities, you should keep in mind that twice as much flash memory does not mean twice as expensive player, because the player itself also costs some money. So you should make up your mind whether to pay a little more for double memory capacity or save it.
8 bits = 1 byte
1,024 bytes = 1 Kilobyte (Kb).
1,024 Kb = 1 Megabyte (Mb).
1000 Mb = 1 Gigabyte (Gb).
Modern computers can store anywhere from 5 to 100 Gbs on their hard drives. Why would you need to worry about the file size of your digital audio with so much space available? Well, your computer needs a good chunk of its hard drive to store the operating system (Windows, Linux, etc.) and instructions for running things like your modem, your internet browser, and your email program. So space is often at a premium. You’ll see.
