MP3 files are compressed audio files that are made from audio formats such as the wave (.wav) format. Wave files replicate the original source with 100% accuracy at the cost of a large file size, while MP3 files sacrifice some quality for a smaller footprint. How much quality is sacrificed can be mitigated by several factors in the conversion process. With the right bit rate and configuration, MP3 files can provide extremely high-quality results that make them very close to their original wave files when played on portable audio players.

The balancing act between file size and quality is a somewhat subjective. To an audiophile any difference will be discernible. Others might not be able to tell a high quality MP3 file from its original wave source at all. In many cases the difference only becomes clear if played through a high-quality stereo system where the smallest nuances of the acoustic environment become clear.

MP3 files are primarily targeted for portable audio players. In this arena quality MP3 files come through with astounding sound given their small file size. Since portable players have limited memory, it makes sense that people want their MP3 files to be as small as possible while preserving as much quality as possible.

To this end the single most important factor in the creation of MP3 files is the sampling rate. In general terms, the more frequently the original source music is sampled, the higher the quality of the MP3 file and the larger the file will be. A lower sampling rate reduces both file size and quality. The idea is to sample the audio source at the right bit rate to get maximum quality without over-sampling. Over-sampling will result in a larger file size without appreciable difference in quality.

For audio voice recordings such as lectures or language lessons, sampling can be set as low as 32 kilobits per second (kbps), though 64kbps will provide better quality. Voices might sound “flat” at 32kbps, though they will be as understandable as the source file. At a sampling rate of 64kbps the MP3 file should sound nearly identical to the original source file.

Non-saturated acoustic music that features simple arrangements should get good results with a sampling rate of 192kbps. If the music will be played on high quality equipment, you might opt for a sampling rate of 256kbps. Music that falls in this category would include ballads, pop “boy-band” songs, folk and easy listening music. Also the work of many classic artists such as James Taylor, Linda Ronstadt, Joni Mitchell, and Simon & Garfunkel.

To make quality MP3 files from classical music and jazz, the sampling rate can vary depending on the song’s characteristics. Soft jazz can get away with lower sampling rates than can classical music, particularly when the latter includes crescendos of volume and sound. For softer music 192kbps will create a good balance between file size and diminishing returns on quality, though 256kbps will sound better on higher quality equipment. Classical music can be sampled at 256kbps for portable players, or 320kbps for a higher quality result that will sound better on top-end equipment.

For saturated music such as rock, pop, electronic and house music, 320kbps will give the best results. Note that most portable audio players will not play MP3 files sampled over 320kbps.

When possible it is preferable that MP3 files be created using a variable bit rate. This allows the encoding program to determine if a particular frame of music requires the full sampling rate. If not, the program reduces the sampling rate for that frame, resulting in a smaller file without sacrificing quality. Forcing the program to “over-sample” a frame can result in artifacts.

While this is a general guideline, one might find that he or she is just as happy with lower sampling rates for specific songs or in general. Many things affect our ability to judge the quality of music, including not just the equipment we use, but our activity when listening. For those that only listen to MP3 files when exercising or walking outside, for example, exterior noise will make it more difficult to pick out qualitative differences. Conversely, audiophiles might prefer to sample everything at 320kbps, regardless of the equipment being used, the source music, or their activities.

If making your own MP3 files, there are also other settings that affect quality. LAME is an excellent MP3 encoder and is free, along with the many graphical interfaces that serve as a front-end for this well-known command line program. LAME allows the user to tweak many settings in order to produce high quality MP3 files in seconds. One can also try various bit rates on a source file to find the best subjective balance between quality and file size.