Ok...
That will require pages and pages of explanation.. let me see if i could compress it ...
To understand this, we first need to understand how all the music we listen to is actually recorded in a studio.
A few years back, all this music was recorded on analog magnetic tape, these were usually capable of carrying 24 individual tracks ( meaning 24 individual instruments and sounds not a song track )
Music was known in the multiple of 24 , like 24 /48 / 72 etc, depending on the number of 24 track machines used while recording.
After each track is recorded , the mix was played and recorded on another machine onto a 2 track stereo tape.
Now a days digital recording is taking over, although digital recording depends on the mixing console, computers and softwares... Analog type recording is still an audiophiles favorite as digital recording will sample a sound wave many times per second allowing an illusion of solid sound waves to be created..... in contrast, analog tape captures a sound wave in its entirety, this attributes to the harshness in the higher frequency bandwidth. Where as an analog recording would be clean and warm.
CD quality audio has a sampling rate of 44.1KHz no frequencies above the Nyquist frequency of 22050 Hz are acceptable for recording to avoid aliasing. Very steep sloped LPF are applied to limit the frequencies above 20KHz. This introduces distortion into the audible range.
Uncompressed audio as stored on a cd has a bit rate of 1,411.2 kbps = (16 bit/sample × 44100 samples/second × 2 channels / 1000 bits/kilobit).... not 128kbps as i had mistakenly written above, excuse me for that.
at this point we also need to know .. Bit depth and Bit rate ...
"Max Lighting" Defines
""Bit depth is the level of data per sample being stored. A recording encoded at 16-bit, means that every sample can store any one of 2^16 (or, 65536) bits per sample.
Bit rate is how many bits of data per second.
Therefore, the bit rate has an upper limit defined by the bit depth.""
So now we know how already a Digitized format stands different from an analog one.
Moving ahead.
MP3.
MP3 in its basics used lossy data compression depending on the algorithm models developed on human perception of hearing.
It is such that it reduces ( deletes/removes) data on the CDAudio.
Imagine stuffing a huge piece of equipment in a smaller box, you would have to dismantle the equipment, try and adjust it inside the box and also may have to omit a few spare parts to shut the box.
MP3 compression is done by removing or reducing certain sections of sound that are beyond the audible spectrum of most humans. This is called perceptual encoding.
Psychoacoustic models are developed to reduce or discard information that would be scarcely audible to human auditory senses.
Different MP3 codecs , encode using different perceptual algorithms. Hence we also notice a difference between different files encoded by different codecs.
Compression efficiency of different encoders are generally defined by the bit rate, because compression ratio depends on the bit depth and sampling rate of the input signal.
A 192 kbps compression would give a bigger file, also it would remove lesser data from the given sample as compared to a 128kbps mp3 file and hence the reproduction would be better too.
Again there is CBR and VBR ... constant bitrate and variable bitrate, in CBR the bitrate remains through out the entire piece of data recording, where as in VBR the decoder reduces the bitrate wherever the audio data content is less, meaing in portions of silence or lesses music, the bitrate is reduced. this helps in making the file smaller.
There are some lossless compression codecs too, but these are not entirely lossless formats like the AAC, FLAC , vorbis etc... although they are better compressions than MP3, they give bigger files too.
-------------------------------------------------------------------------
Just a small experiment you could do whenever you are free..
To see the margin of error increase...
Take your favourite track ( on an original audio cd )
convert it using lame decoded into 128kbps mps file. ( note )
now burn that mp3 on your computer, onto an audio cd format again. ( note )
now use that audio cd to convert into another 128kbps mp3.. do this not more than 5 times... and you will most evidently notice the diminishing quality in sound reproduction on the same audio system.
Bottom line,
Technically and also almost audibly , an audio CD format is the closest to the sound recorded in the studio. ( I said closest, not identical ) ... so further compressions and conversions will definitely lead to data loss and deteriorated quality of reproduction...
Hope this helps...
Fenil Mehta
Sound & Acoustic Designs.
( Personally, I still love listening to audio tapes in a simple 2 ch stereo setup )