Why MIDI music files are lighter than MP3 and OGG?

[Sub-Zero]

Title speaks for itself, just curious is why Midi got lower size

[Commander Makara]

Okay well first MIDI is vastly dissimilar frtom OGG and MP3 so let's start there.

MIDI is based on synthesised instrument sounds which are given a pitch, duration and loudness. These sounds are positioned in a duratin space so on 'playback' they are 'played' in the right way at the right time, mimicking an orchestra of instruments playing together to create music.

For short pieces with few instruments, this can be asmall filesize. With longer pieces and more instruments, the size increases exponentially.

The quality is largely determined by the quality of the instrument synthesis on the device which replays the sound. Voices and natural sounds are typically poor in synthesis.

Other than this difference in the instrument themselves, though, Midi files are lossless representations.

__

MP3 and OGG actually capture the comined signal of the sound. If, for example, whilst recording, a person coughed in the background, their cough could be captured and digitised.
Both of these formates utilise the principles of Fourier analysis to encode and compress in a lossy fashion, but the entire sound range from 20Hz to 20kHz (may be reduced if no high/low freq are present) is considered within a sampling resolution. Essentially this means the number of "instruments" playing at any one point in the duration space is represented by a single value (or dual value for stereo channels of course) the evolution of the waveform identifies the different occpations of different frequencies along the duration.
Some other optimisations such as MDCT are applied in some cases.
Therefore for an OGG or MP3 to have particulurarly different filesizes is mainly due to:

1) Sample frequency (The number of datapoints within a single period of the duration space) - This is the main one!
2) Bit depth (The number of bits used to represent the amplitude)
3) Bandwidth resolution (The actual number of frequency divisions within the overall range)

A higher sample frequency provides lower Quantisation Noise and a more accurate synthesis of the orignal sound, but greatly increases the size of the file. Typically 44.1kHz sample frequency is sufficent for general use - though much of the compression algorithms tend to reduce even this for frames containing less complex sounds to maximise compression ratio even further.

[Marcy]

I would like to add that MIDI has nothing to do with sound. It holds information of which note occurs where, and the velocity and length of said note.

MIDI files can be played directly from any device that has an audio processor capably of doing so. The sound card in your computer is one of them. Usually, the better your sound card, the better the sound assigned to each voice of the MIDI file will be.

MIDI files can be loaded into a synthesizer or DAW where a user can automatically (or manually) assign instruments to each MIDI channel. Depending on the instruments or sounds you've selected you'll get a good representation of the intended sound.

Again, MIDI has nothing to do with sound. It's note information, it doesn't do anything musical on its own.

[tiseno100]

MIDIs are mostly instructions so your player can perform some XYZ tones. Thats why you were able to change the instruments etc etc on these old Roland players.

Mp3's are audio data