Conversion Methods
People working with MIDI music often ask questions like: "How can I record my MIDI music onto CD so I will be able to listen to it by an audio CD player?" or "How to convert MIDI files into MP3 files to be played back on players capable of reading MP3?"
Unfortunately, when you are trying to record MIDI files CD-burning programs give a message "Incompatible file format". Audio CD players are designed for reproduction of digital sound in WAV format, while in MIDI files there are only commands about reproduction of the notes by different instruments but not the digital sound. If you want to record MIDI compositions on audio CD or to convert them into MP3 files, you need to first create an audio file in WAV format that will contain digital variant of the music that we hear when playing back the MIDI file. There are two ways to do it using your computer equipment.
First, that is the so called method of sampling or rendering the music. In this method, the WAV file is being synthesized from the small samples of already digitized sounds of real music instruments. The very fragments of already written sounds performed by different instruments are called samples and the sets of such samples are called banks.
Banks of samples are furnished either together with the sampling program or they can be bought separately. Since these banks contain sound samples of hundreds of notes performed by tens of instruments, these banks usually use many megabytes of sound data. One of the well-known banks format is Soundfont by Creative Labs.
The sampling principle boils down to the following: a program consequently reads MIDI commands from MIDI file and according to their content put the samples into the Wave file.
For example, having read a command to play the note C performed by piano during three seconds, the program will record in a new file consecution of the digital data corresponding to the sample of the note C in piano performing. If needed, this sample will have digital transformations necessary for that fragment gets required duration and volume. Then comes next command and generates next fragment. In fact, the process is much more complicated since simultaneously there may sound several notes by different instruments, thus, before recording a consecution of data into WAV file the program needs to mix the fragments.
The rendering methods use a very complex mathematical algorithm of the sound synthesis, and some of them give a really high quality sound, provided that you use only high quality sample banks that may cost more than a hundred dollars. The new audio file is being made without a computer sound card. The advantage is that the result does not depend on quality of your sound card.
Most sound cards of the 1990's used a so called FM-synthesis (frequency-modulation) or small sample banks and the outcoming sound was mechanical and "nonliving". The modern sound cards already contain high-quality banks and sampling algorithms and in order to get sound better you must be able, at least, to tune the sampling process and then to know what banks to use.
If you are satisfied with the MIDI performance of your sound card, you may use another method of producing the audio file. The method is to record sounds performed by the sound card during MIDI playing back. It works at principle "Recording what is heard". Applying this method, you may be sure that audio file will sound exactly the way as the starting MIDI file. The disadvantage of this method is that its application often requires additional tunings in the sound card mixer.
In the Midi2Wav Recorder the problem is solved by running a test of the computer sound equipment during the first trial of the program. The program is playing back MIDI notes and simultaneously recording them into the Wave file. Then the received data are being analyzed and it chooses the optimal configuration of tunings for mixer and sound devices. |