MIDI AT WORK

Back in the day, the concept of presetting a series of notes to play automatically was experimented with by electronic music pioneers like Don Buchla, and admired by Robert Moog (by the way, pronounced ‘m-OH-g’). Moog used a series of knobs and switches to alter the note’s pitch and length with generally 8, 16, 24 or 32 steps, called a SEQUENCER.

It produced a machine-like melody; think of the end of Emerson Lake and Palmer’s ‘Karn Evil 9 – 3^rd Impression’.  Here’s a cool example by a YouTube poster (click here). Very cool indeed, but it had its limitations. 

With the advent of MIDI, the same principles of Moog’s sequencer were used, except the actual performance on the keyboard could be ‘recorded’ into a sequence recorder’s memory for playback. I put ‘recorded’ in quotation marks because the actual sound isn’t recorded, but the same data of which keys were pressed, when, how long, how strongly, etc., etc. Only data was recorded. The sequencer could then be set to play back this data into the keyboards, and could make them do the same thing that the player did on them when ‘recording’ into the sequencer. This is much akin to the idea of the paper player piano roll of yore. The holes in the paper aren’t the sounds, but only the instructions for the piano when to strike the note and how long to hold it out. Therefore, sequencing is the modern equivalent to making your own virtual player piano rolls, without the paper.

The beauty of the MULTITIMBRAL keyboard and SEQUENCER combination is that you have a virtual recording studio at your fingertips. Essentially, the MULTITIMBRAL KEYBOARD is like having 16 different keyboards at your disposal, which means one can have 16 different sounds played simultaneously. Just like previous discussions of separate keyboards and different MIDI CHANNELS, the MULTITIMBRAL KEYBOARD has the ability to set up 16 different sounds, all driven by their own individual MIDI channel. The sequencer is capable of storing the performance of each of those separate parts onto separate tracks. Since it is not recording sound and only the performance, the player can decide to change what sound is used on any of those tracks later on if desired. (Try THAT with audio!!) A new MIDI term crops up here. Since we’re using the same keyboard to access 16 different sound parts, we need to make the physical keys separate from the synth’s sound generating section. When we touch the keyboard, it plays the sounds of the synth as we know, but we want to disengage this so we have control over only the part we want to work with at a time. This control is called LOCAL CONTROL. It can be set for ‘ON’, the normal way the keyboard works, or ‘OFF’, disengaging the keys from the sound making bits. This now allows us to set the transmit channel of the keys to the specific sound part we want to work on, whether it be the sounds of that very keyboard itself, or connected instruments, and record them into the sequencer. Once it’s recorded into the sequencer, the LOCAL setting is no longer relevant because the sequence is taking over the playback, managing what part plays what.

By following the path of MIDI in this diagram, we can see that the performance data from the controller passes into the sequencer where it is stored in memory. While that is happening, the signal passes out of the sequencer back into the controller, passes through the controller into a slave, thru the slave into the next slave and so on. Since the controller, in this case, is also generating sounds, a problematic phenomenon occurs. By touching a key on the controller, we make a sound and generate MIDI data at the same time. That data goes into the sequencer, back out of the sequencer and into the controller again, trying to replay the note. This is called MIDI FEEDBACK.  To avoid MIDI FEEDBACK, we should set the controller’s LOCAL CONTROL to OFF so the sequencer can do the job of playing the sounds.

Up to this point, I have been referring to REAL TIME sequencing. Real-time means exactly that, playing in REAL time. Sequencers can also record in a unique mode as well called STEP TIME, where the performer can enter one note (or chord) at a time, while determining what note pitch and value should be stored in a sequencial pattern. One of my favorite examples of a recording done mostly using STEP TIME sequencing is by Dave Stewart and Barbara Gaskin called Henry and James.

This diagram shows the typical setup of a sequencer-based MIDI studio. By setting different MIDI channels to each tone module and accessing them individually from the controller, recording each part separately into the sequencer, we can create virtually any style of music with any type of sound, depending on what gear you have at your disposal. Once you’ve figured out the nuts and bolts of how MIDI works, you can record away to your heart’s content. I feel it’s pretty important to have a good handle on what’s going on in the tech world portion before getting too involved working on music, because nothing is as frustrating as trying to put down a great idea when the gear is not cooperating. Certainly your creative flow will be stonewalled.

Although this particular entry refers to a hardware sequencer, the principles are identical when we replace it with a computer. Future entries here will deal with computer-based sequencing and recording, opening up an entire new world of music production.

Okay, go hook it all up, figure out what does what, do some experiments first, then dive right in!