#3 Maths behind MIDI
MIDI has been a world-standard protocol for decades. It has been blamed for a lot of troubles throughout all this time. Vince Clarke of Erasure still stays CV and Gate and it sounds great. But many artists sound real tight and groovy while using MIDI. Suppose deep and funky feel of William Orbit or spectacular chill-out of Afterlife. Almost every synthesizer being produced today is midi-compatible, as well as many cool-sounding vintage synths of 80s. Manageability wins.
- synth with CV / Gate input, the most common format before MIDI
To keep it clean you need to know the rules before you play the game. First, your beat clock has to be super-stable in time. Any hardware or software-induced jitter (variations in time) destroys the feel of your track. Second, as your MIDI OUT is a bottleneck that allows you to send only one message at a time, you need to sequence your messages respectfully and understand how it works.
- Prophet 600 with MIDI interface
The message.
The resolution of MIDI clock is 24 ticks per quarter note. It means that the receiving device gets a message on every 64th triplet note. This value is not absolute - it’s a function of your sequencer’s tempo.
One midi message contains the number of channel, note and velocity. Velocity ranges between 0 and 127. You can send additional continuous commands using midi controllers. For instance output volume is normally a controller #7. You can program coarse volume changes with this controller. (0-127) If you want to fine-tune your volume, you have another controller #39 that sets 127 subdivisions of each position of #7. Any change in controller #7 automatically resets #39 to 0.
Controller numbers from 1 to 32 range are called MSB or Most Significant bits. If you take any of them and add 32 to its number, you’ll get same controllers that allow fine tuning on top of that. Those are LSB or least significant bits. As I already mentioned normally LSB is being reset to 0 with any MSB change.
The logic is like with an old radio tuner - you have coarse tuning to quickly find the spot and fine tune to get accuracy.
The moment.
The stability of the internal clock is one of the most important features of a good sequencer or a drum machine. A famous artist once told me: “When I first heard the TR-909 I didn’t feel that it’s something special sonically. But when we [the band] started to play along to it, I understood that it’s really cool and powerful”.
-Dave Smith with his TEC Award for Musical Instrument Technology
Fairlight CMI’s spectacular sequencer is responsible for many big electronic acts of 80s and 90s like Peter Gabriel, the Art of Noise, Yello, Pet Shop Boys, Seal and many others. MPC2000’s gained popularity among Hip-hop artists and it’s various offshoots not only because of it’s sampling possibilities but also because of the sequencer - it gives a solid foundation for MC’s and break-dancers. Old Otari computers with early Cubase software are still praised by electronic musicians as awesome sequencers as well.
Now that we all moved to advanced operating systems like Mac OSX and Windows 7, started using USB instead of Serial port, software sequencers became more complicated. It’s not always easy to find the source of errors that may alter your timing accuracy and ruin the feel of your track.
- Fairlight CMI’s graphical user interface
There are a few solutions for this: you may go retro and install a PCIe-to Serial slot and hook an old MOTU or Unitor 8 interface. You may use a good hardware sequencer like Sequentix or an MPC station and program everything there. Third solution is to use your audio playback engine (any DAW of choice) to generate sample-accurate beat clock using Innerclock Systems hardware and software, which to me seems an ultimate answer to midi jitter problem. In order to get versatility you may combine all three approaches and have a whole new look at your synths and groove-boxes.
As long as you know you are stable, you will change your approach to programming grooves and sounds. Getting an immediate musical response from your actions is one of the many overlooked subtleties in the digital age.
Latency.
It takes time to transfer the message to a device and generate the sound. For each unit it’s slightly different - that’s why you have to compensate for the latency on a track-by-track basis. The most simple way to do this is to play few midi notes and record audio, then measure the difference between the MIDI message and audio signal. This difference in samples or milliseconds will be your offset value for the track. When you calculate and dial in all the offsets you get a tight and stable MIDI setup. This is even more important when you daisy-chain your devices using MIDI thru.
- USB MIDI-controller
The designer of MIDI, Dave Smith recently noted that you can barely find any other protocol that survived through 3 decades under version 1.0 and still rocks. He launched the standard when mass production in synthesizer industry was taking off. Then it locked-in because of increasing returns on scale. Now MIDI is like QWERTY - a universal standard that everyone has and knows.
Ilya Mazaev, 2012
Check also: