MIDI - Section 2, Signalling and Hardware
Return to MIDI Introduction and
Contents
Return to Studio Technology home
page
This section of the lecture describes MIDI at the level of firstly, the communications or signalling specification, and secondly the interface hardware and electronics. This section will be of interest if you are electronics engineer who may be concerned with either designing or troubleshooting MIDI hardware, or if you are an advanced MIDI user then the information contained here will help you to understand some of the limitations of MIDI. If, however, you are concerned with MIDI only as a user, or programmer of MIDI equipped instruments then you may safely skip this section and go straight to the MIDI Message Structure.
The MIDI specification v1.0, requires that a MIDI interface must operate at 31.25 Kbaud, plus or minus one percent, using asynchronous serial communications with 1 start bit, 8 data bits, no parity and 1 stop bit; one byte thus requires 10 baud for transmission. MIDI signalling is similar to the serial RS232 or V24 specification, in that signalling is active-low, and bytes are transmitted LSB first, and MSB last. MIDI signalling is based on the current-loop principle, with a logical 0 represented by current 'on' (of nominally 5mA), and logical 1 represented by current 'off'. The diagram below shows the timing for a single MIDI byte.
2.2 MIDI Interface Hardware and Electronics
The specification requires that one MIDI output must be connected to only one MIDI input, and that the MIDI input must be opto-isolated. The use of opto-isolation means that equipment interconnected via MIDI remains electrically isolated (unlike audio separates, for instance); a serious electrical (ie power) failure in one piece of MIDI equipment is thus most unlikely to cause any damage in another even though the two are connected via MIDI. The use of opto-isolation means that not even the earths (grounds) are connected. The MIDI cable interconnecting one MIDI OUT to another MIDI IN establishes a current-loop between the two devices, as illustrated below. The MIDI signalling switches the current through the LED in the input opto-isolator on and off, and the signal is optically transferred to the photo-transistor in the opto-isolator. In order to preserve the accuracy of timing data the MIDI specification asks for the opto-isolator to exhibit a rise-time of no more than 2 micro-seconds (which is quite a fast device).
Most fully equipped MIDI instruments (synthesizer workstations, for instance) have three MIDI connectors: IN, OUT and THRU. The IN port is self explanatory, it is the (opto-isolated) input for MIDI data. The OUT port is the output for MIDI data which has been generated by the instrument. The THRU port does, however, require some explanation, for it is also a MIDI output port, but one which outputs a direct (buffered but otherwise unprocessed) copy of the MIDI signal received on the IN port. THRU and OUT are both outputs, but what distinguishes them is that THRU simply presents a duplicate of the MIDI IN signal, whereas OUT is MIDI data actually generated by the instrument. In order to understand why the THRU port is so essential we need to consider that a single MIDI connection can carry data for multiple instruments, which must be daisy-chained together using the IN and THRU ports (because MIDI is not a multi-drop 'bus'), so that each instrument receives the same MIDI data (from a sequencer, for instance) but then 'picks out' those MIDI messages intended for it alone (how this works is explained in section 3 below).
The diagram below shows the relationship between MIDI IN, OUT and THRU ports.
The block diagram below shows a typical MIDI setup with a PC, MIDI controller keyboard, synthesizer module and drum machine. Carefully note the labelling of the MIDI ports, and the direction arrows for MIDI data.
