General:
MIDI: Music: CD Production:
|
Advanced MIDI Concepts
MIDI is actually a lot more powerful and flexible than the GM standard would imply. Below are a few things to think about if you're serious about using MIDI for music. Before we get started, we need to standardize on vocabulary, because things can get confusing quickly.
What MIDI does MIDI is an ordinary digital file that is interpreted by a device as a list of commands. The MIDI protocol provides for general commands that help make global settings on a device, such as patch/channel assignments. Then there are clusters of commands for each note or sound produced, influencing everything from volume, duration, attack, release, channel, etc. Note commands are designed to be sent and interpreted in real time, which is different than some digital protocols which must be assembled and timed after the fact. There is also provision in the protocol for commands only understood by a specific device, called "system exclusive" or SysEx MIDI messages. These can serve to uniquely configure that device. The protocol is serial - that is, it's a single stream of data which could be sent along a single wire, much like a modem uses a phone line. Although MIDI is largely standardized, that standardization is not maintained by a governing body such as the Institute of Electrical and Electronics Engineers (IEEE, which approves and standardizes computer input/output interfaces), but rather is an agreement between manufacturers to always do MIDI communication in the same way, so that devices and sequencers of different manufacture can work together. The protocol was originally intended for use by computers over the Internet, but it is now more commonly used by keyboardists connecting their various "instruments," and by recording artists and composers. MIDI was well-conceived, but not without problems brought on by increased demand on the protocol. Most of these problems have been addressed with hardware solutions (see below). There have been a few attempts to extend the protocol with additional recognized commands (most notably Yamaha's XG), but they remain pretty much exclusive to the manufacturer who designed them. Embedded in the MIDI standard is a list of standard patch-number-to-"instrument" assignments, called General MIDI or GM. This is an attempt to insure that different devices will reproduce MIDI files in the same way. Devices that adhere to such a configuration display the GM symbol. Since the GM protocol provides for no bank change capability, only 128 different patches are possible, and hence GM is fairly limited. In addition, adherence to GM by manufacturers has been reasonably good, but certainly not perfect, so there are differences between how devices handle GM files. What a MIDI device does MIDI devices used to be true synthesizers, that is, they took the basic building blocks of music - timbre, pitch, volume, modulation - and manipulated them electronically to mimic traditional instruments, or produce entirely new sounds, i.e., they synthesized sounds from artificial parts. Today's devices are a little slicker - they actually store mini-recordings of instruments in read-only memory, and then process and combine such recordings (called samples) to produce sounds. Such an approach is easier and cheaper than complex synthesis, and a lot better at mimicking traditional instrumental sound. However, since there are limits on processing speed and storage capacity, and since real instruments played by real people produce nuance and complexity that is very difficult to quantify, MIDI devices are still compromises. Most offer a fantastic number of patches, but do so by maximizing use of the stored voices it has through combination and modification. As mentioned above, MIDI does a lot more than simply send "Note on and note off" messages to a device. The protocol provides for attack speed, release speed, the possibility of changing volume or modulation (vibrato) over the life of the note, stereo pan, and a variety of common sound effects, such as echo and reverberation. The designers of MIDI were pretty smart, since many of the technologies that the protocol can reference didn't exist yet. The MIDI devices of today have surpassed the protocol in capability, but the MIDI standard is not in danger of being replaced. Getting the most out of your device Timbrality: Most synthesizers are at least 16-part multitimbral, meaning they can receive and play data from all 16 MIDI channels at once. Since 16 is the MIDI standard, any device that is more than 16-part multitimbral has more than one MIDI connection point, and is, therefore, essentially two devices. In general, the only way to add channels/patches (that is, to gain extension to multitimbrality) is to add devices. In the world of MIDI music production, this growth is usually aimed in two possible directions: specialty devices (synth modules dedicated to specific families of sounds, such as orchestral instruments, the Hammond B-3, drums, analog synths such as the Moog family, etc.), or samplers with RAM-storable sample sound libraries. The latter is the most expensive and the most flexible, but neither is cheap. There is an exception. As the capacity of personal computers has continued to grow, "virtual synths" - synthesizers residing completely in computer memory - are getting better and more powerful with each passing year. It won't be long before a "device" can simply be a plug-in to a MIDI sequencing program - a collection of sounds stored in memory and triggered by MIDI data routed to it. In such an arrangement, the limit on timbrality is whatever the hardware will handle, which can be much greater than 16. Polyphony: A good general rating for a synthesizer or synth module is 64-part polyphony, though 128-part is not uncommon amongst the better devices. This means that such a device can play 64 voice/pitches simultaneously, at any given moment in a composition. One must be careful, however, to not expect too much, since this refers to simultaneous voices and pitches. Remember that a patch - an "instrument" assigned to a specific channel - may have several voices, and hence a single note triggered in that patch can represent more than one contribution to polyphony. As an example, many synth modules have electric piano patches that sound very full. This fullness is achieved by assigning several different electric piano voices to the same patch - sometimes up to 4 or 5. Hence, a simple 6-note chord triggered on such a patch could represent 24-30 voice/pitches, or almost half the capacity of the device. What does this mean? It means that if you trigger three such patches with the same 6-note chord at the same time, a few of the pitches in one of the chords will not be played. Synth devices are quite smart, and try to do the best they can with the data they receive. Each MIDI voice/pitch command causes a sound to play. The pitch will continue to play and be heard until the MIDI command for it to terminate is received, or it decays naturally. If, while playing a MIDI sequence, there are 64 pitches playing when a command for a 65th pitch/voice is received, the oldest pitch is terminated to make room for the new pitch. Often this is not noticeable, but sometimes it is. Drum sounds add a lot to polyphony, and if the device truncates a cymbal decay to make room for another voice/pitch, it can be very noticeable. Another end result of operating near the capacity of a device is latency - there are hiccups in the rhythm as the device attempts to process a large number of MIDI commands at once. It is usually polyphony, and not multitimbrality, that cause electronic musicians to pull their credit cards out of their wallets. Even a good orchestral composition will only rarely use more than 16 parts at once, but it is unlikely that a MIDI device will be able to handle all 16, since they represent many more voice/pitches than 16. However, adding devices also gains more possible patches and voice combinations . . . Adding Devices, and Beating the MIDI 16-channel limit The 16-channel limit applies to one MIDI device. If you have 16 MIDI devices, you could technically have 256 available channels. But you have to get the data to all of the devices, and all at the same time. MIDI can actually be daisy-chained - many devices offer "Thru" connection points, through which received MIDI information is mirrored and sent to the next device in the chain. However, you must be able to specifically address devices in such a chain, or all devices will play all information. Addressing specific hardware is possible, but adds to the total MIDI data, and the possibility of latencies. In addition, not all sequencers can do this - Noteworthy cannot. It is also possible to induce devices to ignore specific channels, but using this method reduces the channel count back down to 16. A better method is by increasing the number of MIDI connections on your computer. This can be accomplished, on a limited scale, by adding sound cards. However, the best method is by using a MIDI patch bay. Patch bays take computer MIDI information through one of the other, faster I/O interfaces, such as parallel ports, USB, or Firewire. They then distribute the various MIDI channels to different MIDI output points at the exact timing moments required to make sure everything plays together. Such devices can even be used chained together to further increase device count. MIDI patch bays cost between $150 and $1000, depending on how many MIDI connection points each offers (the cheaper ones usually handle two). When the patch bay drivers are installed on the computer, each MIDI output patch on the bay represents a different MIDI device in Windows, and so the sequencing software can route any channel to any point on the patch bay. It works quite well. Keyboards vs. Modules Devices come in two categories - ROM-based (called synthesizers, though they're frequently just samplers with permanently-installed voice libraries) and RAM-based (called samplers, which will receive and store sample libraries in temporary memory). The former is further divided into keyboards and modules, the only difference between the two being the existing of a physical keyboard. Many MIDI composers are keyboard players, so they produce their MIDI music "live" by playing their keyboards and producing MIDI data in real time. This data can be saved and played back later, just like a tape or CD recording, except that the data can be easily modified for creative purposes. Synth Modules, in contrast, have no keyboards, hence they cannot produce their own MIDI data (except in limited ways). They have the advantage of being small, quite versatile, and frequently expandable. If a musician intends on writing parts by hand because they are unwilling to be limited by their own playing skills, modules are very much cheaper for the same sound capability. Even keyboard players use modules as a method of expanding the live sounds they can produce, since only one keyboard is really required to produce the MIDI data which then can drive the other modules. Some modules even have the capacity to store and play back MIDI files. Most cannot realistically be used as sequencers, since they rarely come with the ability to directly edit MIDI data. MIDI Keyboards, in contrast, frequently come with this capability, and often include disk drives or memory sticks to save and load MIDI files. The Drum Module - a whole 'nother device Electronic percussion devices were popular quite early on - perhaps because of the importance and complexity of rhythm in popular music. Drum modules are rated and used differently, since one patch really represents 128 different "instruments," all triggered by different "notes" of the scale. In most MIDI devices, Channel 10 is reserved for percussion. Drum modules, being dedicated to percussion, often will accept MIDI information on other channels, all of which will be treated as percussion. Also, since popular music percussion is quite repetitive, drum modules can be used to construct complex rhythmic parts, by playing on pads either connected or on its surface, or step-wise through an interface on the device. The resultant MIDI information can then be stored internally, or played back to a sequencer and recorded for later use - in much the same way that MIDI keyboards can be used to produce and store MIDI information. Some drum modules can be used to store and play back entire MIDI sequences, thereby becoming the MIDI player as well as one of the devices. |