Appendix; Appendix Xg; Chapter 0: Format Overview - Yamaha MU90R Owner's Manual

APPENDIX

X G
Chapter 1: Format Overview
Yamaha introduces a new tone-generator control format designed to meet the requirements of
the coming multimedia environment. The new XG format — an extension of the existing GM
format — provides broader capabilities suited to the demands of an increasingly sophisticated
and diversified computerized environment. The new format enables a significantly higher
level of musical expressiveness while at the same time ensuring the continued compatibility
of existing sound data.
Yamaha shall use the XG format as the basis for forthcoming electronic instruments, music
software, and tone-generating LSI circuitry, while working to maintain compatibility and
scalability among Yamaha models.
Development Background
Tone generators are utilized in a wide range of devices, from musical instruments to commu-
nications devices and computer games. The first international MIDI standard arose from the
need to enable consistent external control of tone generators on all device types, regardless of
manufacturer or model. Because tone generator voice arrangements tended to vary con-
siderably among manufacturers and models, however, different MIDI devices often produced
different types of sound in response to identical MIDI instructions.
In 1991 the MIDI standard committee enacted additional specifications, referred to as the GM
(General MIDI) standard, for the purpose of standardizing voice arrangements and improving
MIDI uniformity. The GM standard significantly enhanced acoustics compatibility among
complying devices, leading in turn to an expanding base of GM software applications. But
the GM standard also has its limitations. It provides support for only 128 voices, whereas
many users now perceive the need for a greater number of voices suited to a wider range of
musical genres. Users have also indicated a desire for greater control over voice modifica-
tions and effects so as to enable a higher level of expressiveness.
The advent of the computer-based multimedia age has added yet a different perspective,
bringing increased attention to both image and sound technologies. Developments in multi-
media-related sound and music processing parallel recent advances in the area of image com-
pression, and are pointing the way to the future of multimedia.
At present there are two fundamentally different approaches to the handling of sound and control
data. One method is to digitally store sound data at the software side together with the control
data, then send all of the data together to generate the playback. The second method is to have
the software supply just control data to a tone generator on a computer or other such device. The
tone generator processes the incoming data and generates the sound locally.
The first method offers highly realistic sound, but requires immense quantities of data and
locks in a specific set of performance characteristics and voices. The second method requires
far less data while allowing for entirely free variations in voices, tempos, and virtually all
other performance characteristics. The second method is therefore ideal for interactive multi-
media applications such as karaoke and repetitive computer-game sounds. MIDI-based appli-
cations are typical of this second type of approach. As multimedia technology advances, we
face a pressing need to expand this approach to accomodate a larger number of voices and
132