New Sound Technology for PCs

New Technology: Music Synthesis
New Sound Technology for PCs
Introduction
Page 1
  Page 2
  Page 3
   Page 4
    Page 5
Conclusion

Wavetable synthesis

The original sound cards were built around music synthesis chips from Yamaha known as the OPL-2 and OPL-3. These chips use an algorithm for synthesizing music sounds known as “FM synthesis.” The FM synthesis algorithm was an important breakthrough because it made it possible to create a wide range of sounds with little computation and memory. Minimizing computation and memory was important in the early days of digital music synthesis because these resources were very expensive. Thus, FM synthesis opened the floodgates of the digital synthesizers that we take for granted today.

Unfortunately, the very simplicity of the FM synthesis algorithm is also a drawback. Although FM makes it possible to create a wide range of sounds, the sounds tend to be “electronic” in character. It was easy to overlook this deficiency at a time when there was no alternative, but the march of progress now makes it possible to offer a wide range of sounds that are more realistic using a music synthesis technique generally known as “wavetable synthesis” (although it is more properly known as “sampling synthesis”). Wavetable synthesis starts with a recording of an actual sound, so it’s easy to understand why it sounds more realistic. For example, to synthesize a clarinet sound, you start by recording a clarinet in a recording studio. The snippet that you record is stored in memory where it can be accessed by the synthesizer to produce the desired output. That memory was too expensive for commercial synthesizers during the period that FM synthesis prevailed, but today it isn’t. Almost all commercial synthesizers and the better sound cards today are based on wavetable synthesis.

Too bad wavetable synthesis isn’t quite as simple as recording a snippet and playing it back. One complication is controlling the pitch of the synthesized output. You could record snippets of the original instrument playing different notes and then assign each to the appropriate key, but performers usually want more complicated control over the pitch for appoggiatura, vibrato, and other smooth “pitch bends.” Implementing these more complicated pitch changes requires digital signal processing (an operation known among the cognoscenti as “sample rate conversion” or SRC for short). SRC could be considered the essential operation of all wavetable synthesizers.

Another important operation is required to give performers control of the duration of the tone: “looping.” Looping simply takes a small piece of the original snippet and plays it repeatedly as if it were a loop as long as the performer holds down the key. There are other operations as well – envelope (including tremolo), dynamic filters, and reverberation – but a description of these operations is better saved for a different article.

Downloadable Sounds (DLS)

An important problem developed in the early days of wavetable synthesizers over identifying the voices. The synthesizer that the composer used might number the voices one way, while the one that the consumer used numbered them a different way. Needless to say, the performance would be quite different from what the composer originally intended! The solution was to agree to a standard palette of sounds with consistent numbering. This palette is known as “General MIDI” or GM for short. GM specifies 128 musical sounds and 47 percussive ones, and voice #1 is Acoustic Grand Piano on every synthesizer. GM went a long way to creating a more uniform musical experience across a wide range of commercial synthesizers.

Unfortunately, GM didn’t quite achieve the acme of uniformity because each synthesizer manufacturer “interpreted” each voice in a slightly different way. Voice #1 is always Acoustic Grand Piano, but each manufacturer recorded snippets of different pianos and processed them in different ways, making each rendition sound a little different. This non-uniformity provoked a revolt among games developers because they want the sounds that come out when users play their games to be consistent across platforms.

There is now a solution to this problem. The DirectMusic API from Microsoft provides support for sounds that the games developers can download to the wavetable synthesizer. Thus, the developer can be assured that the Acoustic Grand Piano will sound almost exactly the same on every platform because it will be based on the same sample – one that the developer provided himself. DirectMusic will also make it possible for developers to use the wavetable synthesizer to support sound effects as well as the musical accompaniment. For example, the game could download the sound of a laser blast once and then trigger it numerous times using MIDI commands. Or better yet, it could download the engine drone of a race car and then pitch shift it in the synthesizer as the car accelerates or decelerates. Thus, DirectMusic not only makes wavetable synthesis more palatable for creating the musical accompaniment, it also provides a new range of opportunities for creating and controlling sounds in a PC.

64-Voice Polyphony

One subtle consequence of the use of the synthesizer for sound effects as well as musical accompaniment is the need to synthesize more voices at the same time. Current synthesizers typically are capable of synthesizing 32 voices at the same time – described in the trade as “32-voice polyphony.” 32-voice polyphony is plenty for musical accompaniment because it’s usually hard to hear more than 32 musical voices. But when some of the additional voices are gun shots, explosions, or engine drone, they are quite audible. Accordingly, the next wave of synthesizers for PCs will generally provide 64-voice polyphony.
New Technology - Software Architecture Next Page