Chapter 3: Evolution of Tone Generator Systems and Approaches to Music Production

Spurred on by development of the FM tone generator, Yamaha transitioned its synthesizers to digital technologies during the 1980s, and thanks also to advances in integrated circuits, we released products with a diverse array of new functions. One part of this approach involved the development of technologies that would allow digital recordings of the actual sounds of acoustic instruments—commonly referred to as "samples"—to be used as the tone generator. Drums, percussion, effect noises, and other similar sounds are relatively short, making them ideal for sampling; furthermore, very little adjustment of pitch or timbre is required when these recordings are played back. As such, sampling technology can be applied to easily recreate the sounds of these and other acoustic instruments. For this reason, various different instrument manufacturers began to use sampling-based tone generation—also known as Pulse Code Modulation (PCM)—in drum machines and other similar products during the eighties. Here at Yamaha, we call this type of engine the Advanced Wave Memory (AWM) tone generator.

Yet this type of tone generator was not limited to drum sounds: it could also replay samples recorded from pianos, guitars, and other instruments with longer decay times, as well as the sustained sounds of organs and the like. That said, their main mode of use was as a sampler—that is, a device that replays recorded instrument sounds as–is—and they were not used in synthesizers to actually create sound. Many challenges still needed to be overcome before the Advanced Wave Memory (AWM) tone generator could also function effectively as a synth engine that allowed sounds to be designed in a creative fashion, or as part of a highly expressive instrument for live performance.

One particularly difficult task was the development of digital filters that could behave in the same way as those featured in analog synthesizers. Although the theoretical mathematical formulae that described how these devices operated were certainly well understood at the time, when replicated inside a digital circuit their behavior was much less smooth than that of the analog filter. Developers found it especially difficult to give character to "resonance," a highly distinctive feature of the analog synth. Other instrument manufacturers had already released synthesizers featuring digital filters, but the majority fell short of the mark. Many did not feature resonance at all, while others attempted to simulate this unique filter behavior in an artificial manner.

At that time, Yamaha had actually developed a digital filter capable of reproducing the behavior of an analog one, a feature that made its long awaited debut in our SY77 digital synthesizer in 1989. The SY77 was equipped with both an AWM tone generator and an FM tone generator, both of which could be used together with the digital filter to sculpt sound for remarkable levels of expression. These two new approaches to tone generation were christened Advanced Wave Memory 2 (AWM2) synthesis and Advanced Frequency Modulation (AFM) synthesis, respectively. The SY77 made it possible to create exciting sounds using a hybrid-like combination of sampling and FM, and also featured many other groundbreaking functions—for example, the PCM waves of the AWM2 engine could even be used as operator waves in the AFM tone generator.

The digital filter's cutoff frequency and resonance parameters could also be controlled using the keyboard's velocity and aftertouch, and the combination of all of these features together was referred to as the Realtime Convolution & Modulation (RCM) system. With its smooth digital filter and combination of both PCM and FM—the two giants of digital tone-generation of the time—the SY77 seemed almost too good to be true upon its release, and went on to epitomize the advanced state of synthesizer technology of the 1990s.

Another notable aspect of the synthesizers of the nineties was the development of multitimbral tone generators. Multitimbrality refers to the ability of an instrument to play multiple different types of sound at the same time, a capability that is indispensable in music production.

While not so important to the keyboardist performing live, multitimbral capability does allow different instrument voices such as drums, bass, piano, and solo parts to be played together. However, it really came into its own when MIDI sequencers started to become popular towards the second half of the eighties, allowing a single synthesizer to be used to create a full musical arrangement. Against this backdrop, the multitimbral synth saw increasing use in the production of full-fledged demo tapes and backing tracks for keyboardists to play along with. Although tone generators offering multitimbrality had actually existed from the slightly earlier FM era, interest in this feature increased markedly with the emergence of synthesizers capable of the high-quality, highly realistic sounds of acoustic instruments made possible by the AWM2 and other similar tone generators. At the same time, fierce competition erupted between the various synth makers as they strove to deliver products that were cheaper, could play more notes or voices simultaneously, and offered greater choice in voice variations.

Up until that time, if you had wanted to play multiple voices automatically using a MIDI sequencer, you would have had to fork out hundreds of thousands of yen to purchase enough synthesizers for the number of voices required.

Now, however, a single instrument could effortlessly produce all of these sounds simultaneously. Although various manufacturers addressed this need with tone generator modules supporting eight-part multitimbrality, our TG55 offered excellent value for money with a sixteen-part multitimbral sound system, as did the SY77 and SY55 workstations, revolutionary instruments than can be credited with vastly expanding the number of people producing music.

From the start of the nineties, the hardware sequencers that had been used in the latter half of the preceding decade—such as those of the QX Series—were gradually replaced by computer-based software sequencers. In the recording studio, meanwhile, it became increasingly common to see a vast array of digital devices, such as MIDI interfaces connected to computers; samplers, synths, and other tone generators loaded into racks; and a master keyboard for entering performance data, all connected by a complex web of cables. Using this type of music production system with the computer at its core, a single musician could freely perform and record whatever tracks they wanted to produce a complete song in a very short time. This undeniably raised the bar to becoming a keyboard player, and it was at around this time that the course of development the synthesizer split into two different paths based on what was required of it as an instrument.

The first of these was predicated on the synth being used simply as a tone generator. Certain applications saw tone generators played using wind and guitar synths, but the prevailing trend was for them to be used as an integral part of a computer-based music production system. However, the tone generator would be controlled by the computer or another MIDI signal source, meaning that it no longer had to be integrated with a keyboard in the same way as the synthesizers of the past. To address this, Yamaha released a range of rack-mounted and desktop-type tone generators under the TG designation.

The second path saw the synth develop as a workstation capable of satisfying the requirements of both keyboardist and arranger—a format to which most of the keyboard-type synthesizers of the era adhered. Players could produce full-blown tunes using just the workstation, and with no particular computer skills, make complete productions on this single instrument without any complicated cables, be creatively inspired by studio-level high quality audio. The SY99 mentioned previously brought all of these elements together, and came to be seen as the ultimate in nineties workstations; and the SY Series as the genesis of the Yamaha workstation synthesizer.

The sampling and PCM approaches to tone generation that started to become the mainstay of sound production in synths from the beginning of the nineties have at their core the recording and playback of real instruments. High-quality recording and playback make it possible to replicate the exact same sound as the original; however, developing this into a convincing musical instrument was no easy feat, primarily because pitch and timbre must be controlled in real time in order to make these sounds playable. Consider, for example, that the piano keyboard has 88 keys and the same number of scale notes, MIDI allows the strength of playing to be represented on a 127-level scale, realistic sounds need to change over time and join with one another, and controllers must be able to modulate the sound for greater levels of expression. It is not hard to see that multiple different patterns covering all of these possibilities had to be recorded, and the most appropriate instantaneously selected for playback. This in turn called for huge volumes of recorded data to be processed by memory and processors that remained slow and expensive. In this regard, the technologies of the day left much to be desired.

Although Yamaha's FM tone generation systems could produce highly expressive sounds without the need for much memory, our synthesizer development team embarked on a quest to identify new approaches to tone generation that could create much more realistic sounds, similar to those of acoustic instruments. As a result of this effort, physical modeling was determined to be the most promising candidate.

In addition to the Self Oscillation type VA (S/VA) tone generator used in the VL1, Yamaha also developed a Free Oscillation type—the F/VA—and the VP1 virtual analog synthesizer released the following year was driven by this engine. The F/VA tone generator could model many different variations in the striking, plucking, and bowing of percussion and string instruments, but rather than being limited to simulating existing sounds, it could also model instruments that had never even been conceived of. Although the VL1 and the VP1 synthesizers were very technologically advanced and expressive, a number of different controllers, such as the breath controller, had to be played at the same time in order to perform on them, and this meant that their musicians needed to become highly skilled. For this reason, they did not become particularly popular with conventional keyboard players and remained a niche instrument.

In the 1980s Yamaha established itself as a leader in the development of digital synthesizers with the groundbreaking DX Series. With the arrival of the PCM synth era in the latter half of that decade, we successfully developed the AWM2 tone generator, before moving on to the powerful SY Series. Yet all was not smooth sailing for this synth maker.

One of the biggest factors affecting our business was exchange rate fluctuation. When the DX7 was released in 1983, one US dollar was roughly equivalent to 240 Japanese yen; however, this dropped markedly to 145 yen by the time of the SY77 release in 1989. The dollar had fallen even further when we introduced the SY99 at the end of 1991, dipping below the 130 yen mark. From then until the arrival of the VP1 in 1994, the yen strengthened even further, ultimately forcing the dollar below 100 yen in value.

Back in the DX7 era, we proudly delivered high-performance synthesizers at reasonable prices to customers all over the world, but the rapid appreciation of the yen in the nineties severely eroded the price competitiveness of our products. In particular, the synthesizers that Yamaha had developed as entry level models now found themselves in the mid- to high-price brackets in overseas markets, putting them beyond the budget of the users for whom they were intended.

The collapse of the Japanese economic bubble brought further grief. The period from 1991 onward was marked by rapid economic decline in Japan, and sales of relatively expensive electronic instruments suffered greatly. Other Japanese manufacturers endured similar hardship and faced these difficult times by streamlining their product range, sharing functionality between multiple models, and reconfiguring their lineups with less expensive products.

Although the synthesizer had been born of a desire to grant the musician the same level of expressivity as afforded by acoustic instruments, advances in sampling technologies had now made it possible to produce exactly the same sounds as these instruments with ease. As a result, more importance came to be placed on the synthesizer as an alternative to acoustic instruments than on the functionality that it provided for creative sound design. Furthermore, a number of other developments had also made it easier to compare the sounds of synths from different manufacturers using the same performance data. For example, the General MIDI (GM) standard was published in 1991 for synthesizers that produce sounds in response to MIDI messages, and the Standard MIDI File (SMF) was developed as a common format for exchanging MIDI performance data. Consequently, synth buyers focused more on differences in the sounds produced and suitability to music production than on synthesizer-specific functions and playability.