SMAF and the Format Japan Ran On

The proprietary format that powered the world's first ringtone market — and stopped at the border

What Yamaha Built, and Why

When NTT DoCoMo launched i-mode in February 1999 and opened the chaku-melo market — the world's first commercial paid-download ringtone service — the phones receiving those melodies needed a file format capable of carrying them. Yamaha supplied it. SMAF, the Synthetic Music Mobile Application Format, was a proprietary container standard developed by Yamaha specifically for mobile playback on constrained hardware. Where General MIDI provided a standardised specification for instrument patches across devices, SMAF added mobile-specific data layers on top of a MIDI-like note structure: vibration synchronisation, display text, LED control and handset hardware triggers could all be embedded in a single .mmf file alongside the musical data. A ringtone, in SMAF's logic, was not merely audio but a small coordinated event involving everything the phone could do at once.

The format's performance depended entirely on Yamaha's own silicon. The MA-2 and MA-3 chips — Yamaha's Mobile Application series, fitted into Japanese handsets from around 2000 onward — were the hardware counterpart to SMAF's software architecture. They provided the FM-synthesis voices needed to play the format's multi-channel output at acceptable quality through the tiny speakers of the era. A SMAF file authored for an MA-3 chip could use up to sixteen simultaneous melodic voices; an MA-2 device would play a subset. The coupling of format and chip was tight by design: Yamaha controlled both ends of the chain, which gave Japanese operators and handset makers a polished, predictable product and gave Yamaha a commanding position in the domestic supply stack.

A ringtone, in SMAF's logic, was not merely audio but a small coordinated event involving everything the phone could do at once.

Why the World Did Not Follow

Japan's ringtone infrastructure was the most advanced on earth in 1999 and remained so for several years. The SMAF ecosystem was commercially real, technically coherent and already generating revenue when Western markets were still transmitting RTTTL strings over SMS — Nokia's plain-text Ring Tone Transfer Language, which encoded a melody as a character sequence a phone could parse into a single monophonic line. RTTTL required no proprietary chip and no specialised file container; it ran on the general-purpose processors already inside Nokia handsets. The Infrared Data Association's iMelody format occupied similar territory, a text-based monophonic standard designed as a cross-carrier alternative and readable without manufacturer-specific hardware. Neither matched SMAF's expressive range, but both were open and chipset-agnostic.

That openness proved decisive outside Japan. European and North American operators building their WAP-era download services in 2000 and 2001 sourced handsets from Nokia, Motorola, and Sony Ericsson — manufacturers whose existing silicon did not include Yamaha's MA series. Licensing SMAF playback required both the format rights and Yamaha-compatible audio hardware, a two-part dependency that added cost and design constraint to phones already being specified under intense margin pressure. The polyphonic step in Western markets arrived instead through Yamaha's MA-series chips entering selected non-Japanese handsets piecemeal, and through competing synthesis architectures, with content delivered as MIDI or in operator-specific formats rather than native SMAF.

The result was fragmentation. Japan ran SMAF. The rest of the world ran a patchwork of MIDI variants, proprietary operator formats and, eventually, truetones — actual audio recordings that made synthesis formats of any kind irrelevant. The RIAA's introduction of master-tone certification in 2006 ↗ confirmed that the Western market had moved decisively to recording-based products; the question of which synthesis format had won was moot by the time the certifications began.

SMAF's commercial life wound down with the synthesis era itself. As i-mode handsets gave way to smartphones capable of playing back full audio files, the format's hardware dependencies became liabilities rather than assets. Yamaha's format had been precisely right for one market at one technological moment: the world's first paid ringtone economy, running on proprietary chips in a closed but brilliantly organised ecosystem. That it did not spread globally was not a failure of the format's design but a consequence of the incompatible standards choices that had already been made everywhere else before Japan's lead could be leveraged into a universal specification.