Reply to thread

Message: [QUOTE="Shamiso, post: 92982, member: 160"] The Yamaha DX7 uses six operators arranged through 32 fixed algorithms, with each algorithm deciding which operators reach the output and which modulate others. Once you understand how FM synthesis works, those little numbered diagrams stop being presets and start looking like signal-flow maps. FM synthesis operators combine an oscillator with level and envelope control. A carrier in FM synthesis reaches the audible output, while a modulator changes another operator at audio rate, so the carrier-versus-modulator distinction comes from routing rather than from two different kinds of oscillator. In [URL='https://goldmidi.com/community/threads/kaito-sato-expanded-imfmsynth-with-dx7-algorithms.78385/'][B]IMFMSynth 1.1's DX7-style routings[/B][/URL], the familiar 32 six-operator layouts sit inside an eight-operator synth that can also use wavetables. The useful mental model is still the same. Follow the path to the output first, then work backward through every operator feeding it. [HEADING=2]Read the chart as audio flow[/HEADING] A DX7 algorithm chart makes more sense when you ignore the algorithm number at first. Find every operator connected directly to the output and treat those as carriers, then trace upward through the operators feeding them. What an FM modulator does becomes obvious once you follow one complete branch from its highest operator down to the carrier you actually hear. Long vertical stacks create nested modulation. One operator changes the next, which changes the next, so a single audible carrier can develop a dense spectrum from several stages of interaction. Parallel branches behave differently because several carriers can reach the output independently, giving you more obvious layered components and more room to tune or envelope them separately. A [URL='https://goldmidi.com/community/threads/melas-matrix-makes-your-first-fm-patch-quick-to-build.76330/']custom FM routing matrix[/URL] exposes the same idea without forcing you to memorize Yamaha DX7 algorithms. You draw the connections directly, which is useful for learning because the routing itself becomes the lesson instead of a number such as 7, 19, or 32. [HEADING=2]Ratios and envelopes shape the result[/HEADING] FM synthesis programming becomes much easier when the algorithm is treated as only the wiring. Frequency ratios decide where much of the new spectral energy lands, while modulator level controls how strongly it appears. Simple whole-number relationships usually keep the result more harmonic, while noninteger relationships are useful for metallic, bell-like, and deliberately inharmonic spectra. The [URL='https://aes.org/publications/elibrary-page/?id=1954'][B]spectral behavior of FM[/B][/URL] includes sidebands spaced according to the modulator frequency, with their relative strengths changing as modulation index changes. You do not need the equations to use the idea. Small moves in ratio or modulation amount can reorganize a sound far more dramatically than the same-sized knob move on a subtractive synth. Envelopes are where basic FM synthesis examples become musical rather than static. A modulator can hit hard at the attack and then decay, creating a bright transient that settles into a simpler carrier tone. A carrier envelope mainly controls the audible level of its branch, while a modulator envelope changes the harmonic complexity over time. FM synthesis feedback adds another source of complexity by sending an operator back into itself. Used lightly, it can brighten or roughen a branch without adding another operator. The [URL='https://goldmidi.com/community/threads/flite-v2s-fm-engine-goes-far-beyond-basic-fm.77727/']ratio and modulation controls in Flite V2[/URL] also show why gain matching matters, since brighter FM settings can feel better simply because they produce more energetic upper-frequency content. [HEADING=2]Fixed algorithms are starting structures[/HEADING] A lot of the reason FM synthesis is so hard is that changing an algorithm can invalidate work you already did. Ratios, envelopes, and output levels chosen for one routing may behave completely differently after the same operators are rewired. One useful FM synthesis trick is to choose the routing before polishing the tone. Start by deciding how many audible components the patch needs. A single carrier with a deep modulation stack suits a tightly unified tone, while several carriers make it easier to build layered organs, electric-piano components, or mixed harmonic structures. What FM synthesis is good for becomes clearer when you think in branches rather than hunting for a supposedly correct algorithm number. Modern engines loosen the old assumptions. [URL='https://goldmidi.com/community/threads/kaitosato-brought-wavetables-to-fm-with-imfmsynth.77021/']IMFMSynth's wavetable-capable operators[/URL] mean the source entering an algorithm can already contain a richer spectrum than the classic sine-wave starting point, so identical routing can produce very different behavior. The algorithm still answers who modulates whom, but it tells you far less about the final timbre. A practical sound-design habit is to mute everything except one carrier branch, set a useful ratio, then bring modulators back one at a time. The same incremental approach works well for broader [URL='https://goldmidi.com/community/resources/how-to-get-started-with-sound-design-basics-for-music-production.84/']sound design basics[/URL] because it separates routing mistakes from tone-shaping decisions. Once each branch makes sense on its own, the full algorithm becomes something you can hear rather than merely stare at. [/QUOTE]

Name