Main Site | Join Robin Hood Coop | Projects | Events | Blog | Media | Forums | Mailing List | Twitter | Facebook

Vintage music synths in one cheap integrated circuit

Old analog synths from 1970s and 1980s are getting fame again. Their price rises as used gear but they are not manufactured anymore. Analog integrated ICs are made at cheap price nowdays. So many of those old music gear that people want but can t have can be build again in one “collection” integrated analog circuit. Even VCOs (non- digitally controled DCOs) can be used, VCOs need heating element that rise temperature to such that VCOs are stable, but IC is so small so that heating does not use much electric power. Old Curtis and SSM chips are being made again so replicas of old gear can be build with them, but there were lots of synths that did not used them, mono synths etc. gear. Those Curtis / SSM chips were 8 micron tech made and about 2X2 mm sized. If analog IC is 1 micron, its size is 1/64th of those, so even very small analog IC can have several synths (100 voices? 100 different synths? Either different mono synths or different polysynths, or mix of mono / poly synths). manufacturing of that chip does not cost much. With A/D converter such chip can be made as Micro SD card. Micro SD card has 1,88 watt peak power limit at very short time but even continuous power requirement of those 100 voices (100 different monosynths) or 100 different polysynths is perhaps so small that SD Card can handle it (and phone too). 0,5 micron made would be 256 times smaller than Curtis / SSM IC using same circuit. And USB stick can have inside this synth chip also and sold at cheap price. USB stick can have analog additional output so that real analog sound can be used.
Envelope generators and LFOs can be digital, nobody will notice difference are they analog or not if they are in “virtual analog” imitation of analog type mode. Outside software can then be used, like digital / virtual analog synth LFO / ADSR / envelope modules of modular digital / virtual analog software synths, together with real analog hardware, altough miniaturised form in analog IC chip. Even sort of modular analog synth can be build, with hundreds of modules, in one analog IC chip. Connection points in IC are used building “vias” inside chip that can connect outside chip, so it will be like miniaturised modular synth, perhaps made using 1 micron or 0,5 micron tech. Outside this small IC is large patch panel like usual modular synth, cables connect modules, from modules goes wire connections to vias of IC chip.
Even digital synths can be collected to one IC. Yamaha DX7 had 4 micron IC, nowdays in china telephone SoCs are made with 40 nanometer tech that cost few dollars. It is not worthwile to build one Yamaha DX7 chip as one IC, instead it can IC that collects all digital synths of 1980s that are copy free. Digital synths use copyrighted waveforms (romplers), but in 1980s they used synthesis methods like FM / phase distortion, additive and subtractive synthesis. Patents of them are expired. First romplers appeared in 1984 (Kurzweil), 1987 (Roland digital synths) and Korg M1 in 1988, their waveforms are copyrighted perhaps to someone person, so it is lifetime plus 70 years before they become copy free. So in year 2100 about is legal to copy them. But FM / phase distortion synths, digital additive / subtractive synths are copy free. Their factory presets cannot perhaps be used but their hardware can be copied. Freeware presets are however plenty. Synclavier, Fairlight Mark 3, Crumar GDS / Synergy etc can be all copied in one cheap digital IC chip that costs almost nothing, about 50 different digital synths can be in one few dollar IC, including those super expensive synths of the 1980s. Even scientific systems like Sogitec 4X, SYTER, The Katosizer, Interim DynaPiano etc can be in one IC, Technos Acxel, Lyre FDSS etc.
Tchnos Acxel and other synths that used Motorola 5600 or other DSPs can be group to another integrated circuit, so another Micro SD card or USB stick is needed for them. Synths that use same kind of DSP can share the DSP, for example Fairlight and Acxel used same DSPs so they can bots share it in IC circuit. Scientific systems in academia used for sound synthesis DSPs also (mostly). Also analog and digital ICs are different so analog synth collection is in one Micro SD card / USB stick and digital synth collection in another, and digital DSP - based synth collection in third. Manufacturing costs are very small, so price is cheap. Digital synths become all romples from 1989 / 1990 onwards to this day, so they cannot be copied because they are based on copyrighted waveforms, not on some synthesis method like FM or additive synthesis that most early digital synths used. Digital hardware can be copied, digital software is copyright protected up to today? There are examples like Munt that makes legal emulation of Roland synths, but that is not accurate copy and is not the real thing. Analog synths can be copied if they are 15 years old (15 year copyright for hardware) or 20 years if patented. From Minimoog to Novation Bass Station all famous analog synths can be put to one cheap integrated circuit.
Electric organs and string machines are not made any more because they require top octave generators and divide down circuits that are not manufactured anymore. Same goes for ensemble keyboards (Moog Opus, Arp Quadra etc. “analogue music workstations” as stated in sequencer de netpage). Last combo organs were made about 50 years ago. All those can be brought back, in one integrated circuit. Analog synth collection chip can include not only synths, but also string machines, ensemble keyboards, electric (transistor) organs, combo organs, electronic pianos, and analog studio rack effects like filters that were made during 1990s / early 2000s (Filtrex etc). All in one cheap integrated circuit: “best analog gear in the world, all that you can dream of and more, all that you have ever wanted but they don t exist anymore, or are too expensive for you nowdays, all in one chip and with cheap price”. When people think about Hammond organ they think tube valve Hammond but Hammond made transistor organs also. Last real analog organ manufacturer Lowrey ceased trading 2019, there are no analog electric organs then anymore, all are now digital rompler / sample player type fake analog thing.
Not only analog keyboards but analog guitar / bass pedals from 1970s / 1980s or beyond can be brought back in one integrated circuit, that is cheap, hundreds of guitar pedals in just one pedal that is cheap. Guitar synths also, before they turned romplers also.
When synth IC is inside Micro SD card or USB stick, it needs graphical user interface. Together with synth chip in SD card and USB stick is memory chip, cheap, that contains GUI for using that gear that is in IC. So in every phone and in every home PC can have all classic analog and digital music gear at cheap price, it is like playing softsynth, but it is not software emulation, it is real analog or digital synth. User actually plays real synth or electric organ, not software emulation. But what about analog sound, does Moog still sound Moog if it is made to 1 micron or or 0,5 micron analog IC? Analog IC manufacturing quality is very good today, and cheap, so does that manufacturing quality transforms to good sound?

This synth on a chip, which is a collection of 100 or more separate analog synths, either monophonic or polyphonic synths, can have additional features in every of those 100 or more synth voices. Additional oscillator and filter can be added to synth voice. There are “simple audio oscillator” and “simple audio filter” types with very few transistors, resistors etc. It does have not same sound and sound shaping ability than original synth VCO (DCO) or VCF, but at least synth voice now has additional VCO / DCO and filter. Also because ADSR, envelopes, and LFOs are digital, their amount can be increased from example from just one, to 4, 5, or 6, and those LFOs, ADSRs and envolopes can be routed to modulate many different things, unlike some old 1970 / 1980s analog synths that had for those only one modulation route. Also each synth voice can have additional analog features that use very few transistors, resistors etc, like “simple wavefolder” ,“simple aural exciter”, “simple sample & hold”, and simple bass / treble control for each voice. Modern analog synths have usually digital effect section (phaser / flanger / chorus / echo). This synth chip can have for each synth voice this effects section as digital also, or global (all synth voices use the same) analog effect section which uses bucket brigade circuits for flanger / echo etc. Also synth chip can have parametric equalizer and subharmonic generator (synthesizer) that can be global and analog, because those circuits use much components (transistors, capacitors, resistors etc.). Also each synth voice can have own digital sequencer, simple type 4- 8- or 16 step etc. Perhaps it is not possible to use VCOs so VCO synths must be changed to DCOs. Digitally controlled oscillators need digital logic, so if they need 16 bits for economy reasons, instead of integer 16 bits floating point, posit or quire number system can be used. This needs special digital logic designed specially for this analog synth chip. So this 1 micron or smaller analog synth chip needs additional digital logic chip for control duties, digital chip can be made of 40 nanometer etc. tech so it is much smaller than analog chip. Both fit inside Micro SD card together easily. DCOs can use also 24, 32 or even 64 or 80 bit floating point if chip budget allows that. Also digital CPU can be used, instead of specially designed digital logic. CPU uses software so it is not “hardwired” like fixed digital logic. So changing programs change digital synth features, but digital CPU can be power hungry. Something like Intel Quark that uses very little power can be used and put together with analog synth chip to MIcro SD card. This CPU can also have GUI programs for user interface so they don t need to use phone or PC CPU. But GUI needs memory so also (small and cheap) digital memory chip is needed, so Micro SD card has three chips inside: analog synth chip, and for control duties digital logic chip and small memory chip. All three fit easily inside Micro SD card. If USB stick is used instead of memory card power budget and space restrictions are no more. Graphical user interface can be connected to for example PCs own softsynths, so PC softsynth oscilllators can be routed to real analog synth filters in USB stick synth for example, virtual analog modular synth modules of PC programs can be used with real analog synth hardware of USB stick synth. Synth voices can be monophonic voices grouped to polyphonic mode, for example 6 Minimoogs inside synth chip can be used as 6 - voice memorymoog, 4 Roland Promars as one 4 - voice Roland Jupiter 4 etc. just like polyphonic Oberheim synths used individual monophonic synths as polyphonic synth - principle (X-mod), and Rhodes Chroma used something similar. There can be well over 100 synth voices, either different synths or some of the synth voices, or some of them 4 - 8 or more as duplications of same synth voices, like 4-8 voice polyphonic synths. All 100 or more synth voices and dozens of different analog synths can be used as one gigantic synth using graphical user interface. And then electric (transistor) organs, for example Hammond organ models are restricted to transistor models only, not tonewheel or tube valve models, then synth chip can also include combo organs, string machines, ensemble keyboards (“analogue music workstations” of the 1970s, that had in one keyboard monosynth, string / brass machine and electric organ), and analogue rack effects like filters etc. All this in one cheap synth chip, costs few dollars, and contain all old vintage gear that musician ever wants for, in cheap Micro SD card or in USB stick. Guitarists can have just one guitar pedal that has inside all famous analog guitar pedals of the past, hundreds of them, all in one integrated circuit, and guitar synths also. Woodwind players can have wind controller chip in cheap plastic wind controller, this wind controller has all (few) old analog wind controllers in one chip, perhaps digital wind controllers also if they use FM or additive synthesis etc.and are old designs.

Earlier there was phone sound synth chips, used for ringtones, although those chips were in many cases complete synth chips. Manufacturers of those were Yamaha, NEC, Oki, Micronas, Macronix, Vimicro, Rohm, and Atmel SAM / DREAM. Many phone sound chips were wavetable / ROM chips, but Yamaha SMAF and NEC chips both had 128 voice polyphony and those were real synth chips, with sounds from iconic Yamaha and Casio synths. Those chips had very low price, few dollars, and offered to everyone who wanted to buy them, not like real synth chips that are used only in manufacturers own devices. Also customer could use own synth wavetables and ROM waveforms. Free software emulation of Yamaha DX7 took about 7 years of work and international team of computer programmers. At same time those programmers had phone in their pocket that propably had Yamaha sound chip, hardware Yamaha DX synth, sound of it could be used out of the box, no software emulation needed, few dollar price. NEC phone sound chips had Casio CZ inside. Many people buy sample packs, hundreds of dollars priced, with DX and CZ sounds. At same time in their pocket was phone whose sound chip contains those sounds in hardware synthesizer, not just sampled sounds, and it is cheaper too than buying a sampleset. Now those phone chips are not made anymore, Yamaha quit SMAF some time ago. But if someone orders again those phone sound chips they will be back in production. There are two keyboard synth chip manufacturers who offers sound chips for sale, DREAM, and also EMU 8030 is offered for sale still, although EMU ceased operations. Perhaps Ensoniq OTTO is for sale in somewhere still. But very few customers use those chips although price is low. Some odd chinese keyboard module or guitar pedal uses them and other very few manufacturers, although EMU 8030 in that module got rave review because of its iconic EMU synth sounds. Those very cheap synth chips can be put to Micro SD card and then in every phone can have its own Yamaha DX synth or Casio CZ, changing SD card changes synth from Yamaha to Casio for example. Wavetables / ROM waveforms / samples are user defined so different for example softsynth waveforms can be used in those chips or samples from famous hardware synths. Some small memory chip is also needed in same SD card that has GUI for this synth. Its like softsynth in phone but it is hardware synth in phone. Same chips can be put to USB stick so PC can have those synths also. DREAM and EMU 8030 are real keyboard synth chips, not phone sound chips, and Yamaha SMAF and NEC phone chips are later versions of iconic DX and CZ synths.
In previous post was that analog rack effects should go together with analog synths, electric organs etc. in one chip. But analog rack / 500- series effects are for studio sound recording with mixing console, so perhaps own chip is better for them. If analog studio effects are in analog chip, then complete mixing console can also be in that chip. Complete analog mixing console in one integrated circuit. Graphical user interface is used so physical limitations in channels etc. are no more. So professional quality integrated circuit chip mixing console can have 128, 160, 256, or 512 channels, in one chip. It is used using touchscreen or some sort of “mixing desk controller” (like keyboard controller uses softsynth GUI), hardware that is like real mixing desk. And price would be cheap. This analog IC can have large number of rack effects / 500- series incuded, price of that gear in real size in real recording studio would cost perhaps million dollar or more. Original Abbey Road compressor from 1960s was sold in auction for about 50 000 or 100 000 dollars. And that was just one compressor in one channel strip from Abbey Road console. Also analog synth, electric organ etc. gear can be included in this one integrated chip. This professional sound production chip (not amateur / prosumer chip like that MicroSD card analog chip) is then very large. It is possible even using wafer size 200mm analog chip. Modern semiconductor manufacturing has high yield and high quality, and low price. If for example 10- or 8- or 3- or 1- micron manufacturing etc. is used it is not so costly, so wafer size analog “studio-in-a-chip” is possible and for quite cheap price (few hundred dollars? 1000 - 2000 dollars? Or more?). Some parts of it are not working or work only partially because yield is not 100%, but user has to accept that because user gets complete recording studio plus all famous music gear (synths, string machines, electric organs, analog drum machines etc.) in one 200mm size silicon plate. What is electric consumption and what cooling this 200mm wafer size analog chip needs, I don t know. This 200mm chip can be put to one slim rackmount module or to one 500- series module, it needs only power button, everything else is in graphical user interface. So it is “analog audio workstation”, or actually complete analog recording studio with all possible famous effect racks that has ever been, plus all outboard analog gear that can be imagined, in just one chip. Also digital mixing console can be made in one digital IC, but that is propably done already.
When analog synths are chosen for integrated circuit it is not worth to use 1970s / 1980s synths that used large amount of Curtis or SSM chips, because those are back in production, and if someone wants to make clone of those synths those can be accurately copied using original hardware components, in hardware full size synth. Some analog mono or polysynths used IC chip DCOs, VCFs, envelope generators and VCAs that all were Curtis or SSM chips or other ICs. Then were some synths that used only few ICs, and some used none (Minimoog model A and some Teisco keyboards had no integrated circuits at all, others had at least one, transistor array chips etc.). Those synths that used few ICs can be put to one modern IC chip in one large group with complete count of over 100 voices and several dozens of different synths.
Not only synths but analog drum machines, string machines, ensemble keyboards (analog audio workstations mentioned in sequencer de netpage “Combi synths, strings, bass…” page), electric organs and combo organs can be all put together with analog synths in one integrated circuit chip, perhaps made with 0,5 micron or 1 micron technology. This circuit is small and fits with digital control logic (made with 40 nanometer tech for example) and memory chip in Micro SD card or USB stick. Modern fast SD card has 0,72 W minimum and 2,88 W maximum power consumption, and about 450 mA avarage in heavy use (about 1,4 W), so 1,5 W power consumption is not too much?
There can be one chip for guitar pedal that contain all famous analog guitar pedals in one chip if that pedal is older than 15 years, and guitar synths included. Graphical user interface is again needed so pedal must be connected to laptop or something that guitarist has on stage, and control of that analog chip pedal is through laptop mostly. Bass players can have own similar pedal also. Wind controller can have similar chip, inside cheap plastic wind controller that has all famous old analog (and / or digital) wind controller hardware from 1980s.
VCOs when made to IC need temperature compensation, so large amount of tempcos can be in IC size VCOs, or heating element, or both. Small miniscule heating element in small chip does not need much power, if miniature heating elements are placed only in those places that have VCOs in small IC area. Or making VCOs to DCOs in IC chip will solve oscillator drift problem.
Also polyphonic synths can be made to individual voices, for example Memorymoog is not worthwile to make in IC because it can be build using Curtis chips back in production. Instead six Minimoog model As can be in IC, six individual synth voices. Other polysynths can be made individual voices also. Instead of Rhodes Chroma or Polaris can be six Arp Odysseys etc. Arp 2600 is more complicated than Odyssey so only one Arp 2600 can be in IC chip etc. This collection of monosynths and perhaps in some cases polysynths can have well over 100 voices many dozens of different synths from 1970s to 1990s or 2000s, all analog. Analog- digital hybrid synths were few in 1980s, there are about 20 or less of them. Modern hybrid synths are under copyright protection. But those old hybrid synths can have their own analog - digital hybrid IC if needed, or digital control logic of analog synth chip can handle them also.
Each individual synth voice can have instead of just one LFO and one ADSR, many of those, 2-6 LFOs and envelopes. Those are digital for simplicity of chip. LFOs can be virtual analog type in one mode, and purely digital in another mode. Instead of 2-6 channels for LFOs and 2-6 channels for ADSR complex LFOs and complex envelopes (function generators) can be used, complex LFO is like many LFOs in one channel, Doepfer A-143 for example, and function generator is ADSR modulated by (digital) LFO or audio range oscillator. So when LFO and ADSR are in complex form only one channel for LFO and one channel for envelopes will do, and chip is simpler. Modulation routes can be increased, if in original 1970s synth only VCO is modulated by LFO, now it can go to VCF and VCA also. If AM modulation is not in original synth, it is easy to do also.
Each individual voice can have additional features, simple analog ring oscillator, simple sample & hold, simple aural expander (there is synth made only using aural exciters, EXCTR by Samy Karam, this thing can be included in synth collection IC and used as effect also), simple dynamic compander (simple compression circuits exist), simple wavefolder, simple coupled treble / bass control and panning when 2 channels are coupled as stereo. Perhaps even simple distortion can be used, although what good use keyboard sound has for distortion effect? All those circuits have combined component count of transistors, capacitors, resistors etc. of only about dozen per effect in simple form. Also simple digital sequencer and white noise generator is for each voice, and LFOs and envelopes are digital also. Additional simple VCO or DCO can each synth voice have, and simple extra VCF also. Low component count audio VCOs and VCFs exist. But chorus, perhaps most important effect for synth sound cannot be used because it uses delay circuit that has huge component count. So IC has global (all voices use the same) chorus / echo and flanger / phaser circuits, or two of them (for stereo sound), and parametric or graphic equalizer of over 20 channels (two of them), and two subharmonic generators (synthesizers). So now almost unlimited sound sculpting is possible using this analog IC.
To control this mayhem, every synth voice can be controlled and played separately, all voices together with electric and combo organs and string machines in huge unison if needed, if player press one key of keyboard controller all over 100 synth voices and electric organs etc. play this note, perhaps some sort of artificial intelligence is needed to control all this cacophony. XMOS makes cheap AI chip logic that is for sale, and XMOS chips were made for sound control, so including AI to digital logic of this analog synth chip will make human control more easily when all those voices are used in same time. So Micro SD card can include analog synth / drum machines/ electric organs etc. compilation chip, and also in SD card has digital control chip and memory chip. This thing can be in USB stick also which makes power and space requirements easier.
There is also analog programmable logic, field programmable analog array, Anadigm FPAA. If its power consumption is unsuitable for Micro SD card it can be fitted to USB stick together with analog synth etc. compilation chip and digital control logic chip and memory chip. FPAA has unlimited changing potential (one Anadigm FPAA can have 128 filters) and it can change its parameters on the fly, but noise is problem, which is not a problem if FPAA is in cheap low cost prosumer device.
Not only synth based on aural axciter, but drone machines / synths based on astable multivibrators, avalanche oscillators, sweep oscillators etc. can be included in this synth collection IC chip.
Some real complex analog synths like Yamaha GX-1 (including CS80) and Korg PS 3300 can be put to their own analog IC and own SD card / USB stick, because they are complex designs. Complex mono synths like Moog Emerson (it is actually 3-voice?) and Buchla, EMU, perhaps Serge can also be placed in one analog IC, patc cables can be used through vieas of IC chip, amplified signal goes to patch cable and back, so modular synth can be in one IC also. Sound can be expanded using EML polybox / Ethenvar patch chord type solution in IC.
Digital synths can also be grouped to one IC, old digital synths from 1980s before era of romplers / wavetables from 1989 / 1990 onwards. Old classic digital synths, dozens of them, can be in one IC made for example using 40 nanometer process. This chip with memory chip can be in Micro SD card or USB stick, used with graphical user interface.

Nowadays phones etc. have software for digital audio workstation (DAW). But Android has latency which makes music software for Android difficult. What if instead of software DAW, it is in hardware? So instead of software, phone has in Micro SD card complete digital recording studio? This is real hardware digital recording studio, with real mixing console, but it is just inside SD card, as integrated circuit. And price of this “recording studio in pocket” is few dollars. Again, this is not software emulation of recording studio like software DAW, it is real hardware mixing console, that is used with graphical user interface. All limitations of Android operating system can now be bypassed and problems with software etc. Prosumer mixing console of 16, 24 or 32 tracks. Of course all kinds of effects, compressors, echo etc. is included with console, and the synth collection of old digital synths mentioned before. This hardware DAW can be inside USB stick also. Memory chip is together with digital logic chip that has DAW, memory chip has GUI. Both chips are inside one Micro SD card or USB stick. Even real pro mixing console in one IC can be made, it is then free of physical limitations of channels of real mixing console. IC mixing console can have 512 or 1024 channel strips easily. It needs graphical interface and real mixing console that has max. 96 channel strips is used only as remote controller for that IC chip. Price of for example 1024 channel strip digital mixing console in one IC is few dollars, made with 40 nanometer process. Each of 1024 channel strips have usual compressors, limiters, equalizers etc. 1024 channel mixing console needs also 1024 track digital recorder. So price is cheap, but possibilities are endless what can be done to sound with this 1024 channel console. Console automation must be used if 1024 channel console is beyond human control. Perhaps artificial intelligence can be used with console automation to make 1024 channel mixing desk controllable, it is used mostly through GUI anyway.
About digital synths in one IC: original Yamaha DX7 was made with 4 micron process, 40 nanometer process has area of 10 000 times smaller, so lots of old 1980s digital synths must be grouped to one IC so that chip is not too small, those 1980s digital synths had low transistor count. So more synth found whose copyright has expired the better. That means all synths that have no copyrighted wavetables or rompler waveforms. That means old 1980s synths that used FM or additive synthesis, some rare models had digital subtractive. Physical modelling synths appeared in 1994 / 1995. Formant synthesis appeared in 1998 by Yamaha, but it was coupled with rompler synth engine so not possible to copy. There was some big music synthesis systems used by academia in 1980s, but if they used DSPs and not specific ASICs, it is better to put old DSP synths to another IC so when possible the synths that use same DSP type can share same DSPs. Synth features can also be expanded from what they were in 1980s. Even old Yamaha DX7 had in one operator possibility to use user defined waveform from outside of the synth. Instead of using old digital sine wave etc. waveforms can be modernised and changed, even changed on the fly during sound generation process, use bezier curve synthesis with additive synthesis etc. Vector phaseshaping synthesis can be used with phase distortion synthesis (Casio, Con Brio). Each synth voice can have own sequencer, and then when all synth voices of all synths are played at same time some master sequencer syncronises all those hundred or more one voice sequencers and makes sophisticated sequenced patters from 100 or more synth voices.
About analog synths in one IC: to control oscillator drift, either all VCOs are changed to DCOs, or to VCOs are added lots of tempcos, matching transistor pairs are used in VCOs, either heating plate or similar system that transistor array 3046 has, it has one transistor that is used only for heating of the circuit. Such heating transistors etc. can be used in integrated circuit for heating that keeps oscillator drift away. Tuning system that Minimoog had, crystal based, can be used. Automatic tuning system sends tuning note through synth and tunes it in few seconds or faster. Additional features can be put to analog synths in one IC also. Not just sophisticated digital LFOs and envelopes, synth OSC can have extra transistor that is used for extra (primitive) OSC. Digital pulses send to this transistor makes it pulsating and so primitive (extra) oscillator inside oscillator. Also sub oscillators can be made to OSCs, they are simple. Extra (simple) audio oscillator and extra (simple) audio filter can be added to each synth voice of old synth structure of 1970s or 1980s that is put to IC. Added synth features can be used, for example feedback amplitude modulation synthesis (FBAM) that uses allpass filter chain. If it is possible to chain filters in synth structure allpass filters are chained and FBAM can be used (perhaps, FBAM is digital technique).
Analog synth, when it is not miniaturised form inside IC but real hardware made with discrete or Curtis / Sound Semiconductor chips, can have added feature: cassette deck. Oldest synth, Bode synth used sound tapes instead of oscillators. Analog synths are limited to analog waveforms that oscillators can make. Digital synths can use whatever waveform they want. This flexibility can be brought to analog synths of sound source is not the oscillator but sound tape, like Bode used. Cassette can include waveforms (sound) that replaces oscillator in analog synth signal path, cassette sounds are source that are used to modulate filters and VCA with ADSR etc. Analog synths are then more flexible and near flexibility of digital synths. Cassette players, portable or car players, are not so expensive so if added to regular analog synth not cost too much. Using double cassette players makes possible flanging / phasing / chorus / echo effects. And cassettes can be used as “cassette Mellotron” etc. sound effect duties, and as analog samplers. Noise shaping IC circuits can be made, for example mass production of old High Com noise shaping (that was the best?), even small amount can be made using shared wafer / MPW manufacturing. Cassette can also have double speed and even quadruple speed for increased sound fidelity. Cassette decks are expensive, but portable or chinese cassette car players are cheap.
And about analog tapes: Mellotron. Because Mellotron uses analog tapes, it would be natural that Mellotron should also have tape based effect system, tape flanger / phaser and tape chorus / echo. For some reason, although tape based delay / effect systems and Mellotron use almost same mechanism, Mellotron does not have them. Because tape echo and effects are almost same as Mellotron, same manufacturing, components etc. could be used, and Mellotron can then have tape effects, flanger, phaser, chorus and echo, and echo can be extented to Frippertonics super long echo. Also sound sampling (input audio) can be added to Mellotron, making it analog sampler. There was Birotron, Projectron by Alan Parsons, and analog-sampling Memotron (only one prototype). Open reel 8 track tape recorder can be modified to be Projectron, it would then be analog sampler and tape recorder, all together. Unfortunately 8 track tape recorders are not made anymore, 2 track open reelers are made at limited quantity. But analog sampling or “Projectron” may cause some interest that it is economically possible to build 8 track tape recorders again with added sampling feature, 8 track recorders have market up to this today for persons who want to record in analog tape professionally. Perhaps even 16 track fully professional Projectron / tape recorder is possible. So 8 or 16 track sampler Projectron / tape recorder hybrid perhaps will sell. Mellotron can have not only tape effects, but also additional cabinet space over keyboard for eurorack or other modular synth modules, like Bode synth Mellotron is then used as oscillator for eurorack filters, envelope modules etc. So Mellotron plus synth modules in same cabinet then become analog synthesizer, where Mellotron tapes replace oscillators. This kind of new instrument can be called “Supertron” or by some other name. Mellotron with synth cabinet can have empty cabinet for synth modules, customer can himself buy them and fill the cabinet and buy what modules he wants or just leave module cabinet space empty, in “Supertron” is just empty synth module cabinet space over Mellotron keyboard plus additional tape effects / echo added to Mellotron. And this combination can be called “Supertron” or other. Projectron, which is modified tape recorder, can have remote control unit in eurorack format, that control Projectron, patch cables go from tape recorder / sampler Projectron to eurorack modules, Projectron replaces oscillators. Analog tapes can have sampled sounds of digital synths, in analog tape samplepacks, or Omnisphere type “samplesynths” in tape. Now analog synths can have sounds like digital synths, when usually it is so that digital keybords use samples from acoustic or analog instruments, digital samples from analog synths. But now analog instruments can turn this analog / digital sample thing upside down and otherway around, analog samples from digital synths.
In previous posts are that analog delay effects such as chorus have such huge component count that IC cannot have own individual chorus for each analog synth voice. However there is in one netpage “Accurate analog delay circuit” (1997), where is written that circuit does not use much space. Such circuit, if it can be used as chorus effect, may make possible simple chorus circuit for each synth voice in compilation of analog synths in one IC, if chorus / delay circuit is simple and small. Also if analog recording studio is made as one wafer size chip, old Curtis / SSM chips that are back in production used 3 inch wafers, not 8 inch / 200 mm. So made in 8 micron manufacturing tech and 3 inch wafer, wafer size IC is propably cheap to make nowadays. Some percentage of it is malfunctioning, but IC manufacturing has very good yield nowadays, so if some channel strip is not working etc. it does not matter much. Also if analog recording studio in a chip uses graphical user interface, it is then “analog audio workstation”. Only nonlinear sound editing is missing, but can be in analog recording tech too. Bank (many) of 2-track tape recorders, with SMPTE or other sync code, can be used as nonlinear sound editing, computer controlled. There can be 12 - 32 different 2-track tape recorders in studio wall, making downmixing, stems etc. normal things that is done in recording studios, but editing is not done with tape splitting and gluing, instead from 2- track tape sound is copied to another 2-track tape of another recorder, using computer controlled process, without touching tape by human hands. Multitrack recordings use many separate 2- track recorders that are synced by computer. VHS tape hifi video recorders can be used for sound recording also, and many of them for multitrack recording / editing. Separate tracks can use non linear sound editing, and non destructive editing (tape is not split, it is just copied). Only when tapes or VHS cassettes are loaded human hands are needed in process.
There is also optical sample player, Optigan. Modern versions of the type are perhaps in netpage “yehar com / blog”, there is “Laserotator” and “Black deck”.
If old analog synths are collected in one IC chip, then not only western or japanese synths, but also legendary russian synths too (Polivoks, Aelita, Alisa, TOM 1501, Formanta EMS-001) must be included.
And about phone sound chips mentioned in earlier post: those have, for example in Yamaha SMAF, modern Yamaha wavetable / rompler synth sounds, and NEC chip perhaps modern Casio sounds, so if they are in any way related to modern Yamaha and Casio keyboards, putting those sound chips to Micro SD card or USB stick is cheap way to have modern synth sounds easily, in hardware.

Analog mixing console needs analog recorder if sound should be in analog form. Tape (and videotape) for sound recording and editing is not so easy as digital memory. Easy analog editing system is needed. Earlier TV stations had analog hard disc recorders for recording short snippets of TV picture. Analog hard disc would have big platters, each platter is one channel so 24, 32 or 48 platters are needed. And even C-cassette width / speed sound track requires big platters and recording time is short. Earlier 1960s computers had “Transfluxor” memory that was analog. One source says 250 000 bits in cubic inch which is not much, another that 1000 000 000 bits was possible, but what size that memory needed is not written. And does transfluxor has capability to record 24 track sound, in good hifi quality, in long time? Analog recording to CCD / audio echo delay like IC delay chips use, has been used in 1970 /1980s. Sound corrupts the more time passes so after few weeks audio is unusable. But if sound corruption is progressive, meaning first sound corrupts slowly and then at esacalating speed so that corruption speed is fastest near the end, then if sound is constantly transformed between two memories, every second sound moves from one memory to other and back in ping pong style, then corruption is much slower, because the sound is always stored only 1 second, not several weeks in one place. If sound corruption is linear, the same speed in beginning and end, then this system does not work. “How can one make an analog voltage circuit?” electronics stackexchange com. But there has been improved versions of analog memory presented in scientific papers and patents during past 30 years. For example patent 550499 “Nonvolatile analog memory” 1996. Analog memory is in series production in many semiconductor forms nowadays, but that means analog storage of quantized signals (bits) in many memory cells (quantized information). “Switched-current analog memories”. None of analog memory manufacturers is interested in making audio range analog memory. Although IBM has prototype 8 bit memory that is real analog memory, information is in analog form. Such memory should sell as audio store, it can replace vinyl LPs and cassettes as analog sound source. So it would be like analog memory card or USB stick, with analog not digital output. Also those analog synth collections in IC chip can be memory card or USB stick that has analog output, not digital, and even analog input for voltage control of outside synth hardware, modular eurorack synth can be routed after voltage and current conversion to this IC chip. Recording studios could use analog semiconductor memory as analog multitrack recorder, making editing much easier. Short -term analog memory, only seconds or minutes long, is simpler and it can be used as analog sampler in analog synths and in analog recording studio.
Optical memory is another chance, optical film sountracks have been used for ages. Analog version of Bluray disc, with analog electronics and recording like optical soundtrack with dynamic companding, multilayer Bluray disc and large size disc like LP 30 cm or even larger. Multitrack recorder could have 24 normal size 12 cm Bluray disks, like platters in hard disc, for 24 channels. Or 30 cm LP size discs. Optical recording and magnetic tape are only linear, non quantized recording methods if there is no non-quantizing semiconductor memory in existence. Analog sound processing would make possible things like KLT transform for multichannel audio. Anamorphic Stretch Transform can also be used, but I don t know what it does for sound. “What would be impact of an existence of analog electronic memory systems” 2013. There is also optical tape memory, for computers. It has been just research compared to magnetic tape digital memory which is in wide use. If instead of computers optical tape is used for analog sound recording, if optical tape makes better quality than magnetic tape.
For analog synths: every real hardware analog synth can have Anadigm FPAA, for extra features, it is like analog FPGA so synth can change structure on the fly. Even every modular synth module can have Anadigm FPAA. Although there may be sound problems with noise, FPAA is so cheap it does not matter because FPAA brings so much additional features to sound, and even cheap toy keyboards or prosumer keyboards can have it. For electric organs and string machines that are together with analog synths in one IC chip: those can have also same effect section like synths, with ring modulator, sample & hold, aural expander, dynamic compander etc., but instead for each voice, polyphonic or paraphonic electric organ has only four of those effect sections, first organ or string machine voices are divided to two, left and right (hand, when played by keyboard), and then both left and right section has two effect sections for stereo (right / left channel) sound. So only 4 effect sections is enough for those that are not synths. Digital sequencer can however be in each voice. This synth etc. chip can have one master sequncer which uses simpler individual voice sequencers to make one big complex multi - voice sequence.
DCO use usually 16 bit integer, but floating point, posit or quire can be used also. Also accuracy improvement can be used (even with FP, posit or quire), DCO has 20 bit computing by digital logic, but first 4 bits are either rounded or truncated (least significant bits are dropped), then dithering is used to drop 4 bits, finally 12 bit signal is sent to DCO. This 12 bit has almost 20 bit accuracy because it is dithered and rounded.
“Digitally controlled frequency generation using VLRO”, “Tuneable pipelined frequency transform”.
There was polyphonic modular synth, Polyfusion, and Tiptop Audio intented to release 8 voice synth modules, but perhaps price was too much because some of them had 1000 different discrete components. But for IC chip 1000 components is small amount. So either complete modular polyphonic synth or just separate modules can be put to IC chip / chips. It would be cheap also. Shared wafer / MPW manufacturing makes possible small series production, only couple of hundred units, which is suitable for modular synths.