I have the Chipsynth MD, which is a Sega Mega Drive (aka Genesis) based synth from Plogue and it's one of my favorite plug-in synths. It's pretty remarkable how much it nails the sounds from those games.
This video shows some insight into the development process: https://www.youtube.com/watch?app=desktop&v=VLxTHYGLKY0
I bought it yesterday and I'm loving it. It's really good! AND NO SAMPLES! I want to support developers like this who make something out of bounds with math instead of same old same old (samples, kontakt player, half done synth engine) I appreciate the level of detail and effort he put into this emulation, and it ALMOST cures my lust for a THERAPSID mk3.
No samples? Damn.
That was done by using 4bit sample data and using those values to set the audio level.
>I want to support developers like this who make something out of bounds with math instead of same old same old (samples, kontakt player, half done synth engine)
There are countless of synth engines, from basic lfos and filters, to modelling circuits.
This doesn't have any special sauce.
What do you need to run this?
What: I'm using Cakewalk Sonar, but any DAW that supports VST3
How CPU intensive is it?
(VST beta tester here) Seems fine so far over 24 hours. Typically VST's cause spiking and audio glitching right away, so this build seems good, albeit 1.0.0
I think the parent wanted to know how big of a SID orchestra they can expect to build.
This may help (as a place where to build a reference repository)
[see me answering my own question, at bottom of comment]
Is there scientific data supporting the verbal "it's really good" and "we did lots of research" claims?
Such as mdfourier testing and comparison with the original chips, or equivalent scientific scrutiny and comparison of the respective outputs?
The claims may well be true, but the greater the claim to merit, the more robust the supporting evidence should be.
[Answer follows: ]
Yes there does appear to be some, at approx 10m45sec of this Youtube video which is contained within the OP link. https://www.youtube.com/watch?v=pePq68HaI7M
Skimming the article, it does seem like they've built a model (there has been much work over the years on opensource emulators,etc so the basic operation is fairly well researched) and then sampled a bunch of real chips to find variances and behaviours to get profiles that comes close to the real ones.
As for being scientifically correct, Commodore was infamous for cutting corners so resistances (and prob other parts of fabrication) varied wildly so at best you can probably compare to specific chips (even if families of fabrication are closer,esp w.r.t to 8580 vs 6581 behaviours).
Read the article again in that light (as those details are alluded to) or look at the analysis of the C-64 graphics chip colors (There is a nugget in the response from the Sid-chip designer at the bottom of this other article about Commodore inner workings).
> I can tell you that the design was based on the principle that adding a sine wave of a particular frequency and amplitude to an inverted version of the same sine wave at a different amplitude produces a phase-shifted sine wave of the same frequency. The amount of phase shift is directly proportional to the amplitudes of the two sine waves.
I can't follow that. Wouldn't it just lead to a lower amplitude sine wave? Where does the phase shift come from? My visual imagination is failing me here, mostly I don't get why the sine wave would shift forward or backwards if all that changes if the amplitude because that's symmetrical.
Likely bc the "adder" is not an adder. I've analyzed it. It's composed of FETs with nonlinear gate voltage-drain current relationship (that's normal). Often in radio circuits these are set up and operated in a small linear region. They work well enough in a small voltage range, but in SID they are used for larger amplitudes. Result can be modeled as a sum of Taylor series meaning fun, new partials like x^N get synthesized for free!
That makes more sense. You need something that has a temporal element to get a phase shift.
>As for being scientifically correct, Commodore was infamous for cutting corners so resistances (and prob other parts of fabrication) varied wildly so at best you can probably compare to specific chips (even if families of fabrication are closer,esp w.r.t to 8580 vs 6581 behaviours).
It would be really neat if Chipsynth were capable of emulating those different resistances, to tweak them a bit to get a 'warmer' sound or something different, to allow further tweaking of the sound. I'm not sure if Chipsynth is emulating at the silicon gate level, or the analog circuits within the SID. Would be cool if it were.
6581 units are so wildly different re filter behaviour that most game musicians of the time avoided using filters altogether! 8580 is basically a fix, and had way more consistency
>varied wildly so at best you can probably compare to specific chips (even if families of fabrication are closer,esp w.r.t to 8580 vs 6581 behaviours).
Or you could just have some toggle to switch from and offer all of their behaviors.
Yes they characterized them even down to manufacturer/production run using a complex waveform that allowed them to measure and tune the coefficients of their algorithm to match. Although there was also a little sound bite dropped in there that chips made on the same day on the same line did not match up. So it seems they have gotten everything close enough that they are within the manufacturing tolerances. And they did it all as if these differences are even audible to a human being.
I recently came across https://github.com/tildearrow/furnace
Super impressive tracker and emulator for many different 8 bit synth chips.
As a personal project I'm also embarking on a 4x AY-3-8910 chip synth all hooked up to as esp32-s3 for usb midi
More info about my synth build idea https://github.com/tildearrow/furnace/discussions/1605
I already have the chips. 10x AY from AliExpress for £7 and they turned out to be the real deal.
Wow! That is a very impressive feature set. I’ve used Deflemask but this seems to have a much bigger chip roster.
This appears to be proprietary software. If you prefer Free Software, reSID also provides very good SID emulation:
The Plogue stuff is really high quality, though I wonder how it compares to their own chipsounds which already did the SID - I guess it's more all-encompassing of the different SID variants? Anyway, will give it a look later, this is neat! (Good to see they're explicitly calling out PWM, that's one of the key parts that made so much of the sound work so well on the machine)
Wow lots of hype about this out today - it showed up twice in my youtube feed.
It does sound pretty good. Not sure it will see much use outside of demoscene musicians though. I would love to be hearing more c64 music in my life though, glad Chipsynth exists.
Now what am I going to do with the half dozen SID chips I've been hoarding for the last 30 years?
There are bands which use SID chips/sound in their music, and they are not demoscene musicians.
Machinae Supremacy comes to mind [0]. I esp. selected a song which their SID tunes are very audible. Normally they do it way more subtly.
Let me know if you want to get rid of them ;)
(I think the Sidstation is the only synth I regret selling, well maybe together with the Jupiter-6)
TherapSID (mentioned elsewhere in this discussion too) [1], looks to my eye to be a viable alternative to the SidStation. I have a deep love for the SID sound, despite never having a Commodore as a child (I am just a wee bit too young) and the SidStation and now TherapSID tempts me to try to learn play an instrument despite being an adult. They just look like so much fun.
15 years in the making - incredible effort guys.
Will we have to wait another 15 years for the ZX Spectrum version? :)
AY emulation exists in https://github.com/tildearrow/furnace
I already have a SID emulation from Plogue from a couple of years ago.
An overview what is new would have been nice. I think at least the UI looks a bit smoother buy otherwise it is hard to tell if an upgrade is worth it.
my problem with this is that they aren't telling you what they're selling.
they say they're selling deep Sid chip emulation, but for what platform and for what formats? if I knew, I'd either be all over it, or wait until they supported what I was after (in my case, AU synth on macos), but without any information on what I'd be buying, it's a very hard sell.
Because OP is just a news website. This is the product page of the manufacturer: https://www.plogue.com/products/chipsynth-c64.html
There's a system compatibility table on Plogue's main page for it. Standalone, VST3, AU and AAX on macos.
do you have a link? I searched for a while :)
It's in the first sentence of the article.
> Today, Plogue is launching chipsynth C64, an AAX/AU/VST3/CLAP plug-in recreation
It's also discussed in the manual: https://s3.amazonaws.com/chipsynth/C64_manual.pdf
There are unique qualites inherent to the SID itself that makes it impossible without a circuit-level simulation that also models signal leakage and crosstalk because that's why it sounds like it does. Gigs of research, they claim, what's that even mean?
No sir I'll stick with my MSSIAH.
Although its sound is legendary, I can imagine it's not that complicated a chip. Had anyone ever tried to replicate it with discrete components?
The sound of SID comes from its architectural decisions and how the die behaves as an integrated circuit.
IIRC, SID has no real mute on the outputs, so oscillators and output stage is always active. This, in turn, injects its legendary background noise to the generated sound regardless of there's a wave being generated or not.
Moreover, the distortion itself is coming from the internal noise of the die itself, as a secondary effect of being an integrated circuit, plus its fabrication technology.
As a result, the sound of a SID chip is result of its design plus the secondary effects introduced because of its fabrication process.
This is what makes simulating a SID very hard. You need to characterize and formulate these secondary effects and inject them to your pure generation (design) simulation real-time, without much processing overhead.
This is really hard. Maybe not in the math department, but integration, coding, optimization and stabilization department.
P.S.: This is why Machinae Supremacy started making music with real SID chips.
In addition to this (and this is a huge part of what makes it hard to emulate) is that parts of the chip works in the analogue domain, that combined with the added outputs/noise/etc is what makes for the richness to be so hard to emulate.
You're right. I forgot that some parts of SID is actually analog. Thanks for the addendum.
Ah I see that's pretty insane. So the character of the sound doesn't just come from the circuit but is really intrinsic to the chip design. That's some really clever engineering.
It wasn't "clever", it was all about cost-cutting, and from a high fidelity perspective, it was a terrible chip.
But, that gave it character, and people grew to love that sound. Maybe it was a fluke, maybe nostalgia, maybe people at Commodore had a good ear when tweaking the circuit and made something sound good out of the little budget they had because they knew they couldn't go for hi-fi (that would actually be clever).
This is a bit like the TB-303, a terrible bass synthesizer that ended up revolutionizing electronic music. Or overdriven guitar amps that became the staple of the hard-rock/metal genre.
> Maybe it was a fluke, maybe nostalgia, maybe people at Commodore had a good ear when tweaking the circuit and made something sound good out of the little budget they had because they knew they couldn't go for hi-fi (that would actually be clever).
In the words of the engineer that designed the SID chip, Bob Yannes who founded the synthesizer company Ensoniq after he left Commodore in 1982:
> “Actually, I was an electronic music hobbyist before I started working for MOS Technology (one of Commodore's chip divisions at the time) and before I knew anything at all about VLSI chip design. One of the reasons I was hired was my knowledge of music synthesis was deemed valuable for future MOS/Commodore products. When I designed the SID chip, I was attempting to create a single-chip synthesizer voice which hopefully would find it's way into polyphonic/polytimbral synthesizers.”
> “The SID chip was my first attempt at a phase-accumulating oscillator, which is the heart of all wavetable synthesis systems. Due to time constraints, the oscillators in SID were not multiplexed, therefore they took up a lot of chip area, constraining the number of voices I could fit on a chip. All ENSONIQ sound chips use a multiplexed oscillator which allows us to produce at least 32 voices per chip. Aside from that, little else of SID is to be found in our designs, which more closely resemble the Mountain Computer sound card for the Apple II (the basis of the Alpha Syntauri system). The DOC I chip (used in the Mirage and ESQ-1) was modeled on this sound card. Our current designs, which include waveform interpolation, digital filters and digital effects are new designs that aren't really based on anything other than our imaginations.”
> “The issue wasn't budget, it was development time and chip size constraints. The design/prototype/debug/production schedule of the SID chip, VIC II chip and Commodore 64 were incredibly tight (some would say impossibly tight)--we did things faster than Commodore had ever done before and were never able to repeat after! If I had had more time, I would have developed a proper MOS op-amp which would have eliminated the signal leakage which occurred when the volume of the voice was supposed to be zero. This lead to poor signal-to-noise ratio, although it could be dealt with by stopping the oscillator. It would also have greatly improved the filter, particularly in achieving high resonance. I originally planned to have an exponential look-up table to provide a direct translation for the equal-tempered scale, but it took up too much silicon and it was easy enough to do in software anyway.”
https://web.archive.org/web/20070222065716/http://stud1.tuwi...
Makes me wonder if the output could have been used as part of a better random number generator on the 64.
As it happens, you can snoop on the output of voice 3, and in noise mode it could indeed be used for random numbers. It generates a new value about every 26 processor cycles.
It's a partially analog circuit.
Yeah, exactly. So I was wondering of the same analog circuit has been built outside an IC (ie. discrete components) but according to the other response, the IC specifics are integral to the result, which is crazy and really cool.
The original chip was built with 1970s NMOS semiconductor technology.
While you can theoretically build the same analog circuitry using discrete N-channel MOSFETs, it will not capture the tone of the original.
In particular the "operational amplifiers" used in the design were very simplistic things built out of just a handful of transistors and are very non-linear devices.
It is possible to build a circuit that behaves similarly but it won't sound the same.
[dead]
Wow! Some really cool sounds you can make with this setup