All About FKG3 Keyer

FKG3 Keyer serves a wide range of creative image compositing functions.


  • Width, 12HP
  • Mounting Depth, 42mm
  • Power +12V @ 250mA


Other Resources


FAQ: What are the differences between Doorway and FKG3?

  • Doorway’s foreground and background are monochromatic only, but have dedicated grayscale level controls.
  • FKG3’s foreground and background are full color / triple channel, and have preset levels (the foreground is white and the background is black – until you patch something else in.) I find this much more intuitive.
  • Other than this detail, FKG3 covers Doorway’s functionality entirely.
  • FKG3 does new things Doorway cannot do, and has some functionality that makes it easier to use. With all combinations of the rotary and toggle switches, we can say that FKG3 is a finite state machine with 36 possible states, or modes. Doorway only has 2 states (outline on or off.) This means you can get a lot more out of it without changing the patch.
  • Doorway was designed to generate a key or to combine keys with each other. You are then intended to patch the resulting key to some sort of triple fader or compositor, to finish the backend half of “switching between two full color images.”
  • FKG3 encapsulates this in a single module, making what was previously a 26HP+ patch (Marble Index + Doorway, or similar) into a 12HP chunk. Compared to Expedition series, the Gen3 modules pack the “workflow step per HP” ratio of the system into a much smaller space. The resulting effect is that by the time you have 3U of modular, you’ve got an advanced system instead of a basic system.

Just to make sure I understand the different modes/possibilities here, whichever key mode one chooses that ends up a global key for all three channels, right? There isn’t a mode where one can treat the module as three separate mono mixer/keyers with three different mono key sources, correct?


Just to make sure I understand the different modes/possibilities here, whichever key mode one chooses that ends up a global key for all three channels, right?

There isn’t a mode where one can treat the module as three separate mono mixer/keyers with three different mono key sources, correct?

Right. Think of this keyer as having multiple channels, rather than being multiple keyers. Like adding green and blue channels to Doorway.

There may be a “triple mono key generator” type module later on down the road, but that’s likely to work more as an “RGB component amp” with high gain/clipping modes, or a “quad mono keyer” is a possibility as well (that is likely to take a different form.)

“Key” is not a noun describing a signal in Gen3. It is a verb describing the action of one image switching between two or more other images, and how that image is processed by Threshold & Softness: the “keying controls.” This allows you to keep the images on the patch cables full scale and even full color, up until the point it matters, so you can patch the same source multiple places – but with different “keying” processes going on at each destination.

There will be redundancy on functions throughout Gen3 – each module is more about a workflow step than a specific electronic function. The workflow step encapsulated by FKG3 is “keying between two images” – we feel like this is the most essential core modular component of a video instrument’s processing workflows, which is why it’s among this initial set. Want your system to be capable of keying between two images twice? 2x FKG3s, etc.


Okay, cool. Yeah, this is all as expected (can’t hug every cat!). But just wanted to make sure I wasn’t missing something. And I was probably a little thrown off at first because of its naming relationship with the old Triple Video Fader & Key Generator.

But makes total sense since this has a whole logic thing going on (R not BG / B not RG / G not RB) in the signal path to get the chroma keying (as I understand it). There is probably a lot of interesting experimentation to be done leveraging/misusing that, starting with three mono signals rather than an RGB image going into whichever layer is the key source (which is to say: the source for the keying).

Looking forward to seeing what other flavors of mixers you come up with too! The more, the merrier.

FWIW, the way you are using “Key” now brings it in line with how I have always thought of it, yes, as a verb, and makes a lot of sense to me — even though it is different than how you were using it before (which honestly tended to confuse me somewhat).


Yes, this is right! If you look at any broadcast video keyer or any keying function on a video mixer, FKG3 is way more in line with that expectation. It’s a function that necessitates three images: a background, foreground, and a keying source (which may also be the foreground or background). That’s why it is a “Keyer” function and not a “Triple Crossfader With High Gain Common Drive” function for example.

The reason Visionary series TVFKG was the way it was, is because we were trying to maximize what we could get out of a very expensive off the shelf IC, the LT1256. Since moving to discrete circuitry, we have designed a discrete three channel keyer that uses a common current drive for all three channels. It’s just not the same circuit as a “triple mono component keyer” – something that it would not be possible to fit into 12HP without sacrificing the usability or functions of FKG3.

So a “triple keyer” needs to be something else – and to really identify what that module would be, we have to consider it in context. If the context is “I want to be able to key between two monochromatic images three times” – well, you can already do that with 3x FKG3s.

If the context is “I want to be able to key between two monochromatic images three times, but in 12HP”, then we ask “why? where are the sources coming from?”

If the answer is “I want to get different sized hard blocks out of my ramps and shapes” then maybe the right answer is a “triple pulse width modulator / boundary compositor” or just a “colorizer.”


Didn’t necessarily have a specific use case in mind. More thinking in the opposite direction of trying to grok what all it can do and then think about what I could do with that.

I generally work more with image processing than shape/pattern generation (though the line can be blurry), so probably something in the realm of image manipulation/re-colorization, yeah. I can think of a few different ways a compact triple keyer could be useful for that. Would it be more useful than what you have here? Probably not.


this makes a lot of sense but is not how i initially interpreted it. i thought it was more akin to Fox’s Shutter with three independent switching/keying signals all sharing a single Threshold/Gain value. in short i had imagined the RGB channels of the :key: input would individually key the R, G and B channels when set to Luma mode but all channels would share the same key when set to R, G or B mode. not sure where i got that idea but that’s where my mind was at from the time i first saw FKG3 - three stacked Doorways sharing a Threshold/Gain with independent OR shared key source.

in any case - super stoked to play with this but also very much looking forward to the possibility of a “Keyer and/or Fader Bank” type module as described above. guessing in that case it would make more sense to have individual threshold/gain controls per keyer/fader, which of course would make things ultra dense and possibly exceed the scope of a 12-hp module as outlined above… but i can’t help but imagine how cool it would be to have 2x RGB sources fed into something that can key them independently or all together with R/G/B input. til then i’ll hold tight to my pair of Shutters (and my FKG3 that arrives today!) :grin:

why? i fully admit i don’t really have a good reason other than that’s what was in my head at first, but i CAN say that i’ve had a blast using Shutter to effectively key two RGB signals by a third RGB signal… though i also admit the results have rarely been anything less than chaotic :laughing:

that’s a lot of very indirect words to say this but in short: i really appreciate these threads! so much to learn about these tools and this is a great format for it. good to be able to see people asking the important questions that never even occurred to me and getting the important answers in a clear, direct, concise format.

1 Like

Right. Think of it like this. Do you want a 24HP Keyer with triple keying modes that is $799? Or do you want the option to pick 12HP Keyer or 12HP “Triple Switcher” for $399 each? There’s nothing sacrificed in the latter approach from the user’s perspective, but you gain the ability to (assuming you’re filling out 24HP with Keying) have 2x Keyer or 2x Triple Switcher or 1x of each! So that’s three different system building choices and two different workflow steps for $799. Rather than one choice and one workflow step for $799.

In the “Keyer” case the inputs are for “images”. This lets us do the luma and chroma key things. A “Triple Keyer” case the inputs are for “channels” – this lets us do other things but makes it less intuitive to do the luma and chroma key things. So the right answer is both – it’s just a matter of how we approach the second answer in the future. The reason FKG3 comes first is because that it’s possible to cover the “Triple Keyer” use case using multiple FKG3s – but the reverse would not be true (that is, using a 12HP Triple Keyer to do luma and chroma soft keying.)

why? i fully admit i don’t really have a good reason other than that’s what was in my head at first,

Keep in mind that it’s only possible to answer the “what” and “why” question here because this is a new iteration of modules. So we can answer the question in relation to functionality found in (or missing from) previous modules. The goal isn’t to limit scope of what you can patch in, it’s to identify the modular “step” in the workflow to encapsulate, and ensure we are migrating all the important functions into Gen3. Any answers are okay.

but i CAN say that i’ve had a blast using Shutter to effectively key two RGB signals by a third RGB signal… though i also admit the results have rarely been anything less than chaotic :laughing:

Right. And that’s why! They are two separate workflow needs. They’re used for different reasons, in different places in your patch. You want to go full chaotic color component switcher, wait for a future Triple Switcher or however it takes form. You want to luma/chroma key between sources that are presumably two RGB sources from external photographic sources, FKG3 is probably a more appropriate answer for that workflow. Or double them up. Or get both. The goal is to give the user more choices in the system build. Gen3 is designed up front to make “balanced” systems possible, even in small or mini rigs (3-5 modules.) That way if all you want is a small modular video synth, you’re not stuck feeling like you need other modules to make it work. FKG3 is traditionally a difficult thing to patch (multiple modules required), or something like it has been permanently tied to an output encoder (like Visual Cortex).

So if mono channel functions are more your thing and an RGB keyer seems bloated because you’d never use green or blue channels (because you don’t have any RGB sources or don’t like the idea of composite images being inputs or don’t like composing RGB images out of mono sources), hang tight – there will be other modules coming. And we can keep talking about ideas for them over the next few months, too.

And def hang on to your Shutters – it’s a cool module! If that’s working for you, you’re all set. When we do some kind of multi-channel keyer, that doesn’t mean it overrides Shutter. It would be a different take on it. None of us owns basic electronics circuits like “keyers” and “faders” – there’s plenty of room to have different takes.


I’m excited for having example videos eventually. I can’t quite grasp my head around what it does if it’s not just like 3 separate mono keyer channels. I really like it, though, even if I don’t quite know what it does yet. :slight_smile:


This is all I have time for, for now. But I hope this simplified block diagram gives a bit more clarity. (Sorry, my handwriting looks awful without graph paper.) Johnny has a unit on the way, and I bet he’ll have a demo video up soon.

(Note: the “Inversion” / “Outline” switch is part of the Soft Key Gen block in this diagram.)


My Keyer arrived yesterday, but i just made it to the post box this evening. super excited to read the recent posts, updates, andcatch up in general. Can’t wait to populate this into the proteus and see what happens ;}


This looks great! Thanks for the head’s-up in the MW forum! Just put in my pre-order!


This looks amazing!

I’m a noob to video synthesis, but have a ton of graphics experience, eurorack, and some basic electronics as well. I’ve watched the demos several times and think I have a good grasp of the concepts.

I’m wondering if this could be a good way to get started. I would be playing recorded videos for background, foreground, and key and recording the results. Would the FKG3 Keyer work standalone? Or would I need other modules to get started? I’m assuming that I would need to convert the playback from HDMI into component on the input side and the output results from component to whatever format I will eventually record in.

Thanks to all in advance!

1 Like

This Gen3 Primer should help.

You’ll definitely need some other modules to get started. The absolute minimum required would be an Encoder module, just to get a video signal. Everything else in an LZX system is video rate or lower, but definitely not a video signal general video gear will recognise.

Thanks a bunch for the help! Its starting to make sense a little at a time. If I understand correctly, the ESG3 encoder will accept the output from the FKG3 keyer as input and the result is YPbPr. Will I need the opposite (?decoder) to convert to input into the FGK3? I was thinking that I could convert HDMI to YPbPr to input for background, foreground, and key. Sorry for the noob questions!


Two issues with external video inputs–you need them to be synchronous and they need to be decoded into the 1V LZX standard. Currently there are no available new modules to do this but TBC2 will be able to handle this problem nicely. External hardware can solve this in some cases (link) but can be clunkier. You’d still be missing the necessary encoder until at least ESG-3 comes out.

Check these threads out, too:


Thanks again for the help! I will noodle on this a bit and come up with a wish list. It seems like I will need to get a critical mass before I can get cooking! Similar to eurorack, need a build a basic set before anything good comes out of it.


Maybe consider starting with a Chromagnon and build out a system from there.


That’s sounds like a good plan. Chromagnon, Keyer, and TCB2 for additional inputs. I’ve got a lot to learn, but it is slowly sinking in!