Agree it’s an odd “feature” / bug that was not found in the SD ramp shape modules… It’s present on my DSG3 too, but I don’t really mind it. I would hope they can achieve a cleaner shape generation for Chromagnon, but I won’t be sad if it’s the same.
@monads see the YouTube description for module list, here is a photo of where the patch went last night
It’s important to understand that DSG3 is not a ramp generator – it is a shape generator. Ramps are a spatial reference, whereas shapes are a figure derived from a spatial reference. This isn’t the only way to think of it, but it’s why I called DSG3 a shape generator. It is really a complex analog function bank that processes a source vector (that could be a linear ramp or could be something else.) The ramp generators inside DSG3 are very linear and precise sawtooth ramps produced by 10 bit DACs, like Visual Cortex’s ramp generators – but aligned to all 15 video format scalings. So it is the analog function blocks (which process the internal ramps) that creates the deformations in the resulting shape.
In other words, when we release the “HVREF” module, the triangle ramp outputs won’t have the same deformation (this comes from the DSG3’s processing path.)
A guide to system ramp gens
– Video Ramps & Cadet IV are analog ramp gens, they are not precisely aligned to the screen
– Visual Cortex & Diver are 10-bit hi res ramp gens with digital waveshaping
– DSG has 10-bit hi res ramp gens with analog waveshaping
– Chromagnon is 10-bit hi res ramp gens with digital shape selection plus analog waveshaping
– Memory Palace & TBC2 media loaders can produce 8-bit ramps via the media loader
The Y output is a good way to exceed the 8-bit limitations. When you make your source ramps, give them a slight tint – then when you blend the channels in the Y output, you’re not just getting 3x duplicates of the same source, but the color channels mix to form a higher resolution gradient.
I’m pretty sure based on what @creatorlars just said that the Y out is derived directly from the R, G, and B outputs through analog processing, via the formula @wednesdayayay mentioned (multiplying and adding is easy in the analog realm), not via a separate output from the digital brains. So if your question means you want the R, G, and B channels from the same png to go out on the R, G, and B outputs, AND you want the alpha channel to go out on the Y, that’s not possible.
That said, if you can output separate pngs on both halves of the tbc2, you could have a png where one or all of the rgb channels is a copy of the alpha channel from the original png, then you can get the alpha out that way in addition to the rgb channels from the other tbc2 half. So, technically possible if you need it, though not in the way you were probably asking about.
This may be a common misconception. I was certainly wrong about what DSG3 was until I saw the demo videos posted to Facebook.
Honestly, I think the naming of the module is a big part of the confusion. Yes, the DSG3 generates ramps. Yes, those ramps get processed to generate shapes. But as you say, the module is really an analog function bank. Correct me if I’m wrong, but DSG3 is really very close to 2x Arch, except:
DSG3 also generates H and V ramps
Connections among DSG3 function blocks are internally normalled, so a lot less patching
Different options for logic / compositing operations
Different options for rectifier aka folder aka frequency doubler
When my system is fully operational and I have time, I think my first video demo will be about putting DSG3 through its paces. IIRC Johnny Woods’s demo did focus on the internal ramp/shape generation. But there’s so very much more that DSG3 can do with a video signal patched in.
Someone posted on Facebook about the use of dummy plugs; this was a revelation to me. Patching a dummy plug into one of the inputs bypasses the compositing stage. That gives access to the invert, rectify, and exp/log functions on a single ramp or single external signal.
do you all think the ART3 will end up being 499 like DSG/ESG?
this is a module I’m super looking forward to. Repositioning workflows are going to be so neat.
I think you may have said this above, but is Gen3 actually intended to be a stable longer-term set of core modules, in the way the Orion series wasn’t, and the Expedition series was almost? In terms of the expense of these modules, longer term availability without switching up the range becomes an issue of more equitable accessibility.
Yes, this is exactly the plan. I’d like to continually build it up over the next two decades, and prioritize keeping all modules in production.
Early EuroRack frontier days were different – each product release was a limited batch, and an experiment! Now it’s been proven these products have staying power, so we’re interested in developing our “player’s set” or “52-card deck” – and focus more R&D time on lots of continual documentation and content support to go along with it.
So any future revisions to the modules will be Mk2, Mk3, etc. as opposed to entirely new products, and focus on the same format/identity.
Among the first modules we’ve released there are three groups:
“hybrid powerhouses” (FKG3, ESG3, DSG3, DWO3, SMX3, etc) that are combo functions designed to deliver lots of functionality to a small system. Think of these like ambassadors and paragons of their functional groups: keyer, shapes, encoder, osc, mixer, etc.
“triple CV functions”, triple key gen (Keychain), triple hpf (Contour), etc.
Definitely will be in the $499 category, it’s got a huge multiplier count! A 12HP positioner/rotator (Navigator Mk2) would be natural frontend for DSG3 as well.
You’re correct. What makes DSG3 a shape generator is just the HV mix oriented signal path (two input chains going into multiple HV output mixes.) So it’s a shape generator with the hood popped open. By “shape” in the video synth context I mean “enclosed figure”… usually a quadrilateral. That is, a shape generator must have the ability to create a pattern that has an enclosed boundary inside the space of the screen. Other words I guess would be “complex ramp” or “2D ramp.”
But when breaking the normal to the H and V ramps, and inserting an arbitrary voltage, video image, whatever, DSG3 is not “generating” a “shape” in the sense of creating a 2D geometric figure. It’s performing a series of transforms on that incoming voltage. And with two input voltages, DSG3 additionally performs logic/blending operations. Right?
We can still consider it a “shape” generator inasmuch as any function curve has a shape. So although the naming of the module might not immediately paint a full picture of what it can do, it’s not misleading or inaccurate. But I think the messaging can be improved a little here. This is an analog function processor with video HV ramps built in.
Beautiful work, I’m having so much fun. Thank you again.
In the short time I’ve had to play with these, I can say they absolutely pack a huge functionality into a small form factor. Utterly amazing what I can do with just two or three modules! Retail pricing is fair considering the versatility, build quality, forward-thinking “modular module” design, long-term commitment to support, and smart adaptation to supply chain issues. I’ve said it before, but LZX has really arrived, Gen3 exceeds all of my expectations.