Layering DMX effects and model groups / sub-models for moving heads

[TypingOffKey]

What's the best way to set up groups and layers for complex moving head animations?

For example, I would like to be able to use area-mapped effects like the fan, bars, etc. on the RGB, shutter and / or dimmer channels as a group of sub-models while controlling the pan and tilt channels separately, maybe per-model or with another layer or group of submodels.

I have 24 moving heads in a ring and built a 3-channel submodel for the RGB filters (it's an arc lamp with dimmer / shutter, not LED) another submodel for the dimmer channel, and yet another submodel for the shutter / strobe channel. I then put all of the submodels of a particular type (say, the RGB filters for color) into a model group, and tried to apply an effect to that group. No matter which combination of "Default Layout Mode" in the layout tab, and "Render Style" I choose in the Layer Settings tab of the Sequencer, I can't get effects to respect the physical layout of the heads, like a wipe in transition on those channel groups or an actual rotating on / off or color sweep like on the fan effect. The best I could do was to set up the "Default Layer Mode" for my shutter / strobe channel as "Single Line" and then I can at least turn them on and off sequentially with a "left to right" wipe  (see attachment) - but I really want to render that effect as it is laid out in the preview.

Any suggestions / ideas?

Also - what would it take to bump up the maximum channels on the DMX effect? I have some Clay Paky Alpha Spot 1500 HPEs which use 40 channels each. I'm currently skipping some channels I'm not using and controlling it with two models and two copies of the DMX effect.

[TypingOffKey]

I forgot to include the second part of my question.

How should I actually layer DMX effects? Say I have a motion effect (using pan / tilt DMX channels) that I want to keep going over a longer period of time, but I want to mix up the color, gobo, prism, etc. in a layer above it, without overwriting the DMX channels that are controlling the motion. Is there a way that I can tell xlights to ignore / discard the (unwanted / unused) DMX values from a higher layer effect so that they don't screw up the motion on the effect in a layer below?

[Gilrock]

I designed all the DMX stuff but I don't know all the ramifications of trying to use them in groups.  I believe I only designed 2 effects to support DMXModels and that's the On effect and ColorWash because it didn't make sense to run a fan on a model with the equivalent of only 1 pixel.  I'm not even sure what happens if you try to put a Fan on a group of DMXModels.  That effect and all others except On and ColorWash do not have any code to try to route the data to the proper red, green, and blue channels.  As far as trying to blend commands from different channels all I can say is experiment and look at the data on the nodes.  Try different blending modes for the effects.

If one of those showed up at my doorstep maybe 40 channels would show up in the program...lol.

[TypingOffKey]

I understand why trying to route effect data to a DMXModel or group of DMXModels would require some more work, but what about making submodels for just those 3 RGB channels of the DMXModel? Shouldn't a group of 3-channel submodels be treatable like a group of standard RGB pixels and allow layout-rendered / positionally representative effects to be applied to it?

Also - I'd be happy to split the shipping cost (or deliver if you're reasonably close) of sending one of these beasts back and forth to you to play with. They came off of a professional tour, and I have full padded flight cases for everything. The 1500 HPEs are about 110 lbs each, plus another 50 lbs for the flight case itself, and the case is about 38" x 36" x 24". They also pull about 2 kW a piece, and require 200+ vAC power. Even the Alpha Beam 700s that I'm using in the ring technically use 26 channels, but I'm ignoring some that aren't critical to my application. They're a lot smaller, lighter, can run on 120 vAC and I'd be happy to send a couple of them your way instead / also.

[Gilrock]

Not sure our submodel feature would work because the DMXModel is hard-coded as a single channel string type.

[keithsw1111]

I wonder if you could do this with the block of channels model.

But I am not sure what would happen if you submodelled it. It clearly wouldn’t look on the screen right but it may let you do what you are after.


Sent from my iPhone using Tapatalk

[keithsw1111]

I also wonder Gil if it would be possible to move the colour channel locating smarts in on and colourwash into the channel placement code and handle it where dan has handled the new rgbw pixel support.


Sent from my iPhone using Tapatalk

[TypingOffKey]

I wonder if you could do this with the block of channels model.

But I am not sure what would happen if you submodelled it. It clearly wouldn’t look on the screen right but it may let you do what you are after.

I'm not as concerned with it looking properly in the preview (and can't I just assign the same channels to multiple models if I really do care?) as long as I can still name the channels I will understand what's going on when I sequence it. I typically end up spreadsheeting the pan / tilt angles anyway since a lot of times I'm trying to point at specific things or cardinal directions and need to do some trigonometry to back into the DMX values.

If we're talking about feature requests, it would sure be handy to display and allow input of an actual pan / tilt angle instead of DMX value for the moving head models. If we want to get fancy, building in an automatic calculation for aiming multiple lights at a target, or pointing at a cardinal direction (using the pan orientation offset, and inputting a global offset from north or something) would be super slick!

I've never used the block of channels model - will have to try that one out.

[keithsw1111]

I did a video on the state effect that talked about predefining targets with moving head lights.


Sent from my iPhone using Tapatalk

[TypingOffKey]

When I mention pointing at a target, I'm talking about multiple lights that are spaced out in a particular way (this would require actual measurement between lights and target, etc. in the layout) automatically calculating their pan / tilt angles trigonometrically. So, for example, you could have a group of 8 moving heads spaced 5 feet apart in an east-to-west line, which could then be instructed to all point at the same arbitrary point in 3d space at a particular distance, height and angle away from some predefined reference point of the group.

For example, in my setup I have the 24 moving heads spaced evenly in a circle with a 30 foot radius. Off in the distance about 60 yards to the north west and about 35 feet off the ground is a flag on a flagpole. If I want all 24 moving heads to point precisely at the flag, without manually guess-and-checking, I have built a spreadsheet where I can plug in the distance, height and azimuth of the target, and then it calculates the pan and tilt angle for each head. It would be fun to have some more tools like that available right in xlights for moving head power-users.

[Gilrock]

I also wonder Gil if it would be possible to move the colour channel locating smarts in on and colourwash into the channel placement code and handle it where dan has handled the new rgbw pixel support.

I looked and can't really tell how the RGBW stuff works but I'm also skeptical that the technique would work for this.  The render routines try to write into the nodes and for a DMXModel you basically need to reroute the output to a completely different set of channels based on the model parameters.