Automark Refuses to Die Unless I Pay it $100 worth of Time

If you need the light to be somewhere but off, you should give it an intens value that indicates to both you and the console (auto mark) that the light is active and needs to stay put. I make an intensity palette of "Off" that is at 0.1% for every light, and use that when I need the lights to stay put but not output any light. This is enough intens to make the console think the light is active, and not try to mark it.

As eos moderator has pointed out, on Element you can't disable auto mark, but using this 0.1% intens trick will make it stop trying to intervene.

Right, the super dim thing is what I've been doing occasionally, but from a software development perspective, this is very very dumb. Users shouldn't be forced to keep a light on just because the software can't understand simple, simple requests. 

I'm the first to rail on ETC for poor design choices, but you've got auto mark on and so it's trying to mark the lights automatically.. You're in Element mode, so the console assumes you're a poor programmer trying to do simple things. If you want to change that assumption, change modes and disable auto mark globally.

But even then, using dark moves is bad programming practice. It looks like a mistake, and when a new programmer subs in they're likely to clean it up and break your show. Flagging the light with IP "Off" indicates to future you and potential subs that the light is doing something important in the dark. I don't really see how any feature ETC could write would be better.

Using this also lets you have the lights mark for the dark position. Say in cue 10 I need the lights to do a fly out, so I have them "Off" at the starting position in cue 9 and FL in the ending position in cue 10. I can put a mark flag on the lights in cue 9, so that sometime before cue 9, the lights will mark to that flyout start position. It's much cleaner than manually moving them to that position in cue 6 and hoping you remember you can't use those lights when making edits.

The feature they could write that would be better would be not allowing me to set Mark Time to Disabled if setting it to that doesn't disable mark. I'm still not sure what that field actually does. I'm in Element because the effects window doesn't properly display in Eos mode on Nomad. 

But holistically, specifically in the context of timecoded design, providing the option to use keyframes as opposed to cues completely erases the entire concept of marks altogether. A keyframe is just an inverted cue. 

Mark Time Disabled means the marks inherit the cue time. Best practice is to put your mark flags on part 20 of the cue so that you can set the mark time discretely without affecting the rest of the cue timing.

If you set Mark Time to say 2, every mark will take 2 seconds, regardless of cue time.

I'm sorry you're having issues getting the effects window open, this isn't an issue I've encountered. Do you have the CIA locked? Bottom right lock icon. If that's locked closed, tapping [Effect] once won't open the effect status screen. Double tapping [Effect] will still open the Effects tab.

You can also just remove the concept of channels, cues, and everything else and play back the raw sACN stream. We abstract these things away for speed of programming. The mark system as written allows us to not only know where a light is pointed, but why it's pointed there, and not have to worry about when so much, the light will automatically move into position at a good time. I use timecode all the time, and still rely heavily on my timecoded cues marking lights at the right time for me.

I can't reproduce the effects issue right now of course, but it was always intermittent. I would press Effect Effect and only half the time it would come up with the actual editor. The list of effects would show, just not where you edit the effect. Could only edit the effect with softkeys. I'll reread your post again next time I start a new show file. 

With cues, you need good marking strategy and good marking execution to keep everything where it's supposed to be in the dark so that you don't have unwanted visible moves. With keyframes, it just works. No strategy needed. 0% of consciousness needs to be taken away from art and put into marking fixtures. 

With a keyframe, you just say you want it to look like this at this time and hit record inverted cue. The stage will look like that at that exact time. You need a clock for it to make sense though of course.

Ah, same issue, locked closed CIA. Even if it's locked closed or just closed and not locked for whatever reason, there a pointed up or down triangle next to the lock that you can use to toggle the CIA open or closed as well.

I'm not familiar with "keyframe" style marking, and searching for it in reference to other popular consoles has yielded no results. What console have you used in this way?

Now I'll know to check the lock icon. I definitely had CIA open, but it refused to display the effects editor. Thanks so much for the tip!

Keyframing isn't a style of marking, it's just a way of thinking about controlling motion that completely erases the entire concept of marking altogether. It's an animation term.

A cue changes behavior after. A keyframe changes behavior before. With a cue, you say at this time, start going to this look. With a keyframe, you say make it look like this it this exact time. Cues are requests to start going to a look, keyframes are snapshots in time. 

Imagine a video on a large screen on stage. In the video is one of those screensaver logo screens where the logo bounces from wall to wall. You can freely scrub through the video and the video's timecode is synced with the light board. That's the setup.

Your job is to program a mover's beam to copy the logo's path the way a spot op would. How might you do it? Well, you're going to run into a severe problem. The way light boards work, you can't tell the board to put the mover on x position at x time. It just doesn't work like that. You can't train the mover at logo collision point A and hit record cue. Sure, you can do that, but you're then going to have to set a highly precise transition time and then you're going to have to set the timecode stamp to exactly the right time to get it in that location at exactly the right time. Sure, you can get it done eventually after enough time, maybe all day for a 60 second clip. Sure. 

Here's the solution: keyframes. A keyframe is how you tell the computer you want the beam at this precise location at this exact time. You scrub to collision point A, train the mover to that spot, hit record keyframe, then you scrub to the next collision, move the mover into position and hit record keyframe. In 2 minutes, you're completely done. 

Now in real life, you wouldn't be trying to get the movers to copy the motion of a logo bouncy screensaver screen, but there are numerous real life examples where you can only solve a problem efficiently using keyframes. For example, a dancer is making wide arm gestures and you want the mover to copy her movements move for move in perfect sync. Can't do it with cues. At least not without a ton of frustration. 

LD's tend to think about what can be done with light design within the confines of cue-based-thinking. Simply being exposed to the possibility of keyframes opens up a whole new world of possibilities. It changes the way you think about motion.

jordanhenshaw said:

I can't reproduce the effects issue

Would you happen to have performed a software update in the meantime? What you're describing was a bug that was fixed in 3.1.0, where the Effect Editor wouldn't correctly populate the CIA if you tried opening it while you had a selection in the command line.

As I mentioned this should have been fixed for close to two years.

Ok, so to be blunt, it's not a real thing. I'm familiar with video keyframes. Lighting doesn't use them, but I of course welcome you to develop the technology.

In your bouncing logo example, I would do one of two approaches, depending on how video programmed it. If the logo is an object they are translating across the screen, I would ask them to connect it to PSN and send me the datastream, and use something like Zac Tracks or Blacktrax to read the PSN and point a light at the logo. If it's just a movie and I can't have position information, I would set up timecode to start with video, ask what time each move started, add a cue to that timestamp, and then how long until the next move, set the cue time. Probably 60 seconds per move worth of programming. This does however assume linear motion, and takes some amount of programming time, so PSN is highly preferable. Regardless, the entire 60 second clip is programmed in maybe 15 minutes using existing techniques and technology.

Your proposed keyframing solution doesn't really remove any of the friction - I still have to know what time each position happens at (which I can still learn from paused timecode with normal cues on Eos) and point a light there., it's just doing the "Cue time X Enter" command for me in the background.

You're not going to make a ML follow a dancer's arms in sync no matter how you save the data - you have to have real time tracking. Get Zac Tracks.

These are solved problems with existing techniques and technology. I'm always willing to be wrong and adopt better technology as it develops, but I don't think any of your examples require me to abandon the way current cues function - you can always set cue time to equal the difference in time between two cues in your timecode sequence.

The real opening your mind to the possibilities of motion is real time tracking and PSN. You tell your light to "point at X", and just give it the data on where X is. This is routine to use for followspots, video objects, and scenery - you can calibrate with known positions (Chandelier 1 is at XYZ when the motor is 20' spooled out) or use radio/IR trackers to allow complete freedom of movement. Do a little zoom calibration and your fixtures can even keep a consistent size as they move around with complete freedom.

It's very close to being a real thing.

Linear start/stop motion is easiest. Now try to track a 3D audio object circling the room, curving, and doing all sorts of turns without ever stopping. Obviously you'll have gimbal lock issues, so you'll also have to swap movers mid-move without making it obvious you swapped. 

That's a real problem I worked on once. Couldn't finish it because cues made it impossible. 

Last week, a Visual Lighting Sequencer wasn't a real thing either. Next week I'll be working on getting Intensity/Pan/Tilt keyframable inside the sequencer.

It's very close to being a real thing.

Linear start/stop motion is easiest. Now try to track a 3D audio object circling the room, curving, and doing all sorts of turns without ever stopping. Obviously you'll have gimbal lock issues, so you'll also have to swap movers mid-move without making it obvious you swapped. 

That's a real problem I worked on once. Couldn't finish it because cues made it impossible. 

Last week, a Visual Lighting Sequencer wasn't a real thing either. Next week I'll be working on getting Intensity/Pan/Tilt keyframable inside the sequencer.