Shutters, surfaces, & mapping

[Luke Walrath]

Hi guys,

We have used Qlab for years for our audio playback and recently for simple video playback. We are doing a production now that requires projecting different video files onto different surfaces during the show. The locations of the surfaces change throughout the show. Here's the catch: we have one projector (though it's powerful).

How do I use QLab to do this? I realize software like MadMapper can handle the mapping, but it has no show control to it (at least as far as I can see). Is it possible to map it all out in MadMapper and then fire it in QLab?

Any help would be appreciated. Thanks!

[micpool]

Qlab /Settings /Video Create a surface

Assign a new screen and select syphon

Syphon output appears as input in Mad Mapper

Mic

[Joshua Langman]

You can probably do this entirely within QLab 3, though more details would be helpful. One important question: do the surfaces ever move while video is playing on them? If not, easy. Just set up all the surfaces you need under video preferences. If, say, you're projecting on a flat and the flat moves during a projection cue, you'll need to use the full projector as our surface and map the video onto the flat with custom geometry (which works in three dimensions). Then you can use Fade cues to follow the movement of the surface in space.

[Luke Walrath]

Joshua, thanks. Below are all the details (I apologize in advance for the length).

The main surfaces we have onstage are: (1) 3 legs hanging on either side of the stage; (2) the main cyc along the back wall; (3) a flat (screen) that will fly in and out occasionally roughly over center stage. 

We also want to project onto the following surfaces that get moved in and out during the show: (1) about 20 boxes that are being held by actors (these actually move while being projected upon); (2) 2 walls that get created on either side of the stage with those boxes; (3) a large cape worn by an actor standing on a staircase; (4) the actors' bodies as they stand in a line.

What's tripping me up is the difference between "surfaces" and "screens" in Qlab 3. We only have one projector, so I'm assuming I only have one "screen" since that's my only output destination, correct? If that is the case, how would I map these videos onto the exact locations of 20 boxes spread around the stage? Do I manipulate the "screen" or the "surface?"

Thanks!

[Joshua Langman]

"Screens" are the physical display devices that are connected to your computer. You have one screen.

"Surfaces" are imaginary and unlimited. They basically correspond conceptually to the things that you're projecting onto. You can have as many as you want. So each leg, the cyc, the cape, etc can all be different "surfaces." You make surfaces in the sideo settings. You can use "shutters" (drag from the edge of the video proxy in the surface settings) to make rectangular surfaces within the throw of your projector and/or use "masks" which are arbitrary b&w image files that block out part of the projector's throw to narrow it down to a certain shape. These can have soft edges as well. (Note that with the current implementation, a mask on any surface that's in use will block out video on all other surfaces as well. I think this might be changing in a future update …?)

So, you have:

— legs: totally doable as surfaces in QLab

— cyc: easy, just make a surface

— flat: easy, just make a surface

— "about 20 boxes that are being held by actors": uh … with QLab? … not so easy. I would not attempt this with QLab. Now you're into the realm of Isadora or MadMapper etc. You probably want to do something like put infrared LEDs on the boxes and track their movement in real time, right? QLab will NOT do that for you.

— two walls made out of boxes: easy, make them surfaces

— a cape: pretty easy to do as a surface with a custom mask, unless the actor moves a lot

— actors' bodies: depending how much they move and how precise you need to be, this is also iffy as far as QLab goes. Maybe doable? But see note on boxes.

Does this help?

[Joshua Langman]

Just to elaborate a little bit further about the boxes thing:

If you had, say, twenty boxes scattered around the stage and you knew exactly where they were going to be, you could open a blank Photoshop or Illustrator document on your projector display, trace the boxes in white against a black background, and use that file as a mask to block out everything on a specific QLab surface except the boxes. BUT this would essentially only allow you to treat all the boxes as a single surface (though you could move videos around with custom geometry).

OR you could define twenty specific surfaces within QLab, one for each box (wither using shutters or masks). Then you could easily send a video to a specific box. BUT (with the current implementation of masks), you could NOT send video to more than one simultaneously.

With both of these scenarios, you also could NOT follow the boxes as they move. SO you could use the full throw of the projector as your surface and use custom geometry to map video files to boxes. No limitation on the number of video files you could play simultaneously (other than your hardware), AND you could move them with Fade cues. However, this still supposes you always know exactly where the boxes will be. Which seems pretty unlikely, especially when they are held by actors, which is why I assumed you'd want a way to track them in real time.

Like this: http://www.marcotempest.com/screen/PublicMediaPlayerV3/language/en/mediaAssetId/1298466065220

He uses infrared LEDs at the corners of the projection surface to track its movement in real time. Don't know what software he's using (something custom that he built for himself, I think). And as a I said, QLab won't do that.

[sprsndman]

Just curious, I know you can't play the same video on multiple surfaces at this time, is that a feature that might be coming? or not really?

Thanks!

[Luke Walrath]

This discussion is great. Thanks so much for all the help. Yeah, I figured the box scene was going to be a tough one, which is why I began to look at MadMapper. One of the problems I see with MadMapper is that you can't use more than one input at a time. Since we were going to film each of the 20 actors and project that onto their boxes, I don't know if MadMapper would work.

And even if it did, I still don't know how to integrate it with QLab. I followed micpool's instructions, and that allows me to fire a video cue into MadMapper and then, I imagine, directly out to the projector. But what if the mapping configuration changes during the show? Do I have to go into MadMapper during the show and change the presets? Seems messy and risky. I'm a complete newbie to all this, so I may be way off.

I may do what you suggested and create 20 different surfaces mapped to the location of the boxes and PRAY that the students hit their marks. I can then group the 20 different videos into one group cue and fire that. I'll at least try it.

Another surface question, though:

I'm still confused how you make keystone adjustments, etc, with surfaces if the only cropping tool you have is the horizontal and vertical shutters. Is there something I'm not seeing? I see how to make the X, Y, Z adjustments using custom geometry, but that's not as precise as using corner pins.

On Thursday, March 27, 2014 7:21:22 PM UTC-6, Joshua Langman wrote:

[micpool]

On Friday, March 28, 2014 4:40:16 AM UTC, Luke Walrath wrote:

I'm still confused how you make keystone adjustments, etc, with surfaces if the only cropping tool you have is the horizontal and vertical shutters. Is there something I'm not seeing? I see how to make the X, Y, Z adjustments using custom geometry, but that's not as precise as using corner pins.

I'm not sure Im following you. There are corner pins on surfaces. The trick is to make your surface at the projector resolution, then you will always have the corner pins visible You can then move the four corners to create the smaller surface. Does that do what you want?

Mic

[Joshua Langman]

"I may do what you suggested and create 20 different surfaces mapped to the location of the boxes and PRAY that the students hit their marks. I can then group the 20 different videos into one group cue and fire that. I'll at least try it."

Note that if you create these surfaces with masks, you will not be able to play video on more than one box simultaneously, because one surface's mask will block out the other surfaces. Not sure of this is true if you create the surfaces with shutters and corner-pinning.

[micpool]

> Note that if you create these surfaces with masks, you will not be able to play video on more than one box simultaneously, because one surface's mask will block out the other surfaces. Not sure of this is true if you create the surfaces with shutters and corner-pinning.

The area outside the corner pins is transparent so you can do any number of corner pinned surfaces on the same screen/projector and see them all

Mic

[Luke Walrath]

When I create a new surface, the "screen" is showing corner pins. The "surface" only allows me to use shutters. I realize it works fine if the physical surfaces are flat to the projector, but what if they're at an angle to projector? I understand that adjusting the corner pins on the "screen" will help adjust the perspective of the image, but how to I use the corner pins to actually adjust the shape of the projected image?

Or is that not possible without creating a mask in PS or AI?

[Luke Walrath]

In regards to MadMapper:

How do I allow multiple inputs in MadMapper? It seems like you can only use one input, which would be fine if you're using variations of the same image/video on multiple surfaces. But what if you want to project different images/videos onto different surfaces at the same time?

[micpool]

Corner pins are by definition for producing 4 sided shapes. If what you are projecting onto  has 4 sides, regardless of their lengths and angles you will be able to shape your image to the screen.

If it is not 4 sided you need to make masks. Unfortunately the black portion of masks in Qlab is not currently transparent so effectively you can only apply one per surface.

In your cube example the best you could do in Qlab is use three surfaces per cube for each 4 sided face visible to the audience.

If you use Mad Mapper you can change the presets using DMX.

It is possible to change the presets via MIDI. Unfortunately this feature won't be fully implemented until V1.6. It is a hidden feature in 1.5 for testing. If you search Mad Mapper Midi you will find the terminal command which enables this feature but I don't think it is officially supported.

Mic

[micpool]

You make a composite video with all the video material you want to project.

e.g if I had 10 circular screens at various angles  I would input a source video with 10 undistorted circular images in the same video frame.  I then draw round these images in the  input screen. Each of the circles I draw on the input screen will have a corresponding circle on the output which can be distorted, rotated and scaled to match the projection surface.

That is the essence of mad mapper. You draw a shape on the input and apply transformations to that shape on the output.

Mic

[Luke Walrath]

Mic,

Thanks. Just punting here: is there a way to program a hotkey (via MIDI or some other method) such that you can fire a cue in QL that syphons out to MM and advances the preset in MM at the same time? 

[Drew Schmidt]

Yes. There are other methods of triggering MM. You'll have to download the beta from their blog though. If you enable MIDI in MM and, QLab then recognizes MM as a MIDI out. They have OSC too but I haven't ised that. Just saw it existed.

I've used this beta in a production setting by force, and it never failed me.

[Luke Walrath]

Again, thank you...all of you.

So I have the 1.6 Beta loaded and I've figured out how to fire different cues into MM by creating different Syphon surfaces.

I've got the preset advance hotkey set in MM, how do I sync that to QL3?

[Joshua Langman]

If MadMapper takes OSC (not sure, I've never used it), then the easiest way would probably be with an OSC cue in QLab. You could also maybe use AppleScript cues in QLab to send key presses to MadMapper, but someone else would have to explain how to do that.

[micpool]

Heres a couple of screenshots

an example mad mapper preset including building one of your boxes

an example Qlab 3 setup for 3 faces of a cube in front of a cyclorama projection using 4 surfaces (open this in preview.app select actual size and use the scroll bars to move around it to see what is going on in the screenshot)

Mic

[micpool]

In the edit menu you will find the MIDI mapping option

Set up the window as in attached screenshot

In Qlab set up a midi control change to  send the preset number to  cc101

I only have this working on my 1.5.3 demo so 1.6 beta may be slightly different. On my versions the presets change but the preset pane doesn't, so there is no indication in mad mapper of what preset you are on.

I am sending the MIDI from Qlab to IAC bus 1. It seems to listen to all the IAC busses. Again your version may vary.

Mic

[Greg Leeper]

Or as an alternative to create your composite video live: Qlab is your media player, MM is your media mapper:

-Use Qlab to fire multiple videos

    -Use 'Display & Geometry: Custom Geometry' in Qlab to scale them down & position them so 20, for example, will fit in a grid inside one 1920x1080 output.

-Send that to MM with Syphon.

-Then use MM to do all of the geometry correction, scaling, positioning, etc to fit your boxes.

Greg

[micpool]

Indeed if you require the flexibility to change the surfaces individually that is the way to go . Simple Quick and dirty  example attached

The important thing, whether you do your image compositing in pre production, or on the fly in Qlab,  is to be absolutely rigorous with templates  in your workflow

Mic

[Drew Schmidt]

OSC would be the easiest option. But MIDI in MM came first, is a bit more well vetted (was introduced in a previous beta), so would advise using that in a beta over OSC.

[Luke Walrath]

Thanks guys. I will most likely create composite videos in either Final Cut or Premiere and then splice those up in MadMapper using the Syphon output from QLab.

I'll also look into the hotkey mapping suggested here either through Midi or OSC. I'll be honest, I have no experience in either.

Thanks all of you for this great advice.

[Luke Walrath]

Hi guys. I've struck out regarding sending a MIDI trigger to MadMapper while running in QLab. Micpool, I'm not seeing the ability to set things up like you have in my version of QLab 3. 

I've setup MadMapper such that the <RETURN> key advances the preset. However, I can't seem to get a signal from QLab to MadMapper. Any ideas?

[micpool]

Which version of Qlab 3 are you running?  It's sounds like you are missing features. If you haven't got MIDI then it sounds like you're on the Free Version. You need Audio Basic for the MIDI. You need to have Video Pro for the Syphon Features.

Heres the comparison table which lists which features come with which version.

http://figure53.com/qlab/buy/

Mic

[micpool]

On Monday, March 31, 2014 12:23:45 AM UTC+1, micpool wrote:

 You need Audio Basic for the MIDI. You need to have Video Pro for the Syphon Features.

Mic

 Sorry, Should have said the MIDI features come with video basic too.

[Luke Walrath]

Actually, I just figured out you were referring to an actual MIDI cue. I'm now seeing it, but am not finding a place for IAC buses. The only option it's giving me for MIDI Destination is "MadMapper In"

[micpool]

That's the one, if you can see a MIDI in for madmapper you don't need IAC. They must have added it in the new beta.

MM

[Luke Walrath]

Gotcha. Unfortunately, though, when I fire the MIDI cue (with an auto-continue to go to the next video cue), nothing changes in MM. When I go the "Controls List" in MM, I'm able to setup a "next preset" command in MIDI on Channel 1 and a Control Change number of 101. 

Stil, though, nothing happens. Any idea why?

[Luke Walrath]

micpool,

Ok. I got it to work once. Not sure how. I had to set the MIDI cue in QLab to a CC Value of 0, run it (nothing happened), then reset the CC Value to 1 and run it again. And it worked. When I tried to do it again....nothing. 

On Sunday, March 30, 2014 5:28:47 PM UTC-6, micpool wrote:

[micpool]

The cc value corresponds 1 to 1 with the preset number. So 0 won't do anything. In previous versions of MM changing presets didn't change the highlighted preset in the list but did change the actual preset in use.

Try making half a dozen presets just using the various checkerboard patterns in MM and then change cc values to match the preset numbers.

Mic

[Luke Walrath]

Okay. Tried that. Get this: it is now jumping to the Preset # that is DOUBLE whatever the control value is that I set in QLab. Eg: if I set a control value of 1, it jumps to preset 2. If I set a control value of 2, it jumps to preset 4.

Did I set something up that perhaps is causing this?