matkeane

To expand on JFK's suggestion - I usually then put the live input and the drop shadow/border layer together in a composition so that I can move and scale the whole thing as required without 2 sets of Tweens to manage. Placing the live input in a composition doesn't seem to affect the latency.

Hi Josef, Yes, and that is a really nice time-saver, but it would be even better if it were possible to change the 'advanced' properties of multiple cues (blend mode, etc) at once - similar to the way it's now possible to edit the specifications of multiple displays at once.

The recent update to thread about the Photoshop import script reminded me of a feature I think would be a big improvement to Watchout: a scripting API for show creation and modification. When building shows with Watchout, I seem to spend a large part of my time double-clicking and copy-pasting. I’ve already created scripts and small apps to speed up some things, but the current ‘copy-and-paste’ API has its limits. A scriptable API for Watchout would make automating repetitive tasks simpler and quicker. The creator of the Photoshop import script has also created a clipboard manager app for storing frequently used Tweens, etc, which is a good example of what could be simplified. An example I encounter quite often: I have a folder of client videos, each with a different duration. I create a Task with the first files, add the audio file with a generic input for the volume, a countdown and fade out for the end, and so on. Then I duplicate the Task and replace the video file, try not to forget the audio file, drag out the video file to the new duration, move the countdown, slide the opacity tween points to the new end point, rename the Task and so on and so on with my folder of videos, all the while hoping I don’t get distracted in the middle of the process and skip something. If, instead, I could write a little script (Python, Java, Javascript, Lua, whatever...) to loop through all the videos in a folder and build the tasks, with the video set to the correct duration, generic inputs created and assigned, it would be a big time saver. Imagine if, at conferences with hundreds of participants, I could point a CSV file at the Watchout text tool and build all the names with one click! When I’ve shown Watchout to people in the past - especially those familiar with other media server systems - their reaction is often that the Watchout UI is a bit ‘clunky’. A scripting API wouldn’t require adding any complexity to the UI, so it wouldn’t change using Watchout in the normal way. But, when I look at the eco-system of plugins and scripts that has developed around After Effects, for example, I think Watchout could benefit from the added flexibility. It could make it easier to add certain feature requests using scripts - re-usable tween curves anyone? Check out the Flow script for After Effects! Need to change the blend mode of 97 video clips without going crazy? Check out this little script… etc, etc. In the meantime, if it were possible to copy the duration of a video clip from the Media window, that would already be an improvement.

Quite often, yes. There's another thread on this forum about strategies for show backups/spares, but quite often I run shows with main & backup player machines controlled from 1 producer machine (usually with a spare producer laptop nearby). Occasionally, on bigger shows, there are redundant systems for everything - production, players, projectors - but then it all syncs to one timecode source... It usually comes down to the budget for the project.

It is still possible to create data-driven PSD files, although I think the name of the feature has changed in the CC versions of Photoshop - the data & variables functions allow you to create layers or files from a CSV data source. More info here: https://helpx.adobe.com/photoshop/using/creating-data-driven-graphics.html

@Mike Fahl: That makes sense now, thanks! I just did some quick tests and it works as Mike describes - but uses the STOP, not RESET, command. I opened up a recent project in which I had 6 audio jingle tasks, so I added a new composition into which I pasted a control cue, and then duplicated and modified it to create 6 cues to stop each of the audio jingle tasks. I pasted the composition into all of the audio tasks and now, starting any of those tasks will kill the other tasks, but not the task from which it was called. Are control cues in compositions actually supported behaviour though? They can't be created directly in compositions, but can be copy-and-pasted into place and then modified. This has turned out to be extremely handy in the past, but I'm never sure whether it's supposed to work, or just a useful bug. And, as Mike says, you still need to manually add a control cue for each task you want to kill, but at least the composition trick makes it reusable, and updatable in one place.

Somewhere way back in the epic Feature Request thread, Mike Fahl suggested that this is possible, but I've never quite worked out what he meant - perhaps he can chime in to enlighten us!

Hi jfk, The source and target cues were identical (except stage coordinates). It was somewhat related to the other thread about large resolution video playback... Long story short - I hit the HAP import limit at 8000px on the production machine I was using, so split content into 4 UHD slices. Each video clip was made up of 4 separate cues on the timeline but, of course, that then meant that any colour correction had to be applied identically to 4 clips, which is why I was trying to copy-paste hue tweens with a shared generic input. I set up (and tested) the tween and formula on the first video cue, and tried to paste it to the 3 other slices, but no joy. It was about 3am at this point, so I didn't really have the energy to go through all the options methodically to work out why it wasn't working but, as I mentioned earlier in this thread, once I had time to calmly reproduce the problem at the office (with the same media and also some small test images), adding a single keyframe (even if it's ignored by the tween formula) is enough to enable copy-pasting.

@jfk, @Erik - Thanks for that info about the HAP limits. Is the limit imposed by the GPU in the production machine, or is Watchout checking with connected players what their GPU limits are? Just curious, as I often run shows with production laptops which are much less powerful than the player machines - does that mean I might hit the import limit for HAP on production even though a player machine would be capable of playing back the file?

The maximum resolution at which a HAP video file can be imported into Watchout is currently 8000x8000px. I don't know whether that is a software limitation, or whether it is limited by the GPU. Either way, HAP media at higher resolutions must be split into pieces for playback. In general, the graphics company will deliver content at full resolution and quality (i.e. uncompressed, if feasible, or using lossless compression) which might, for example, be a ProRes video file or an image sequence (which makes partial content updates/corrections easier). The file size and data rate will usually make this unsuitable for playback. However, the next step is to use this as your master for slicing up and compressing the content, using a codec more suitable for playback in Watchout - HAP, MPEG2, h264, etc, depending on circumstances. The delivery format may also depend on your physical proximity to the graphics company - lossless codecs are more suitable for delivery on a hard-drive. If files are being sent over the internet, you may want to compromise with some form of compression to speed up transfer times, or use image sequences which can be transferred image-by-image rather than as one huge file.

Hi, It looks that 25Gb file is the full 13824x3464px resolution though. I thought your plan was to split the content into roughly UHD-ish chunks (in any case, Watchout won't accept HAP files above 8000px). On a recent project, I also ended up splitting content into roughly UHD resolution pieces which, at 25fps, were running at around 450Mbps on average. That was a 25 minute show but content was in 1-3 minute scenes which made rendering/copying/importing/updating easier to manage than dealing with content of the full duration.

In my experience, h264 - even with constant data rate and a fixed keyframe interval - is not necessarily decoded in perfect sync across multiple outputs. Also, you haven't said anything about the duration of your media. If your content is long, say 30 mins to an hour then, at the resolutions you're talking about, you might have to think about disk capacity if you use the Hap codec (512Gb SSDs can fill up fast). On the other hand, I've seen long-ish (17 minute duration, encoded by the client) h264 files require 2-3 minutes of preloading to get a clean start without stuttering. I generally encode video with either the Hap codec or as MPEG2 (you can find encoding tips on the forum). I haven't tried playing image sequences in a show yet. For very large resolutions like your video wall, it might simplify content preparation (no slicing required), but you would have to test carefully to make sure the data rate didn't exceed the capacity of the player machines.

Thanks Josef - I can't believe I missed that in the release notes!

I think I have narrowed this down... I first encountered it during a show just before Christmas, using Win7 Pro producer machines, and was able to reproduce it at home (also on Win7 Pro production) in both the show file and a new test file. I've just started a new test project and I can happily copy-paste opacity and scale tweens but not hue-and-saturation (copy option is greyed out in the menu)... that is, until I add a keyframe. Once there is a keyframe (for any of the hue/saturation/gain values), the 'copy' option is available in the menu. Same applies for other tweens. Why do I want to copy a tween with no values set? Because I'm using it with a formula linked to a generic input, so that I can control the hue/saturation of several clips at the same time. It looks like adding an initial keyframe with a default value allows me to set up a formula and then paste it to another clip, but it seems like copy-pasting of tweens is not always consistent though - opacity, scale and z-rotation tweens can be copied without keyframes, but not other tween effects. Scale and z-rotation I can almost understand, as they can be set in the UI without actually adding a keyframe, but I can copy an empty (no keyframes or formula) opacity tween track... Perhaps only tween tracks with a single property value can be copied without keyframes? So I guess it's not a bug, just user error (would be nice if copy-paste behaviour were consistent though).

Multiple, related, requests in one: Can we please change the defaults for 3D object import, so that they aren't silently scaled to 1000px? I first noticed this in 6.1.6, but 6.3.1 has the same behaviour. If Watchout insists on scaling a media file on import, can we at least generate a warning message so that the change doesn’t go unnoticed (and I dumbly start scaling stuff on the stage before I realise the source of the problem)? The Media window displays the object's pixel size, but it never occurred to me that Watchout would modify media specifications without asking/warning me. It would be nice if, after I have set the media settings of the first object imported to be at 100%, subsequent imports used the same settings, to save having to make the same change on every object. Currently, when importing several 3D objects where it's important to respect the relative sizes, I have to change the settings for each one. Better still, could we please add the ability to select multiple media files and change settings with one click? That way, I could just drag all the 3D objects into the media window, select them all and set their scale to 100% in one go. Or, for example, when the graphics person sends me 80 title PNG files with a weird alpha channel, I could select all the images in one go and change the alpha interpretation.

It would be really helpful if there was a simpler/quicker way to enable/disable displays and assign them to player machines than double-clicking each display one-by-one. Once the output ID is set in the display dialog, it would be great to be able to drag-select a group of displays (on the stage or in the stage list panel) and change the player, either in the contextual-menu, or maybe by dragging onto a player machine in the network window. I know that I can assign a player name to a different machine in the network window, but that's not exactly the issue... On a recent project, I had 21 displays, plus another 21 for the backups. Pushing media to all 10 player machines was slowing down show programming enough that I disabled the backup machines while working on the show with the client, then enabled them at night to push media to all machines. I realise I could have cut-and-pasted the backup displays into a text file as a workaround, but a quick drag-to-select then right-click to enable/disable would be quicker still. On the same project, a group of 14 displays required geometry correction. I did the warping on the main displays, then copy-pasted the displays onto the backup tier, but then had to manually reassign the player machines for all 14 displays one-by-one. If, for example, I could drag to select displays 1-4, and right-click to 'assign to Backup 1' etc, it would be a real time-saver. For managing backup displays, it would be even better (from a UX point of view) if I didn't have to duplicate displays on extra tiers at all - if it were possible to associate displays with more than one player, I wouldn't need to copy-paste geometry between sets of displays, I could just right-click and 'add player' to assign an additional machine. In the meantime, anything which speeds up the process and reduces the wear on my left mouse button would be a welcome improvement.

In Watchmaker version 6.3.1, I'm no longer able to copy/paste certain Tween FX between cues. After applying Hue & Saturation to one cue, and selecting the Tween track in the timeline sidebar, the 'copy' option is greyed out in the menu and the keyboard shortcut no longer works. I just opened the same project in 6.2.2 and 6.3.1 and I can copy/paste tween tracks as expected in 6.2.2. I haven't tried all the available tween effects, but Contrast & Brightness suffers from the same bug, while Opacity tweens copy OK. I didn't notice anything in the release notes about changes and, suffice to say that, on a project where the client wants colour correction applied to multiple cues, this regression is a real pain.

I'll take Mike's word for it that it's OK to force all conditions to zero - although I've also added a switch for it, just in case the default behaviour is useful in some circumstances. And, after a busier couple of weeks than planned, I found time to Mac/Windows builds of my helper app, which can be downloaded from my website: https://matkeane.com/project/watchout-conditional-layers-utility. The readme file in the zip has some instructions but, basically, conditional layers can be switched through the UI, or via UDP/OSC commands. The UI controls can be disabled if only network control is required. In addition to switching individual conditions, they can be assigned (in the UI) to groups and presets for more flexible recall and switching.

I'm quite happy to share once it's ready - I already have a couple of other utilities for Watchout on my website for download. I'm working on loading/saving settings for the app, since setting things up manually each time gets a bit boring and after that it should be usable, even if it's not going to win any awards for UI design! The apps are all developed in Java, since I generally work on Mac but want something which will work on Windows too. I have thought about switching to another language, but I've yet to find something really suited to cross-platform desktop development with a decent UI library (If anybody has any suggestions, I'm all ears). I have a project coming up in a couple of weeks where the app might be useful, so I should have something to share quite soon...

I'm pretty sure I added a feature request a while ago for the ability to change a single layer condition from within Watchout. Like you, I'd like to be able to change one condition without affecting others - effectively impossible without storing and recalling all the other layer condition values. This would make it easier to use conditions to manage the visibility of content. I also use conditional layers to manage the 'visibility' of layers containing control cues, which allows some simple switching logic within Watchout. I agree that 'setLayerCond x 0/1', 'enableLayerCond x' or 'toggleLayerCond x' would be useful additions. With this in mind, I've been working on a little external control app for managing layer conditions. Still a bit rough around the edges, but it allows for: toggling or setting a single layer condition, creation and recall of presets, groups of conditions with different behaviours (radio buttons, synched, etc). Layers can be switched through the UI, by UDP commands (from Watchout or elsewhere), or by OSC commands (not from Watchout!).

That's interesting and potentially very useful, but I wonder if it's intended behaviour, or whether overflowing the expected maximum value just happens to result in all conditions being set to disabled. I wonder, because I'm working on a little utility to make managing layer conditions easier for me and, shortly after trying 1073741824 instead of 0 to turn all conditions off, my Watchout production machine became unresponsive to further enableLayerCond commands and eventually crashed WatchMaker altogether. It could just be a coincidence or another bug in my app of course...

I've seen this behaviour before too when trying to jump between control cues in a timeline, in earlier versions than 6.2.2 and I think as far back as v5. The workaround I found was to split the background (unchanging part of the loop) from the foreground in two separate tasks so that I could fade the background image in and out in one task, and then jump between cues in another task. It works most of the time (still not entirely reliably) and complicates the programming a bit since you now have twice as many tasks to manage.

Just guessing, but is your 3D object a very complex model? I've experienced Watchout production becoming unresponsive in the past when I imported a 3D object with a very high vertex count. It's just possible that in your case, the production machine managed to import the model, but the display machines are struggling and appear unresponsive. The Watchdog process running on the display machines may have launched a second instance of Watchpoint, if it appeared that the first instance has crashed.

When I was looking at different playback software to try and find a solution to the seamless loop issue (http://forum.dataton.com/topic/2838-getting-out-of-a-loop-at-the-end-of-loop/) I discovered Mitti (https://imimot.com/mitti/) which not only makes the looping question very easy to manage, but can also output NDI with an alpha channel and can be controlled with network/midi/dmx commands (although sending OSC commands from Watchout is still a sore point). I managed to get a 6400x1200px HAP alpha file playing smoothly over my Watchout background using an NDI input. More expensive than VLC - obviously - and some of the alternatives, but I'm still tempted to buy it because it seems both versatile and very simple to use.

Did you by any chance open an older project in which the audio files were already imported? I had a similar problem recently with audio not playing in a old (probably 6.0.x) project from last year when opened in 6.2.2. When I looked at the audio channel configuration for the cue, no audio channels were recognised. I had to re-import the media (double click the cue in the media window to reselect the source file) to force Watchout to update the media completely - a simple refresh didn't seem to do the trick.