JG

Sigh...nearly everything in this rig is consumer grade, which is at least half the problem. There are two different models of extenders in the system, and I would agree they are the most suspicious part of the problem. I will say that the issue seems to occur per computer, not per display, and the center display computer has two of each type of extender. I think your point about trying to run a test on a simplified setup is wise. I'll give some thought about how we can achieve that with the way things are currently rigged and dressed. Jim

Production mode operation. (2) Netgear GS108 switches. One on deck linking the three display machines, one at the tech table linking the production machine with designer laptops. Then a trunk line between the two. I did have a bad power supply on one of the switches earlier in the week. I wonder if the other one is going and causing momentary packet loss... Jim

We have a system that's a little cobbled together, and I acknowledge that. I have 12 LCD displays being driver by three display machines. Two machines are identical with: Radeon HD7900 with the v17.4.3 of the Radeon driver 3.5 Ghz six core i7 The third machine is unique with: Nvidia Quadro M40000 with v385.69 of the Nvidia driver 3.5 Ghz quad core Xeon E5 The displays are (5) LG 55LS560e's and (7) LG 55LJ5500's The displays are cabled using HDMI over CAT5 exteneders, (6) Cable Matters, (6) Geffen Toolbox. We are running Watchout 5.5.2 The issue I'm running into is that the three display machines are noticeably out of sync with each out. It's slight, but definitely there. They seem to all go to black in sync, or close enough that I can't see it, but coming back to content there's a definite order of the four screens on the stage display machine, then the four on center, then stage right. I'm not sure where the best place to start trouble shooting is, but would appreciate any thoughts folks have. Jim

In the past, we've always had great luck triggering the main timeline using MSC from our ETC Ion (and before it our Obsession). On this show, I had an element that needed to be triggered on cue, while the rest of the timeline was still running. My solution to this withing Watchout was simply to put this extra element in an Aux Timeline. That's when I started to run into problems with the midi. If I set my "Midi Playlist" setting to "ignore", nothing happens. If I set it to "Map to Main Timeline" I can't trigger the aux timeline. If I set it to "Map to Aux Timelines" I can't trigger the main. My first try was to make a timeline called "1" that had all the midi triggers and it, in turn, trigged the main or aux timeline as appropriate. This was intermittently dropping cues, though. I tried putting all the content in timeline "1" and the extra element in timeline "2" but that was still inconsistent. Has anyone run in to this in the past? Ultimately, this became a convenient way to convince the director that this particular cue was a poor choice, but it is something I would like to be able to do in the future. We're running 5.5.1 on Windows 7sp1 Jim

Diving a little deeper into codecs, do I have the sources of these codecs correct? Implemented by Dataton: MPEG-2 H.264 Animation WMV Implemented through Quicktime Image sequences (Photoshop, JPEG, PNG, etc) FLV Others? MPEG-2 is still the preferred format, Animation is still the only option with an alpha channel, and image sequences are specifically discouraged. H.264 and WMV allow HD playback with smaller file sizes, but are much more processor intensive than MPEG-2. Jim mentioned MPEG-4, is this H.264, or MPEG-4 part 2, or something else? Are any decoders implemented through Windows Media? Anything worth mentioning that I missed? Jim

Excellent. And what about parallelism? If a display computer has a six-core CPU, can Watchpoint process, say, four tasks simultaneously? Is there even any need for that level of computing with the way the program handles playback? Lastly, assuming the program can run multiple threads in parallel, is there a theoretical limit? Could I build a machine with four twelve-core Opterons, for example? I freely admit this specific example would never by financially practical, I'm more curious about the upper bounds of the software. When does more computer equate to better performance, and when does it become just...more? Jim

Okay, so two follow-ups: So video files are decoded during playback? When still images are rendered to (let's call it) Watchout Format, does Watchout render each object seperately, or does it render all the images into one frame of playback? Jim

Am I correct in my understanding that when the display computer receives a file from the production computer it transcodes it into a Watchout proprietary format of raw RGB values and that this internal file is what's played back when the show is run? Jim

Jim, thanks for more clearly explaning sync variation. In weighing Radeon options vs FirePro for future systems, is screen geometry processed on the CPU or by the video card? Are there other parts of the workflow besides 3D tweens that are offloaded to the GPU? I suppose there's also a question, perhaps not objectively answerable, of how far do you need to go with 3D processing before you start stressing game orienteted consumer cards? Jim

We're currently running AMD Phenom II X6 @ 3.2GHz for both our display machine and our production machine. We have had no complaints with the performance and it was a big savings against the comprable Intel processors. We also have an AMD FirePro V7900. When I put this system together I was just learning about Watchout, but had been working with video cards for CAD for years and had a personal bias for workstation cards vs gaming cards, so I went with the FirePro. I have re-debated this choice ever since I made the purchase. We paid a very real premium to go with FirePro over Radeon. The V7900 does support an S400 module, but we don't have one and we've had no problem playing back 4 XGA outputs across the card. As I've learned more and more about Watchout and become more proficient with it I've come to understand, perhaps incorrectly, that the processors handles all of the video encoding/decoding, and the graphics cards processes 3D geometry as it relates to tween tracks. For our specific case, crunching the video is the main task for our applications. We do tend to do a fair amount of geometry correction, and I'm not clear to what extent the graphics card is involved with the 3D calculations there. Perhaps if Jonas or Jim have time they could speak a bit to that end. In any case, for our particular experience, AMD processors are completely analagous to Intel. FirePro offers very real advanages over Radeon, but they aren't necessarily advanages that all Watchout shows can leverage.

Anyone tried one of the new Decklink Mini Recorder cards, yet? Do they seem to hold up to live feed use in Watchout? Jim

What's the status of this bug? I'm about to go back into production and I'm trying to plan for the best way to be able to use alpha masking on both screens of a two screen setup. Jim

Thanks, Mike. Gives me a lot to think on about the best waying to generate alpha channels for photo style footage. Jim

The guide lines for codec settings list recommended settings for mpeg2, windows media, and h264, but not for animation. Any suggestions? Straight or Matted? Number of key frames? Jim

I had a similar problem recently, using Photoshop files. Two outputs 1920x1080, both alpha masked by a white png square, both showing the same picture with similar tween tracks applied to them. Left output (which I think was output two) stayed transparent no matter what the opacity setting. Right output worked fine. This was also only on one cue out of maybe twelve or so that used alpha masking. Using AMD Firepro v7900. Jim

Mike, I'm very inrigued by this idea and want to make sure I understand the concept since I'm not near any equipment to test with. The gray cue in the Aux timeline creates a seamless black field. Because it's on top of all other timelines, it will stomp on all negative space. But because the blend mode of the gray cue is set to "lighten" any content underneath this black field will still show through, leaving you with your pleasantly blended images floating on an even field of black. Hope I've interpreted this correctly. Jim

While we're on the topic of feature requests, I would love a little more flexibility in entering times. Personally I wold love if entering values from 60 - 99 were automatically interpreted as 1:00 - 1:39, if values with three or more digits were automatically interpreted as being in hh:mm:ss format, and if you could use + or - to enter incremental/decremental values relative to the start of the media file (or tween points) or the cursor position (for media files and compositions). Jim

We've been running into a similar problem. For us the problem seems to be specifically limited to pausing on masked-still images. We've found that having an auxilary timeline playing, even one with nothing in it, solves the problem. We've sent our files to Dataton and they've confirmed the problem but are still trying to determine the cause. Hope the time line fix proves useful. Jim

Out of curriosity, what is the warranty on the physical key? Life of the software version plus 1? Jim

For the sound un-initiated, if our display machine has built in multi-channel out are there any negatives to utilizing that instead of an external interface? Jim

Interesting comment, we ended up repositioning things to eliminate the edge blend because we were able to get a nice blend on whites and images, but we were struggling with getting a good blend on black. To be fair, the two projectors were fairly mismatched, so it was in no way an ideal situation. Jim

Fredrik, I think I have the problem solved. It appears to have been a combination of unnecssarily complicated mask files and needing to explicitly set transparency settings and not having enought breathing room in between my tween points and my pause cues (it looks like you need almost a quarter second on either side of the pause before you can make a tween adjustment). All of which I guess boils down to, me being new at this. If it rears it's head again in a clear maner that I can point to I will send tech support a show file. Thank you. Jim

After further work, I've narrowed the problem down not to the compositions but to the alpha masks themselves. Trying to specify the transparency in the media properties didn't seem to have any effect. Does anyone have any further ideas? Jim

I've run into this same issue. At first glance, it looks like the "primary" display is always one. The next attached display is two, regardless of what connector it's in, and so on. I agree, not very intuitive, but probably very home user friendly. Jim

I have a composition set to be free running that covers the full with of a two projector edge blended display. Both projectors are run by the same computer running v5.1 under Windows 7/64. When the composition is set to free running only the first output renders, the other remains dark. As soon as the composition is set to normal the problem goes away. The preview window on the production machine displays correctly. I'm afraid I'm at a loss for what to do. Jim