Jump to content
Hash, Inc. - Animation:Master

luckbat

*A:M User*
  • Posts

    2,750
  • Joined

  • Last visited

Everything posted by luckbat

  1. Hi dre4mer, thanks for remembering me! The "Ebon" project was put on hold due to lack of resources. The test scene I had been working on remains 80% complete or so, but getting that last 20% done proved beyond the limits of the technology I was working with at the time. Although I've moved on to a new celshaded project since then, I'm still strongly considering finishing the "Ebon" scene when I have some more free time, and posting it publicly. At this point, it would probably only require a couple more months' work. Just wanted you to know that I'm still out there, still doing toon-style animation.
  2. Man, you think I'd at least test my own links before posting them. Sheesh. Fixed. Thanks Rodney.
  3. Okay, this is going to be a little weird, animating backwards in chunks like this, but you work with what you have. So, remember shots 4 and 5, posted way back in August? Well, to that pile I add the newly completed shots 1 and 2. Not only is there no lipsync in these shots, making the title of this thread increasingly inaccurate, but since there's no sound in this new clip, I couldn't really cut it together with shots 4-5. So you'll have to do the editing in your head. Watch: Shots 1-2 (1280x720, 10.2MB) You'll notice a bit of shot 4 at the end, there. Shot 3 was removed during the storyboarding stage long ago, so with these last two clips, you now have the first four shots of the scene.
  4. They were slain (offscreen) by an unknown assailant. Ebon lost over 20 of her "people" during the same incident. It's all in the script: http://ebon.org/about/
  5. New (and hopefully final) 720p version of shots 4 and 5: http://ebon.org/images/forForums/Shots4-5_...-2_1280x720.mov Changes in this edition: DOF rack-focus now more visually accurate Slight remodeling of Ebon's lower lip; new "mouth-corners" pose added Laborious keyframe revisions so minor as to be imperceptible to everyone but Robcat Room tone, some SFX added; all dialog replaced with alternate takes for some reason
  6. Oh, they are, but fast is relative when the shot is 25 seconds long, and you're rendering at 1280x720. The toon lines take a little over one minute per frame to render, which is good, but that's still 600 minutes, which definitely makes it an overnight render. The set is rarely a problem to render, because for most shots, it's only a still frame. That, plus the fact that I made sure that all the lights in the set are either Z-buffered or non-shadow-casting. The real time-killer is the raytraced sunlight. I gave up when I started seeing 30-minute-per-frame render times. In the end I went with ambient-light maps on the walls, and a separate volumetrics layer rendered out at quarter-res, scaled up and blurred. Even at quarter-res, the five-second volumetric layer in shot 5 took 36 hours. The set uses true DOF, since I don't have a depth map. The characters are rack-focused in post.
  7. Just for "fun," I decided to render out shots 4 and 5 at 1280x720, the final resolution that this project will be mastered at. It was a demanding test of my entire production workflow, and it forced me to refine many of the techniques I've been using to put this animation together. It also brought my computer to its knees--the 25-second sequence you see here represents many, many overnight renders composited together. Just how big is a 720p animation? Oh, it's pretty big. (I took the liberty of generating a 720x400 version for those of you without the latest hardware--at least you'll be able see what these two shots would look like on DVD...) Download: FULL-RES 1280x720 version (h.264) Download: DVD-RES 720x400 version (divx) NOTE: These clips are out of date--There's a newer one at the bottom of this page...
  8. The line thickness is 0.7, and the bias varies from group to group. I render lines separately at 720p for these test renders, and downscale them in post. The final short will likely be done at 720p, so I'll render the lines at 1080p for that—they'll look a little thicker than they do in this test. No settings for the toon shader; I use regular lighting.
  9. Here is the full 16-second version of the shot. I still have a few tweaks left to make, but this is pretty much it. Three lines of dialog down, eight left to go. http://ebon.org/images/forForums/lipsync_shot11-2.mov
  10. That's correct--you may be thinking of this thread: http://migrate.hash.com/forums/index.php?showtopic=15373 That was a couple of server migrations ago though, so some of those older movie clips are lost to the sands of time...
  11. That wouldn't surprise me. But with keyframes already built for 24 fps, re-optimizing the animation for 12 would be no small undertaking. I promise to do some tests, though, and if 12fps ends up looking better, I'll consider making it my standard on subsequent shorts. Yes, they are. Agreed. I was supposed to do a blink pass, but I forgot!
  12. Hey folks. Just wanted to pop by, in case anyone was wondering where I'd gone, or heard I'd switched to another app, or something. Rest assured I'm still plugging away, spending my nights and weekends in A:M, bringing "Ebon" to life. Here's a new lipsync test, for example: Six-second clip (1.6MB): http://ebon.org/images/forForums/lipsync_shot11-1.mov There are a number of reasons why I haven't been around much lately--one of them is simply that I've been posting my work-in-progress details on my own production blog instead of here. Nothing personal, honest! It's all part of the publicity-building machinery, laying the groundwork for even more ambitious animation projects to come. In the future. If I ever finish this short. Anyway, I hope you're all well, and do let me know if you have any questions...
  13. Sorry folks, the company's making me work over the weekend. Gonna hafta postpone it again. Looks like the response has been pretty sparse anyway, probably because of the holidays. Maybe January would be better. Apologies...
  14. See: Control Point Weights
  15. Hey folks, we're long overdue for another meeting! What say everyone to Saturday, November 18th at noon? Let's vote!
  16. Sooo amazing... Why must you set your multipass render setting so low...? We want to see every detail...
  17. Okay, that settles it. I checked the weather, and you're right. It's going to be a miserable day tomorrow. Let's postpone it due to inclement weather. I'll post an updated schedule shortly. Hopefully more people will be able to make it to a November meeting. Thanks for the update.
  18. Okay, we've had both cancellations and confirmations. Has anyone's availability changed since they last posted here? At the moment we've got Davie, Gerry and Joe. Kattkieru and I will also be there, which makes five people. Dennis may be unable to attend due to work commitments, and JP said he'd try to make it, which is JP-speak for "I'll be waking up at 5:00pm that day." Five NYNJAs is plenty, but if there are any more cancellations it would probably make more sense to just change the date. So if you're planning to attend, please speak up!
  19. Carl Albrecht-Buehler's "Rigging a Pterosaur" video is on the Siggraph 2004 training series. http://www.hash.com/products/tapes.asp
  20. The hell? Dude, the guy's name is right on the image, and he doesn't even use A:M.
  21. Note that you're assuming that Lee isn't gay, when you say that.
  22. Every day for a year if not more? Seriously, if you think you need 365 shooting days for a feature, it may be time to hire a producer. Apocalypse Now was shot in less time than that.
  23. If the Squetch installation process were going to be overhauled, I'd prefer to see it adopt the TSM method, in which a "base skeleton" is positioned by the user, after which the installer integrates the rig skeleton overtop of the base one. In TSM's case, all the model's CPs are assigned to the base skeleton, which makes it fairly easy to update your rig without having to start the entire rigging process over again. However, since the Squetch rig provides so many fan bones, this method wouldn't be quite as applicable. Still, the ability to change the base skeleton, run the rigger again, and generate an updated rig configuration would be most convenient.
  24. Understood. If too many people can't make it, we'll move the date, so keep checking back...
×
×
  • Create New...