Rodney

I note that there is an interesting phenomenon going on that some have actually struggled against but that can be used to nice effect for glow/transparency. Note that in the second image above how the glow is greater in the area where no ground plane is. This suggests that if we want to increase the glow in certain parts of our model while decreasing it in other areas we simply need to place (or remove) some geometry directly behind the model. As this could be useful later on... note to self... note to self... note to self.

Failed attempt with an ultra simple model... Things that I thought worked: - Using a checker material to try to get the color variation. - Modeling only one half of the brain and then duplicating the other half in the Chor (scale the second instance of the half brain -100% to get the other side of the brain) - Flipping the material in the second model to achieve the reverse effect of color Things that don't work so well: - Doesn't capture electric feel - Surface Roughness doesn't achieve brain folds effectively. - Need more transparency throughout* - No glow* *Note that in these last two cases they can be achieved by adjusting the surface properties. Added a render with glow and transparency tweaked. Note that for the material I just started adding random nodes/colors. I didn't make any attempt to be accurate. As such some nodes could be deleted with minimal effect over all. ColorfulBrain_Failed_Attempt.prj

Looking good Paul! The primary area that seems to stand out as 'off model' is the origin of the shoulders/arms which in your model is really low and in most of the imagery/references appears to begin just below the head. Some minor simplification of splines might help in making him more squashable and stretchable which is something I'm confident you'll be wanting. The other option might be to lower the head down to the shoulders. In drawing the rotoscope I had to make a compromise in the angle of the arms going out to the side in that they curved downward more than they remained horizontal (as in your model). I'm tempted to suggest that wherever you have a shape you might strive a bit more for a curved line. By this I mean to say that your model appears to have real bones and anatomy whereas Humphrey's construction only hints at that underlying structure in a minimal style. The example of this might be seen in your model's knees which in most reference aren't nobby knees but two parts connected together in a classic Disney/Warner Bros line. Can you post a render or two from the front and side views? That will help with the analysis. He's definitely a bear right now... and a good looking one... now he just needs to be pushed a little bit more toward the likeness of Humphrey. I'd say the eyes (mischievious), shoulders (hunched) and knees (streamlined) are the primary character identification areas to focus on. Added: I've attached a crude drawing that attempts to raise the shoulders and simplify the mid section. By raising the shoulders/arms you could then remove some of the midsection to achieve easier squash/stretch in animation. I would also make the toes more pointed. In some instances it's almost like he couldn't bare his own weight whereas with your model's wider toes he'd have no problem there. You might be able to achieve that tapering effect simply by scaling the toes down smaller toward the ends. There should be some of that tapering/sharpness in the fingers as well. There aren't too many places on Humphrey that are sharp so you'll want to place some emphasis there. It's a bit hard to see detail in the face of your model and that is a clue in and of itself that suggests you might want to further exaggerate the whites of the eyes and add those largishly dark and expressive eyebrows.

These renderings just keep getting better and better.

Thanks for posting that Robert. I've been itching to purchase a new computer (a desktop) for quite awhile now. If I could find the right setup I might spend some money.

Sure is. He's a busy man these days so go easy on him. He goes by Yoda64 (presumably to distinquish himself from the other 63 yodas that are considerably less active than he is). http://www.hash.com/forums/index.php?showuser=218 Added: What you've got going sounds like a very cool project.

Keep in mind that the current trend is to move away from UV mapping entirely (at least in the standard sense of that term). It is interesting to note that while the world in general is moving away from classic UV mapping, some A:M Users are trying to move toward them. It's a fairly awkward situation and my gut feel is that a more perfect solution/workflow should be explored that is more spline/patch-centric. For where Disney/Pixar is trying to push the industry reference: http://ptex.us/overview.html

Tore, It's great to see you here again. I expect you want to talk about the items on your list (and that is a good idea because some things might already exist or have effective work arounds/approaches already but ultimately keep in mind that in order to make any wish list a reality you'll want to add those line items into A:M Reports as feature requests. You never know when one of your requests may sync perfectly with something Steffen already has planned to bring to reality. Keep in mind that when posting 'wish lists' to the forum you are not requesting new features so much as saying, " I really want to talk about/explore my thoughts on this". The reason there are not 'wish list' topics is that wish lists are to be filed in A:M Reports so that they can be better understood, considered and prioritized. Those reports help define the future of A:M.

That was fun. It's good to see an animation contest running in the forum again.

Stefan, If you can get in contact with Steffen Gross, he's got the current (or most current) Software Development Kit (SDK). There should be some good info on how to interface with Actions files there.

So very goooood! Love it. Good grief man. If you can do that on your first attempt there is only one thing to do. Model more!

I'm arriving late to the party as well. Hope it was a happy one Lloyd and Jason!

You are on a roll. Very nice!

Paul would have been 61 today. It seems appropriate to reflect on the man, his talent and generosity even as we are reminded of how much we miss him here in the A:M Community.

Thanks Paul, I only see the flaws (which are many) but hopefully that will give you a starting point. One of the reasons I wanted to do an animated turnaround is that it reduces those flaws considerably. For instance, in between the front and side views of the head we have something of a problem with the smiling face on the front and the frowning face from the side. The comic book simplification is a quick overtrace of a Humphrey face I saw posted in a comic book. The corners of the mouth really irritate me and even now I want to go back in and resolve that by having him smile (or frown... or both) from all views. There are a few 3D recreations of Humphrey already out there in the wild and those would be great to get up close and personal with. One is a figurine based on one of the old animated short films. One is a teddy bear. And the other is the bottom character on a life sized totem pole in Disney's Wilderness Lodge. As far as reference goes, I haven't looked through all the Humphrey shorts themselves and I'm sure there are some very nice views to choose from in those. Example: kcaQ4P7qn38 Hooked Bear Q7JhGMXvDP4 Donald's Fish Fry Perhap's the classic Humphry the Bear short (with a ton of bears duplicated all over the place) is in the 1956 short 'In the Bag': 4ol28kPTqa4 Of course there are many others to include 'Bongo' which is a great short to study because of it's higher quality. Interestingly, one of the directors went on to create other, almost carbon copy versions of Humphrey for other studios. Those can provide decent reference as well. And there is Yogi bear too which is something of the ultimate oversimplification of Humphrey. In yet other news there is Humphrey B. Bear who only looks like Disney's Humphrey on occasion.

What I find particularly nice about this workflow is that these two files, the image/decal (.png) and the model/spline roto (.mdl), are generated from exactly the same file; an adobe illustrator (.ai) file: AI_Import.mdl

Don't forget that there are several ways to slow down video via youtube. One is to view the video in HTML5 enabled view. Of course then you'd lose Marcos's relaxing music... (I suppose you could play the faster version in the background though) You can watch it even faster too! (Watching someone model fast is a good way to get rapidity going in your own methods)

Here's a spline roto that came from the raster to vector test I did with Xara Xtreme (which over here the Tinkering Gnome noted was free until 20 May): Why use a standard roto when you can start from a spline roto instead? Note: The results here are using default settings across the board. Simplifying the results in Xara would reduce the splines coming into A:M considerably. Edit: I removed the model for now because it is entirely too spline heavy. I'll adjust the settings and upload it again later today. Added: Uploaded a simplified roto:

Here's a first draft... I started messing with it and messing it up so I figured I should post it before going any further. I haven't given a lot of thought to where spline might be placed and that will be an important consideration. I did an initial test to see how the lines might come in as splines via the AI Wizard but I should have used the black and white image first... the plugin choked on the detail. If interested I can run that through again as that would give you an initial placement of splines that you wouldn't have to lay down. This was more digital manipulation than drawing but I tried to get some variation in the various views. Some decisions will need to be made on whether to keep fingers and toes pointy or rounded and you might want to borrow Keekat's chest hair to get some bear hair going. I noted that dependng on the artist/animator the styles for Humphry can vary a lot and I even found some comic books. If interested I can smash all the various reference images together and post those. Added: The reason I have Humphrey jumping is that for a moment I considered doing a turnaround that would roto him in a variety of poses. I think maybe we'll save that for the 3D version. Also: I note that Humphrey's nose is often cheated from front to side view. In some images his nose appears slightly heart shaped (on the top). If we need to perfectly align those I'll inbetween them and find a happy medium. As they are right now they don't align in the front/side views. In the side view I would say his head is tilted back a considerably. Tilting it forward should get it into closer alignment. Back to you! Thanks for the vote of confidence Robert!

Unless someone really wants to take this on I'd like to take this challenge on. It may be possible that no drawing is necessary (if I can find the right reference) but I'm in the drawing mood so... ...just say the word and I'm in.

Okay, who let Stian take over David's brain? Impressive. Very impressive. And it'll be even more impressive in motion!

Happy Birthday Teresa! Live long and animate. (and as they say in Oz... Woot! Woot!)

Nicely done Dan! Regarding the return of Epic Gaming.... what can I say except... Yay!!!

Great looking building! Don't forget the use of proxy models in the short term as well. You can substitute a low count model (proxy) while working on the scene and then substitute the hero model back in for final rendering. The proxy model can even be a cube or simple shape. There are several ways to swap out proxy models. One (probably most used) is to change the shortcut in A:M. Just point it at the desired model. Another (my preferred method) takes some consideration: Create three folders to facilitate your 'proxy pipeline'. Folder 1: Working Folder (This is the folder you will point A:M at to find the models Folder 2: Proxy Models with little detail Folder 3: Hero Models with full detail The benefit to this setup are many but the primary one is that you don't have to change any shortcuts in A:M. A:M will always point to the working model. Because the location of the proxy and hero models never changes there is little risk of corrupting those models. Note that a script is usually ideal for copying files to the Working Folder. When you run the script it simply copies either the proxy models or hero models into the working folder. So... if you are working with only proxies then you'd launch the script/utility to overwrite the Working folder with proxies. If I was a programmer I would plus this idea of proxies up a little more to take advantage of what A:M can already do for us with 'psuedo proxies'. This psudo proxy system (as I tend to call it) is yet another method we can use in A:M that doesn't require us to swap out models. We simply use Object Draw and Object Render modes in the PWS to adjust our model's proxy status which in turn will speed up response from the interface considerably.

The way to tackle that is just to insert a couple frames where everything except the face freezes in place. You've already established that a little roadrunner-style pausing is plausible so that shouldn't be an issue. Here's where A:M' nonlinear rendering capability can be put to good use. Let's say your sequence has 250 frames... (I'm using a random number because I can't check the actual number at this time) Let's say the guy reaches the bottom at frame 150 what you can do is save that Chor/Project as a second Chor/Project and then go back and adjust frame 150 to manipulate the guys face. It helps to think in terms of three frame sequences here. Frame 15a might be eyes closed. So render that out with a unique name (like... imagesequence0150a 000.tga) Render as many frames as you feel are necessary to convey the change that will occur on his face. (I'd say this should probably be no less than 24 frames (1 second) but not more than 72 (3 seconds) *adjust for 30 frames per second as necessary Render out the next look that crosses the guys face as imagesequence0150b 000.tga. Again... 24 to 72 frames. Render out the final look of the guys face as imagesequence0150c 000. Now, in your operating system (or video editor) you can go back in and rename/resequence those frames as you see fit. Splicing this inner sequence into the main sequence can be done the same. If brute forcing the edit you'd just have the following: Frames 1 through 150. New sequence Renumbered final sequence beginning where the inserted sequence left off. Of course you can do this all in A:M without all the juggling but the preceding is what is happening behind the scenes either way. You know how many frames you've budgeted for this scene and if all of the images are already rendered you can renumber and reposiition them as needed. When in doubt you can simply go in and duplicate the frames you need to see longer (again suggesting frame 150 so... name the dupes something that will keep them in computer sortable sequence. Then simply replace those frames if needed. Bottom line: It takes approximately two seconds for a subtle change of expression on the face to fully read so it would be nice to have a slightly longer delay at the very bottom of the guys decent on the tree. If there isn't going to be much of a change of expression then you could likely get away with just inserting a few duplicate frames to pad the held/frozen frame.