Leaderboard

Simon Walker noted on Twitter that the Maxon Training Team guys likely have some solid training in the works for this stuff. It wouldn't surprise me if we get one of their 4 or 5 week series on this. Say what you want about subs but when things are going smoothly, getting the software update with a polished top new feature and accessible bulk training with it all on your desktop with no mucking around is an attractive combination. I'm looking forward to when this comes out.

Makes me feel so old.

Has anyone seen Sandman ? I haven't seen any VFX breakdowns of their 1st episode intro but it immediately reminded me of those AI animations. I have the feeling it could still be made with traditional AE but it was so well made and hadn't seen anything like this before. I'm talking about the first few transitions before the craw enters the dream world. It could be some kind of datamosh effect or good-old morphing...

c4d lists the width. The height is determined by your render setting aspect ratio. so enter 27.99 into the sensor size setting of the camera and then use a suitable render resolution with a 27.99:19.22 ratio. eg 3840x2637

AI is a whole new beast and hard to add to 3D DCC companies for the following reasons: Companies like to implement well-founded technologies, or they will have to invest in R&D which is a risk and costs time and money. Deep Dreaming has been around for a long time and gave the foundations (data bases) for text-to-image. The evolution to 3D needs more data and tricks and this is just the beginning. AI is based on training data. This means that the algorithm needs access to a huge dedicated data base on a specific problem in order to solve it. In the case of UV mapping that would require a big database of well-made 3D models and their respective UV maps. The data also require some normalization and classification this is a big headache for data analysts due to the quantity of the data. This is done during the build of the data base. Classification means that the data need to be described in terms of tags and normalization means they have to be conformed to certain restrictions like the resolution size of the UV maps, number of polygons or even color space. It's a whole new and different field of expertise for companies to look out for hiring. Not just programmers but also technicians. Training an AI requires some big guns in the hardware arsenal and the gun provider is only NVidia. The technician's job would be to setup a GPU cluster or farm with all appropriate necessities that requires (cooling regulations, security monitoring etc). The AI needs to be trained. This is also quite a task because the AI still needs some human guidance to make sure it spits out correct results.

Works on a single axis so any cylindrical object will work but you have to adjust the radius. For rolling a ball in all directions, that would require more elaborate steup.

when you see the steep developement in 2D, then this is just the start and we will see much better results in the near future. I would not be surprised if we get a shop that generates the assets for you. that of course in the future will cost quite some modelers the job, but thats how it always was. The end of letter setting comes to mind. If we really have to go into this direction I don't understand why comanies like autodesk or maxon do not integrate AI algorythms into their software. for example to wheight characters or make great UV maps. It looks like these algorythms always start with the final image but they also could just be great tools for unpleasant parts of the 3d work. for me it looks like photographers will be the first to have a hard time in the near future. Not because there will be no jobs for them anymore, but because there will be much less jobs. Speakers also will get problems quite fast. for 3d, for animation we might have some years left, because making a animation needs a deep understanding of the background workings, But modeling will deffinately be challanged in the near future.

Btw, I am sure Maxon wouldn’t announce something too early. Wonder when this is coming out?

That McGavran guy just seems to dance to his own tunes!

Many of the previously missing icons are now present, but some are still missing. In my layout, the following are missing: Wireframe, Isoparms, (those two bother me the most) Skeleton, Display Filter, Save as Startup Layout, Close Project, Close all Projects, Go to First Marker, Go to Previous Marker, Go to Next Marker, Go to Last Marker (To be fair I never use those marker buttons anyway and should remove them from my layout) Use as Render View, Go to First Key, Go to Last Key, Lock Camera Pivot, Calculate FPS, Frame Project Settings, Frame Preview Range, Go to Start, Got to End, Frame Range, 50%, 200%, Show Text, Show Line, Pixel Aspect, Vertical Tags, Object Information, Convert Object Selection to XRefs, Unfold Selection, Fold Selection. Wow I wasn't expecting this list to be so long, but to be fair again many of those last ones I listed I use text menus only and I just turned on the icons to see if they were there or not.

Since you both are the only ones showing any interest I wonder if you would want to test the updated plugin, before I provide a public release. I currently have an updated version of the plugin which runs in R23, and another for R25. Both Windows-only for now. Let me know if interested to test? I will then provide the beta plugin via private message for you to test. Feedback of the testing can be provided in this thread.