The course is the process of unwrapping a robot model which I did in another Udemy course - Hard Surface Modeling and Sculpting Course in 3D Coat. It covers everything you need to know to get useful geometry and UVs from voxels.
Its a big job and I discover some surprises on the way, like not having enough graphics card memory and how to resolve the issue and how to use scripts to manage what becomes a pretty enormous list of retopo and UV parts.
I thought I knew 3D Coat inside out, but I learned a lot from this process,l including some little known workflows, and you will too.
This is not a course on general 3D retopo best practices but more about how to manage retopo and UV worflows in 3d coat when dealing with many many scene objects.
While most of this project uses autoretopo, we do some manual too and this gives a basic overview of the tools as well as some lesser known ones.
We finish up the overviews of manual retopo. We use it in an applied way in various of the videos coming up as we retopo the robot.
Retopo is a way of simplifying form. Here we look at the advantages and challenges of pre-simplifying geometry before doing auto-retopo. This includes eliminating symmetry and 're-voxelizing'. A look at the controls of the auto topo dialogue and basic curve hints.
Manual re-topology really makes more sense for organic characters; however, here we make some manual retopo on a simple piece of the arm. As you will see, it takes some time to solve a simple part. In the next video we do the same thing automatically and review the time and quality differences.
We finish the manual retopo then show how much easier it would have been done using some curves and auto-retopo.
This piece requires quite a few curves to solve but the result is pretty nice. This one done before the release of 3D Coat version 4.5.1 (still in beta as i write this) which has a new autotopo algorithm implemented which can retopo this whole piece in a couple of minutes with no curves. However, the principles covered are useful for more complex geometry which you are likely to encounter on a regular basis.
Check out the new algorithm video at the end of all these videos to see some of the parts redone using the new autotopo method. Such a joy
We begin manually retopo-ing another reasonably simple part - a shoulder piece. A simple piece yes, but manually retopo is always time consuming.
We finish off the manual retopo with a reasonably satisfactory solution. Now on to the autop-topo to see the efficiancy difference in results and quality and time.
2 or 3 iterations of auto-retopo on the same shoulder piece examining how different curve drawing styles create different automatic solutions
This is a complex piece of geometry which needs quite a bit of cleanup in voxel room before we attempt retopo. The actual retopo needs a bit of manual massaging to make it work.
All the thumb joints are quickly curved up and auto re-topoed. Again, these solutions are great, but with the new algorithm probably no curves at all would be needed
We work on the main instance of the finger. It needs some cleanup using the surface fill holes tool. Curve drawing and auto retopo. Quite similar video tot he previous one on the thumb but it may be useful to see how solutions are found for difficult pieces.
Simply finishing off the finger joints from the previous video
2 different manual retopo solutions for the main cylinder which underlies the finger joints
We export the retopo as it stand now to Maya where we review the smooth preview and see how the surfaces stand up to sub-ds and smooth preview
We do an auto retopo. Half of it is good but the top half is weak, so we combine techniques and manual retopo the top part - so a hybrid approach if you will. i tied the same piece with the new algorithm and it solved in a couple of minutes totally perfectly with no hybrid method and no curves needed - amazing.
Still, its worth a look to see how to resolve more complex pieces
The manual retopo continues. Resolving areas of differing resolution can be a challenge. In some ways its like a jigsaw puzzle trying to find the solution Videos like these might be boring perhaps but it may still be worthwhile watching how to resolve it but again the new algorithm makes these pieces much easier and much more precise.
I seriously can't believe this retopo isnt finished yet. Manual retopo can be a drag but there are a few good mesh optimization tricks like collapsing edge rings, finding hidden extra vertices and that type of thing.
Straight ahead autotopo of a complex form. We trouble shoot some of the curve issues and identify optimal placement of the curves to get decent autotopo.
Rapid autotoping of various parts of the leg...all pretty successful without too many issues!
We change the autotopo algorithm to 'hardsurface'. It is remarkably successful. It can be finicky to know when to use this. it can often be a case of trial and error. To be honest with the new algorthm as of 3D Coat 4.5.1 you can get better retopo without using the hard surface option.
There are some pretty challenging parts here. Occasionally the voxel mesh will need a quick cleanup. Autotopo with some brush clean up as we go.
Mainly auto retopology with some manual cleanup work collapsing edges to get more useful solutions.
We blast through a whole bunch of the torso parts as fast as we can!
We continue the high speed retopo of the torso region
Quickly and succefully retopo the rear part of the head and the pipes. Nothing fancy here - just curve drawing and autotopo.
The last parts of the helmet done nicely except for one big piece which turns out to be quite difficult...see next couple of videos for details
This and the next 4 videos were pretty nightmarish, yet educational. As you will see in the final 'new algorithm video', these can all be done now in a couple of minutes with no guides - totally amazing. For this video I actually had to change the voxels to get a workable solution.
Huge amount of manual editing of the auto retopo to make it useable. I think this might be quite educational to watch.
Finishing the helmet and starting this difficult hybrid auto/manual rib piece of the robot.
This is difficult mainly because there seems to be a scale issue. I dont know if its 3D Coat or my own inabilty to visualize the parts well but it is quite difficult to resolve converging lines from 2 different direction which have an inclination to get smaller. Is it a measure of both? as I feel the camera in 3D Coat never lets me see enough of the model to find a solution quickly, or is this representative of an underlying design problem. Maybe we should never have pieces which have this much variation in shape. Certainly this would never be manufactured and probably wouldn[t 3D print without breaking. Something to think about as you design your own models.
Finally we get a solution despite the tools getting a bit crazy at the small scale. If anything this is a testament to the power of patience -and simplification- winning the day. Never let a polygon get the better of you. Ever!
So annoying, yet comforting that the new algorithm solves this in like 30 seconds with no manual retopo. Unbelievable
I felt so happy finally finishing this off. Its so easy now with the new retopo algorithm, still you will probably encounter difficult pieces just as I have so its good to see maybe.
An introduction to basic UV tools and how to use retopo symmetry to your advantage - you only create the UVs once then duplicate with symmetry in retopo after UV creation. It can be difficult to get perfect edge loops with some of the spiral autotopo, but 3D Coat has the tools to get it done all the same.
We add multiple UV groups per object which is a great way of getting different maps for each piece. Later I have to simplify this for memory considerations but its good to see how to get it done anyway.
The work continues. Clearly there are some information management and speed issues with working with so many retopo groups and uv groups. Several videos from now I discover a little known script which changes the whole workflow and makes life SOOO much easier
We start wrapping up the retopo and find one or 2 errors on the way.
We pack and actually unwrap the UV sets
Wherein I wax eloquent about how happy I am about these new retopo scripts - what a discovery.
I continue adjusting UVs as I go
The UVs are united but the retopo groups ALSO need to be united to make baking out a a more straight forward process. This video shows the process
We show the final process of how to get the retopo into the paint room - finally!
We see how to add subdivisions on our retopo which allows us to paint with displacement and color in the paint room which is another little used but INCREDIBLY powerful feature of 3D Coat. Seriously, its ridiculous what is possible with this.
As i've mentioned in several of the descriptions, 3D Coat version 4.5.11 in Beta as of this writing has an amazing new retopo algorithm. It cleans up with ease all of the parts which were very painful to do in this course. In this video I redo a couple of these pieces and marvel at how much easier this course would have been if it was released already. Still, itsthe difficult parts where you learn how to actually work with retopo IMO.
I have worked in game design, and 3D since 1995. I started my career designing multi-user games using Flash and Director for a big national website then spent 12 years focused on producing animation and VFX for commercials for television. I now teach these topics at university. While I have an M.A. degree from Bournemouth University focused on teaching animation and game skills - really, its all about loving and being fascinated by the process.