sgerrey wanted to know how to get high resolution scans into fusion....

[PhilProcarioJr]

@Anonymous

I worked on your scan data and here is where I am at with it. Just want to make sure this is what you want before I make the tutorial on it.

This was kind of difficult because the control mesh for the T-Splines was 16,000 faces. But it's in and converted to a solid and editable.

I just want to add, contrary to popular belief Momento and Meshmixer will not work for things like this if you want to achieve CAD accuracy. I will explain why in my tutorial.

[PhilProcarioJr]

[PhilProcarioJr]

[TrippyLighting]

I'd be interested to know how long it took on your machine for that 16k mesh to convert into a T-Spline.

I've a prohject that includes a hair over 19k faces and the conversion from a mesh into a T-Spline took 5 minutes. Afterwards you can work with the solid but editing the T-Spline is out of question, at least on my almost 6 year old i7 imac.

And, yes, again, I can only support what @PhilProcarioJr has said. If you want CAD like accuracy ther is no free lunch here. That requires retopologizing the raw scanned mesh and that also requires a certain skill. 

[PhilProcarioJr]

I just went back and did the conversion again to check the time...it took 11 seconds to convert the 16000 polygon T-Spline to a solid.

In all my tests if you are trying to real-time edit anything over 5 to 7000 faces it's not going to happen on any computer. These high polygon meshes are for T-Spline to solid conversion only.

But let's be honest, how many other CAD apps can even do it?

[TrippyLighting]

Whaaaaaaaat ? 11 seconds ?

Now I am wondering if that has to to with the tubular grid like structure of the mesh I imported and Fusion 360 and whetehr the complexity of the mesh plays a role.

Just to be sure, I am talking about the conversion from the imported mesh body into the T-Spline.

Totally agree on your last statement!

[Anonymous]

@PhilProcarioJr

WOW! 

I think I may have misled you a bit on the title of the file. So that results in the geometry you see. 

Let me clear this up.

mirroed 202. Stl meant that the scan had been mirroed. It was done so that the left handed stock that was scanned could just be mirroed to produce a right handed stock. This saved me a lot of money and the results were perfect for what I needed. That was until taking up with fusion. Then my scan files became redundant until your work flow. 

The detail you have got into that. I am speechless, amazing! 

You do relise I wish to become a disiple of your work flow. Regardless of what come from the mesh tools your work flow has some major implications to my future projects. 

Did you get the other files I linked You into? 

The camo stock is much more of the kind of stock I make. 

The 202 is a bit of a curve ball as the stock is in two parts. 

The butt stock (which is ether left or right handed) and a forend which is generic. 

I also apploigise for mentioning that I would like to join the butt and forend together. It wasn't very clear on what I have done. There is a stl of what I have tried some where on a old laptop. I will dig it out and sent it to you. But the camo stock would give you a better idea of what I intended. 

From here on in. Please consider me your padwan. Awesome work that. 

Autodesk should have you on retainer! 

[PhilProcarioJr]

@TrippyLighting

Oh Mesh to T-Spline....sorry that took 16 seconds. 

Something else I am not sure if you have done but when you create your mesh make it in pieces then convert them to T-Splines.

Make a copy of the T-Splines then stitch them together and keep the copies for realtime editing. You may spend a little extra time stitching the pieces together but you will be able to edit the T-Spline mesh. Actually here is a little secret ALWAYS break your control meshes into pieces....How many? Depends on the crazy levels of detail you want. There is one tool in the T-Spline toolbar that doesn't much seem to care how many faces you have while editing.....the merge edge tool.

Cheer

Phil

[PhilProcarioJr]

@Anonymous

The mis-communication is fine, no worries. 

Thanks for all the kind words, I'm just trying to break a few molds and get people thinking outside the box...this old age CAD attitude that floats around is depressing!

Out with the old ways and in with the new!

The point here is what you want to do can be done but not with Fusion alone, and not with Momento or Meshmixer.....Those apps definitely have their uses though.

Actually what I have done here is really no big deal, just a little bit of work. What I will show later.....now that will be a challenge.

Remember I am not doing your work for you I'm just going to show you how to do it yourself.

Trust me you will have a lot more pride looking at what you did. Give me a little time to get things rolling on the camo...

Also what did you use for these scans? 

I have to say these are some of the best scans I have seen.

Anyway Cheers guys more to come....

Phil

[Anonymous]

@PhilProcarioJr

I totaly agree. 

Lets not get stuck In the past. 

I also don't want the work done for me. I want to learn how to do it myself. Not just for the fact it can save me money. But for the sure benifit of improving myself. 

So please teach me. I will be a eager student. 

The scan files were created by a Faro v6 and the arm to go with it. 

This tech is state of the art. There are many different avenues into scanning and you get what you pay for. 3D scanning is some thing I want to get into. 

First things first though. I need to get the cnc to pay for its self. 

With my old cam that would have taken 30 years. Fusion cam cut my run time down by two thirds. But as you know it won't cam from a mesh. So that leaves me needing to rebuild my expencive scan files. I have paid for this work before. But even then the the results can be mixed. 

As as far as cad skills go. I am pretty much a blank sheet. So I am very willing to learn a work flow that suits me. 

I do think autodesk are looking to bridge the gap though. During my email with Colin and Jake I get the feeling that getting the mesh workspace right is of real concern for them. 

I have a call booked with Jake next Wednesday before I get on a plane to visit the USA. 

He has seen the same camo file you have. His first comment was the same. How did you get that file? 

Lol. 

Look forward to talking with you soon. Are you on Skype? 

Cheers 

Simon 

[PhilProcarioJr]

@Anonymous

No worries I will show you, it will be up to you to refine your skills after that.

As far as the mesh tools go I see one huge problem, all the information I have seen from the people involved with them are focusing on ways to do the current CAD stuff.....

IMHO that's all bad, why, because in order to be innovative you have to break away from whats currently known. (The only thing I have seen that's inspiring is Fusions ability to handle much denser meshes. That I will admit is huge. Kudos guys for that one.)

For instance Meshmixer and Momento are NOT good for CAD accurate modeling.

Some of the videos out there on the mesh tools focus on things in those two pieces of software being in Fusion 360. And for some things that's fine, but for CAD accurate models not so much.

Everyone is missing what the really needed tools should be, and all of them deal with mesh flow. For the life of me I can't see why the world doesn't see this. I have only talked to maybe 3 or 4 people that do see it.

Now don't get me wrong I know they have to start somewhere, but if they develop based on what is already out there and being done they will back themselves into the exact same corner as all other CAD software is in.

That means somewhere down the line all the logic or framework or both will have to be rewritten or massively overhauled.

I have seen it so many times over the last 20 some years. The problem here is people don't know what tools are needed for this kind of stuff because each person is good at their one thing.

So this guy knows this and that guy knows that...so on....the problem here is simple you need to know this, and that, and that, and that to develop the new tools.

I have the upmost respect for the Autodesk employees but the company suffers from a lot of the same problems as others.

This team works on this and that team works on that and they are not consistently working together.

For instance the team that handles the render works on the rendering for all Autodesk products, but then Fusion has it own people for deciding what goes into Fusion renderer and how it is used and what options are available to the end user. So even if the renderer supports X function the group behind the Fusion UI for it may not support it at all or they may dumb it down because they feel it is too complicated for the average user and so on.

Now keep in mind I'm not bashing the Fusion team, so far all my dealings with them have been fantastic and they are a very talented group of people. I just feel they are in a raft in uncharted waters with a bunch of hungry sharks....don't stay on the raft with a leak when there is a cruise ship 50 feet away. 

Just my 2 cents

[PhilProcarioJr]

In theory retopoligization works great if certain conditions are met. Now what I am about to describe is a VERY simplified form of this theory.

In order for T-Splines to interpret the correct surface flow there needs to be as many control points located to the real curvature of the surface as possible.

The farther away the control points are from the actual surface the less accurate the final surface will be.

So how do we know if we have enough control points?

In simple terms you need at least two actual surface points within the tolerance you need to hold.

So if you need to hold a 0.1mm tolerance your scan mesh needs to have actual surface data within that tolerance.

This is problem 1- High resolution mesh needs to hold enough polygons to accurately represent the surface.

Now your T-Spline needs control points that reflect the curvature of the surface, so the sharper the bend the more control points are needed closer to that bend.

The broader the bend the less are needed, this is why mesh flow is critical. Anyone with a Sub-D modeling background understands this, well they understand to get the look they want they have to apply this concept.

If you have a fillet that's 1mm on your high resolution model you can't have control points 10mm away from the fillet.

This is problem 2 - Keeping control points close enough to the actual model accuracy to define the surface correctly.

Programs like Meshmixer and Momento apply a quad grid over everything and it's unified. Looks like someone wrapped them in graph paper.......

Why is this bad?

Because looking at our theory we see that in order to accurately mimic a surface we only need the dense polygon count around the sharp edges and not the smooth flowing long curves. 

With both of those apps so many polygons are in the wrong place and it all amounts to wasted computation. 

In order to accurately represent the surface (especially around fillets and sharp edges) you would need billions of polygons for some models.

Try importing a mesh with a billion polygons once and see what happens.

In it's simplest form you only add control points where they are needed.

This is problem 3 - Mesh Flow

There is a lot more to this but this is a simple breakdown......

[Anonymous]

@PhilProcarioJr

So what your saying is the more detailed the mesh in areas of tight corners. You will need to add more control pionts. 

Or I supose in a t spline view you would add more subdivisions to get a better curve. 

WhenI look at the model you made using the scan of a 202 butt. I was amazed to see such a uniform grid pattern. Yet when it is re rendered it looks flawless. 

How did you get the curve where the butt meets the action block to have such definition? 

[PhilProcarioJr]

@Anonymous

I'll show you....

So I got a min to start on the camo stock...I really wish it didn't have carvings in it....anyways

[Anonymous]

Yes those patches of checkering could do with ironing flat. But that what happens with such fine detail scans. 

[PhilProcarioJr]

@Anonymous

I could capture the detail...that's not the problem I wanted a blank slate so I could add my own. 

I can probably get rid of them...we'll see.

[PhilProcarioJr]

[daniel_lyall]

@PhilProcarioJr remove the checkering @Anonymous should be able to do them by hand it's a old but still done gun stock making thing everyone likes it a little bit different, @Anonymous yer more work for you but have it as a option some people don't like it. 

I don't but the big guns I have used are match rifles so you wear a glove 

[Anonymous]

@daniel_lyall

thats spot on. I keep the stocks that fail for practice on. Checkering is a art. But I saw a robot at the Mach machine show last Thursday that could cut checkering so fine that you would never want a human to do it again. 48 lines per inch with no boundry cuts. I have been studying checkering since starting this project 4 years ago. Boundry less cuts are still evading me. 

The way the robot did it you could hardly see the pattern but boy could you feel it. Amazing stuff. Some would say to good. 

I prefer stiplelng as I find it easer to do. But some people think it's cheating and the diamond are the only thing worth considering. 

But then they shrink away when you tell them how much it will cost! 

I will ill send you a PM today mate. Sorry I haven't got back yet. I have been on nights, but now back in the land of the living. I'm off to johns open house next week. That should be a lot of fun! 

[Anonymous]

@PhilProcarioJr

I am beginning to see what you mean. Control piont to the places they are needed. 

I can't wait to have a crack at this.,I can see a way forward for this now. 

I also except that this is not a five minuet job. 

But what a skill set o have. 

Thank you for taking the time to teach me.. I promise to invest the time to practise. I won't let you down.