Full Version: V4 beta Feb-27-2019 available now

Show messages:  1-2  …  83-102  103-122  123-142  143-162  163-182  183-202  203-218

Hi Michel, thanks for the v4 feedback!

re:
> PDF export crash report
> Mac Beta version crashes on my side when exporting a rectangle to pdf.
> I tried on OSX 10.9.5 and on another machine on 10.12.6 and both crash.

Was a crash report file generated? That would be under ~/Library/Logs/DiagnosticReports/

If one was generated could you please send it to me at moi@moi3d.com or post here?

One thing that you might check though is your setting for "Canvas megapixel resolution" in the PDF export options dialog when generating a hidden-line drawing. If that is set to a particularly large value but you do not have a corresponding very large amount of system memory then that could result in a crash currently. Since MoI v4 is 64-bit now the limit of how much memory it will try to consume has also been increased, that could be why you see a difference in behavior here between v3 and v4.

It takes up about 34 MB per megapixel and putting it at something like 1000 will consume about 34 GB.



re:
> 3D view rotation speed observations
> Also, I find that in 3D view space rotation and zoom in/out seem slower in v4 ("bumpy"?) when the grid is visible.
> Much smoother when grid is masked or when using Parallel (isometric) view.
> However, in Parallel view, display artefacts are more apparent.

What is your setting under Options > Grid > "Grid sections" ? The default is 150 but if you have changed that to put in a high value there that might cause these kinds of problems.


- Michael

---

Hi Michael,

re:
> 3D view rotation speed observations
>What is your setting under Options > Grid > "Grid sections" ? The default is 150 but if you have changed that to put in a high value there that might cause these kinds of problems.

My setting in Options > Grid > "Grid sections" was way too high !
I turned it down to 1000 and now it's working fluidly.
Thanks !

re:
> PDF export crash report
I just sent you a crash dump.
This one still a mystery to me.
The MacBookPro has 16gb of memory and the other machine I tried was an 8gb iMac.
When exporting lines (no solids or surfaces), the "Canvas megapixel resolution" option is not there.
However, when I select export a Solid, the "Canvas megapixel resolution" option is set to "30.0".

Michel

---

Hi Michel, thanks for sending the crash dump file. It appears to be a bug in converting a style name into a layer name for the PDF file.

Can you please send a .3dm file that will trigger the crash? That will help me to track down the bug.

- Michael

---

Everything (basic functionality that I use) seems to be working great. New version install process is so smooth with it picking up CustomUI automatically, sweet!

---

Hi Tover, thanks for testing it!

- Michael

---

I like this theme, but I now have two issues with me install. Once I installed this theme I noticed my buttons on the command bar are round and it created two rows. I will prefer ever thing to be one row. How do I fix this?

Image Attachments:
Annotation 2019-07-04 133404.png 

---

Hi Ron, I answered you on this other thread here:
http://moi3d.com/forum/index.php?webtag=MOI&msg=9242.25

- Michael

---

A very minor observation:
Circle TanCopy.js and Flow.js both use the "shorthand" "var ui = moi.ui;"
Circle TanCopy.js continues to use the "shorthand."
Flow.js uses the "shorthand" once, and then reverts to moi.ui.........

- Brian

---

@Michael- Is there any way that we can adjust blend parameters after it has been created? Currently the blend updates if the blend references surfaces move or update. It would be cool if you could double click on a blend and adjust blend parameters- ie the initial surface creation parameters appear again so that you can adjust the bulge. It would be cool to do this on other features like revolve- double click on a surface to re-tweak parameters- cap or not, revolve angle etc:

---

Hi parel, sorry no there is not currently any way to adjust those kinds of parameters later on. It is something that I want to explore in the future though.

- Michael

---

Love the interface as is- look forward to where you take it.

---

Hi Michael,


I was just wondering, will there be any other improvemenst or new features for V4 beside the measurements?

Don't get me wrong. Measurements is a big addition to MoI by itself. Was just wondering.



Thank you!

---

Hi Stargazer, there will be an assortment of various bug fixes and a few smaller things but it's primarily annotations.

The back end for the annotations is pretty much done now, the main things left are UI for controlling them and seeing if I can get some export functionality.

Then it's time to wrap up v4 and move on to v5, I would like to work on a wider variety of new things for v5.

- Michael

---

Cant wait to see the final V4... :-)

---

Hello Michael,

I'm sorry if this has been both asked and answered already. If it has, maybe a shorter thread with just FAQs about the new version can be pinned.

I have a few questions I'd like to ask and issues with MOI 3.0 I'd like to report so that they're fixed in 4.0 :

1. MOI 3.0 sometimes has trouble properly importing some of my more complex .sat or .step models, even though the models loaded and displayed correctly in other software such as eDrawings 2019. Once imported, the CAD models would be missing surfaces or would have some surfaces distorted or snapped / joined to edges they shouldn't be joined to. And could also have disjointed edges as well. Can I provide samples to help ensure these issues are corrected in MOI 4.0? Here are a few examples :

a) the propeller.stp model you can download from here : https://grabcad.com/library/swept-blade-propeller-1

b) the cylindrical outer shell of the mould.quadlobe.toy.stp model you can download from here : https://grabcad.com/library/mould-for-the-plastic-toy-02-model-1 . It has a small orifice at the bottom which is supposed to allow fluid into the mould but when you open this in MOI 3.0 you get a flat square surface which plugs the inflow orifice.

If I export the model as a .sat file instead, it's much worse. The entire outer surface of the outer shell is missing.

I do all my modelling in a different CAD software and mainly only use MOI to export to .obj so reliable, correct export of CAD models really what I want most from MOI, above anything else. The only other operations I use MOI for are twist and flow.

2. Can I install MOI 4.0 alongside and separate from MOI 3.0 and not have it interfere with MOI 3.0 in any way, including not overwriting or changing MOI 3.0's configuration file? Otherwise I'd rather not switch to MOI 4.0 until the release version. I have the thread limit set to 1 and the undo levels to 0 to save memory so as to be able to import and export more complex models and meshes.

3. When exporting to .obj, without choosing to weld vertices, MOI 3.0 sometimes introduces unnecessary vertex duplication / seams.

Like with the flat bottom plate of this CAD model when you export it to .obj :

https://grabcad.com/library/impeller-design-12-1



or the flat bottom plate as well as center top curved surface of this CAD model, likewise when you export it to .obj :

https://grabcad.com/library/impeller-design-05-1





Normally, I export my models with vertex welding disabled and I also usually make use of both the 'avoid larger than' and 'avoid smaller than' options. Not sure if those constraints are forcing this behaviour to occur when it otherwise might not.

4. Is MOI 4.0 64 bit and can make use of more than 4 GB of memory?

5. What will be the upgrade price to MOI 4.0 for MOI 3.0 customers?

6. Any ETA on the release version of MOI 4.0 and the release version of MOI 5.0 ?

Attachments:
problem.models.7z

Image Attachments:
redundant.edge.png  redundant.edges.2.png  redundant.edges.png 

---

.stl files aren't like .obj files because they cannot contain information about normals, don't use triangle indexing and cannot contain information about UVs.

You can use Meshlab to convert .stl files to .obj. On import of the .stl mesh file into Meshlab you'll be asked whether you wish to merge duplicate vertices. You'd want to choose yes because .stl doesn't use vertex indexing and each triangle is basically regarded as separate and on its own. And every vertex is consequently duplicated for each triangle that has a point at that vertex. Then you would go to Filters => 'Normals, Curvatures and Orientation' => 'Re-Compute Vertex Normals'. Then you can export as an .obj from File => 'Export Mesh As'.

You can also just use Meshmixer to open the .stl file and export as .obj. It should automatically attempt to generate averaged vertex normals, from my experience.

---

You can import your .stl file into Meshlab and you can use Meshlab as follows :

1. File => Import Mesh => .stl file.
2. Filters => Remeshing, Simplification and Reconstruction => Subdivision Surfaces: Catmull-Clark - 1 iteration. This will convert all triangles in the mesh to quads.
3. Optionally, Filters => Normals, Curvatures and Orientation => Cut mesh along crease edges
4. Optionally, Filters => Normals, Curvatures and Orientation => Cut mesh along crease edges
5. Optionally, Filters => Normals, Curvatures and Orientation => Re-Compute Vertex Normals
6. File => Export Mesh => .obj file.

Alternatively, you can use Meshmixer to quickly and easily convert your .stl file to .obj and only afterwards import it into Meshlab to apply Catmull-Clark subdivision and the other steps. It will even generate averaged per-vertex normals automatically. When you apply Catmull-Clark in Meshlab you may afterwards need to rebuild the normals anyway.

You can also break a mesh into shading groups in Meshmixer as well, like in Meshlab - by specifying a threshold dihedral angle.

---

Hi Larry, yes MoI v4 is now a 64-bit program and so can use all your system memory instead of running into 32-bit limits. So it can handle much larger files than v3.

Also there have been some other improvements such as using multiple CPU cores for processing objects so usually it's a lot faster loading files too.

There is a collection of all the V4 beta release notes at http://moi3d.com/wiki/V4Beta.

However, CAD file translation is a very complex area, the thing that tends to cause the most problems is that different CAD systems represent trim boundaries in somewhat different ways particularly on closed surfaces and so that means there can be some processing needed on trim boundaries at import time. It's a somewhat fragile area and when you see the type of distortions you mention it means something didn't go quite right with that. This can usually be corrected by doing some repair operations like untrimming and retrimming the messed up surface. There is a tutorial on these techniques at http://moi3d.com/forum/index.php?webtag=MOI&msg=446.17 .

Your propeller.stp model has the same issue in v4 as in v3. Unfortunately because it is a delicate area and we are at the very end of the v4 beta time period, it is not an area that I feel comfortable with making changes to because of the risk of introducing other bugs. It is an area that I would like to tune up in the future but it will have to happen earlier in a release cycle rather than at the end.

What happens with propeller.stp if you use IGES format instead of STEP/SAT ? If your software has an option for using IGES type 144 surfaces that could potentially have a trim boundary closer to what MoI needs and so might have less processing needed on it.

re: mould.quadlobe.toy.stp - the same plug problem is happening with v4 here as well. V4 does now import all the colors. If you delete that square plug you can then show naked edges (http://moi3d.com/forum/index.php?webtag=MOI&msg=6051.2) to see where it was supposed to be. There is a very skinny ring there, only 0.0017 units wide. That can generally cause problems having such a small width. But if you select the object you can then use Construct > Planar to fill it in and then it all seems to be ok after that.

re:
> 2. Can I install MOI 4.0 alongside and separate from MOI 3.0 and not have it interfere with MOI 3.0 in any way,
> including not overwriting or changing MOI 3.0's configuration file?

By default they will install to separate directories but they will use the same moi.ini file. But there isn't any problem with them using the same moi.ini file, v4 will add some more entries to it for new things in v4 but those are just ignored by v3. It is possible to make v4 use a separate moi.ini file, you can do that by either putting a moi.ini file in the same folder as moi.exe, then it will use that one. You'll need to make v4 run with administrator privileges for it to be able to write to the file though. Or another thing is if you pass a path to a moi.ini file (put " " around it if there are any spaces in the path), it will then use that ini file.

re:
> Otherwise I'd rather not switch to MOI 4.0 until the release version. I have the thread limit set to 1 and
> the undo levels to 0 to save memory so as to be able to import and export more complex models and meshes.

Moi v4 is far, far better at dealing with complex models so if that's an area you're working with I would definitely recommend using v4 now instead of v3. There are improvements in a lot of areas for this - no 32-bit memory limit, multi core object processing on load, a file progress indicator on load rather than everything just freezing until it's done, and improvements in the display engine.


re:
> 3. When exporting to .obj, without choosing to weld vertices, MOI 3.0 sometimes introduces unnecessary
> vertex duplication / seams.

If I understand you correctly those seams are in spots where a surface has a hole in it. Most polygon mesh formats do not have the concept in them of a polygon with an internal hole in it, so when there is a hole in a surface after it has been subdivided, the polygon will be divided through the hole so as to bridge the outer and interior loop so it can be represented by a regular polygon instead of a polygon with an internal hole "island".


> 5. What will be the upgrade price to MOI 4.0 for MOI 3.0 customers?

The upgrade price is $100 between versions, although if you ordered MoI v3 within the past 8 or 9 months or so you will be able to get a free upgrade.


> 6. Any ETA on the release version of MOI 4.0 and the release version of MOI 5.0 ?

I don't have a specific date yet, there will still be one more v4 beta release before the final one. I don't have any idea yet when v5 will be released. It should not be anywhere as long as the gap between v3 and v4 though which has been from a major rewrite undertaken for v4 to get it to be 64-bit and also a Mac native build.

Thanks, - Michael

---

Hi Larry, here is the propeller.stp after doing some untrim and retrim repair work.

For mould.quadlobe.toy.stp delete the problem face and then select the object and run Construct > Planar.

- Michael

Attachments:
propeller_3dm.zip

---

Hello Michael,

Thank you for your reply. I really appreciate how response you are with so many people and the extensive time you take answering inquiries and requests, even though it's probably very time consuming.

Regarding your points / replies :

1. Hi Larry, yes MoI v4 is now a 64-bit program and so can use all your system memory instead of running into 32-bit limits. So it can handle much larger files than v3.

A: Thank you, this will be tremendously useful! There are certain models which are simply too big to export to meshes with 3.0, even with 0 undo levels and 1 thread limit and horrible preview render tessellation.

2. Also there have been some other improvements such as using multiple CPU cores for processing objects so usually it's a lot faster loading files too.

A: Much appreciated as well!

3. It's a somewhat fragile area and when you see the type of distortions you mention it means something didn't go quite right with that. This can usually be corrected by doing some repair operations like untrimming and retrimming the messed up surface. There is a tutorial on these techniques at http://moi3d.com/forum/index.php?webtag=MOI&msg=446.17 .

A: Thank you, I'll review the article. However, I did suspect that the issue was due to the modelling software I use, Autodesk 123D Design, possibly not exporting CAD models correctly. So, to rule that out, I viewed the models which didn't import ok in MOI3D with eDrawings 2019 and they display correctly in that. Just now, as I'm typing this, I also opened the problem models with freeCAD as well and they also displayed correctly in that. They also display correctly in grabCAD's own .sat / .step viewer, even though that sometimes produces artifacts of its own due to not using enough triangles when tessellating.

Maybe a different geometry kernel might help, like Open CASCADE or Coin 3D?

4. Your propeller.stp model has the same issue in v4 as in v3.

A: Just now, looking through my old CAD files, I managed to find an earlier version of that propeller model in 123D's own native format, before I removed some unwanted solids from the scene which I'd only used the sides of as sketching surfaces. When I exported that earlier version of the model as a .sat file, it appeared to have imported correctly in MOI 3D, along with the unwanted solids, but when I checked, the propeller was a joined SRF rather than a solid. I've attached it to this reply for comparison.

4. It is an area that I would like to tune up in the future but it will have to happen earlier in a release cycle rather than at the end.

A: Can geometry kernel improvements to importing be made in 5.0? Or maybe switch to using Open CASCADE or Coin 3D for handling file imports / conversion to .3dm, if MOI handles .3dm better than .sat or .step? Is there a particular CAD model file which is easiest for MOI to handle? I can try changing my workflow by converting to that format from .sat or .step using freeCAD as an intermediary or exchange format for reliable correct import into MOI, rather than importing the .sat or .step files I export from Autodesk 123D design directly into MOI.

5. What happens with propeller.stp if you use IGES format instead of STEP/SAT ?

A: Usually, models imported from .sat appear to have more issues than .step. But I've had cases where the reverse was true.

6. If your software has an option for using IGES type 144 surfaces that could potentially have a trim boundary closer to what MoI needs and so might have less processing needed on it.

A: I don't have any option to choose what type of version of IGES the .sat files I export will use. If I export an earlier version of the propeller model as .sat it appears to display correctly in MOI but is labelled a joined SRF. I attached that file as 'propeller.sat'.

If I import the propeller.stp model back in Autodesk 123D Design and export it again as .sat instead of .step, the result is worse than the straight export of the earlier version of the model. I attached that file as 'step-reimport-to-export-as-sat.sat'. All of the problem models display correctly in Solidworks eDrawings viewer 2019 and freeCAD. Which is why I think the issue is with MOI.

Could it be that the import issues are caused :

a) by an index variable value range overrun? Maybe an index variable is used in a loop somewhere in the import routine for .sat or .step files and the value it reaches exceeds the value range of the data type used? Maybe switching all index variables and counters in the code to 64 bit integer data types would fix obscure bugs and possibly also this one?
b) by floating point variables with insufficient accuracy? Maybe there's snapping occurring when the software performs array indexing by checking a given float variable's or float array position's value to others before deciding whether to add it as a new array position or just as an index to an already existing position, which is actually supposed to be slightly different in value but with a difference too small for 32 bit floating precision? And so what should be two different positions in the array collapse to a single position?

Or maybe MOI can be switched to a different or already existing geometry kernel in a future version, one which might be more battle hardened?

7. If you delete that square plug you can then show naked edges (http://moi3d.com/forum/index.php?webtag=MOI&msg=6051.2) to see where it was supposed to be. There is a very skinny ring there, only 0.0017 units wide.

A: This observation is what made me think that the import issues might be due to insufficient floating point accuracy. Maybe double precision is needed instead of single precision. I have no idea whether freeCAD, 123D Design or eDrawings use double precision or single precision floats but if changing all float variables in MOI's code to double precision is an easy change with no complex or problematic implications or ramifications, it might be worth it to see whether it fixes the issue.

There shouldn't be any square plug. I'm not sure which of the surfaces shown in this cross-section view in 123D Design are getting turned into that square plug :



8. That can generally cause problems having such a small width. But if you select the object you can then use Construct > Planar to fill it in and then it all seems to be ok after that.

A: I'm really really used to working only with solids, only in 123D Design, not with individual surfaces or edges / curves, except at the sketching stage, in 123D Design. That's why I hardly do any modelling in MOI, except for Twist or Flow ops on imported models. I think a better solution for me would be if there was a particular CAD format which is known to be the easiest for MOI3D to handle (either its native internal representation of CAD models or the closest format to that) and import and which could be expected to import most reliably. Then I could try converting to that format using freeCAD, from .dwg, .dxf, .sat or .step files which I can export from 123D Design, before importing the converted model into MOI.

Conversion to an intermediary format before actual import might also be a good workaround in the future. Maybe you could use Open Cascade or Coin 3D to import 3rd party CAD files to an intermediary format or to MOI's own native format or the closest format to its internal representation. And then actually import the file resulting from the conversion to the intermediary or native format of MOI rather than original, initial 3rd party CAD file itself. Just speculating, I might be completely off-base.

9. If I understand you correctly those seams are in spots where a surface has a hole in it. Most polygon mesh formats do not have the concept in them of a polygon with an internal hole in it, so when there is a hole in a surface after it has been subdivided, the polygon will be divided through the hole so as to bridge the outer and interior loop so it can be represented by a regular polygon instead of a polygon with an internal hole "island".

A: I'm ok with seams, I want the seams and I always export models with vertex welding de-activated because it avoids vertex normal averaging where it's best to not average vertex normals. The issue is that there are some redundant seams being created. I've highlighted what I mean in the following image. The blue lines are seams where vertices were duplicated upon export from MOI. The models are were exported 'triangles only'. No ngons are used. The blue seams are between triangles that make up the mesh surface. However, there's more seams being created than should be necessary. I highlighted the ones I think are redundant in red.



Thank you for all of the information, the work on 3.0 and 4.0 and for all the time you're taking to address people's inquiries and requests.

Attachments:
propeller.sat
step-reimport-to-export-as-sat.sat

Image Attachments:
cross.section.png  redundant.edges2.png 

---

Show messages:  1-2  …  83-102  103-122  123-142  143-162  163-182  183-202  203-218