OTOY Forums

is there an Octane for Inventor test version?

I'd like to see if the huge assemblies I have here even start to render before I might also purchase Octane for Inventor. Currently I'm using Cinema (and Octane) but it takes a lot of time to import all the single parts to Cinema and then render parts in Octane. It would improve my workflow a lot if I could render directly from Inventor. Yet I have some doubts whether it is even possible with the huge Inventor model I have here. (parts of a industrial site with complete bays, piping and machine installations... some 2500+ inventor part files.)

Hi, I just put together a demo version of the Inventor plugin, which you can download here....

https://render.otoy.com/plugindownloads/2c/d3/f5/5f/OctaneRender_for_Inventor_2014_2016_3_0_6_13_64bit_DEMO.exe

In general, loading a normal Inventor assembly into Octane is very quick, however I have been sent scenes where there are very very complex parts, used many many times in the assembly, and in this case, the load time into the plugin can be slow. The demo version will give you a true indication of the load time. Pls let me know how you go.

Paul

thanks for the link.

will get back to you how (and if) it works with my assemblies...

seems to work fine here on smaller test scenes.

what I wonder - is there an indication how many objects there are to be exported? I see the number of exported objects increase yet I'd love to have an indication how many there are in total. this way I get at least an idea how long I need to wait for it to finish. would be helpful.

with my first tests I got the feeling that the Cinema Exporter is quite a bit faster but I will test that as well later on.

what I wonder - is there an indication how many objects there are to be exported? I see the number of exported objects increase yet I'd love to have an indication how many there are in total. this way I get at least an idea how long I need to wait for it to finish. would be helpful.

Unfortunately this would require parsing the whole scene first to count the objects, which would substantially increase the load time.

with my first tests I got the feeling that the Cinema Exporter is quite a bit faster but I will test that as well later on.

In theory, the Octane plugin is iterating through the assembly/parts tree as quickly as any other exporter - although the C4D exporter might do something fancy with nested scatters (instances) that Octane cannot do which makes it export quicker.

Paul

Hi,

I checked a model in Inventor vs. an export from Inventor (.stp to MoI, then .obj to Cinema4D) in terms of export times to Octane until the render window pops open.

In Cinema (no matter how it is imported, whether I have 4 objects or 2400 for the same model) - it takes ~5 seconds until the render window opens.

doing the same in Inventor with the standard export settings takes 4:15min until the render window opens....
lowering the tolerance to 1 (instead of 0.001) takes ~19 seconds but some rounded tubes are more edgy now..
lowering the tolerance to 0.1 (instead of 0.001) takes ~24 seconds. the tubes are nicely rounded now.

What does this tolerance refer to? 0,1 times the actual unit (in my case mm)? Just for my understanding so I have an idea what I influence with this adaptation.

btw: when exporting the thing from cinema to Octane it has ~430k triangles. the same goes for the export without changed tolerance setting. when exporting with a setting of 0,1 I get ~130k tris.

seems like this is getting useful. I just need to check the tolerance setting before export.

something useful that might be nice to have: make the input in the tolerance setting take '.' and ',' (point AND colon) to separate digits lower than 0. when I put 0.1 using my number pad the value is changed to 1 as the '.' is not accepted. This is only visible when entering the value, closing the settings and opening them again.

The "Tolerance" factor is best described from the API Inventor manual - pls see below.

From my experimentation - the valid tolerance range is 0.0001 to 1. Numbers higher than 1 are treated as 1.

The tooltop reads:

Specifies the Level of Detail to load into Octane. 0.0001 = high LOD and slower load times, 1 = low LOD and very fast load times. The Inventor LOD cap is 1 (so higher numbers than this are treated as 1).

In Cinema (no matter how it is imported, whether I have 4 objects or 2400 for the same model) - it takes ~5 seconds until the render window opens.

Using the OctaneRender for C4D to load the scene into Octane is a totally different process to using the Inventor plugin - since the C4D model has already been de-instanced (so it's just a single mesh), whereas in Inventor the plugin need to convert all the part instances to meshes, which is a more complex process.

something useful that might be nice to have: make the input in the tolerance setting take '.' and ',' (point AND colon) to separate digits lower than 0. when I put 0.1 using my number pad the value is changed to 1 as the '.' is not accepted. This is only visible when entering the value, closing the settings and opening them again.

This is related to the regionalization settings you have setup for Windows. Can you post a screenshot of your Windows Region->Additional Settings window pls?

Paul

okay. thanks for the clarification!

there is a problem with the export element # display - with models containing more than 10k scene elements the 5th digit gets hidden / overlayed by the 'cancel' button.

right now the huge model I have here is still exporting (40min and going) at a count of 12.000+ elements. I'm interested if this will be able to be transferred to the graphics card as the polygon count might simple be too much for it.... nethertheless this would be great for rendering stuff directly from Inventor as I cannot export this model in a convenient (and not time-consuming) way to Cinema to render it completely.

if you're asking for the input settings: it is set to EN (us int.)


edit:
at the end it took 49:50min to export, ca. 15sec to voxelize it and renders fine.
Now I need someone in the company to give me money to spend. wasn't expecting this to work in the end. (50 minutes would be a lot less time I'll spend than exporting and importing in Cinema)

and it renders like a breeze - just 520mb of graphics memory used (tolerance setting of 0.25 gives me 3.73millions tris)

Good to hear. Yes, triangles don't use a lot of VRAM. It's a shame that the Inventor API doesn't offer a flattened geometry structure (instead of the multi-level instanced architecture in place at the moment) for the plugin to access the scene geometry more quickly. 3.7million polygons would take 1 min on any of the other Octane plugins.

Paul