BartHays

I hear you, Peter. Unfortunately, there we still see plenty of organizations that have maximum file size limitations, often at 10 or 20 Mb/upload or email

It sounds like you are looking for Screened Poisson Surface Reconstruction a la Meshlab: ( but built in VW)

OK, I understand the misgivings about Acrobat, but this solution seems like a much easier approach (to me, a Windows user) Set up a "Fixup" in acrobat (https://graphicdesign.stackexchange.com/questions/156933/rasterizing-pdf-vector-objects-acrobat/156935#156935) Publish from VW as PDF, and run Fixup. I just tested a 43 page 100Mb VW PDF the Fixup rasterized the pages to 150 Dpi, low-quality JPG, and preserved vector text to a total file size of 11 MB - still 43 pages

I'd have to guess it is based on the layer/layer height. Can you share more about your file set up? How are you setting the height of the chairs?

Yeah, as others have said, it depends greatly on the CNC machine control software. Generally, there are 3 steps to CNC. Draw the object in CAD software -> export to Processing software ( DXF, DWG, STEP, SAT,IGES, etc) Process the object(2D or 3D) in software that calculates optimized tool paths->Export as GCode optimized to the specific machine Run GCode tooolpath file on CNC machine's Control Software Some software can do both 1 and 2 (Enroute, Vectric) Some Machine Control Software can do both Step 2 and 3( Shopbot) Some of these may do all three steps if they are packaged with the machine you are using to do the CNC work. Vectorworks can only do step 1 (unless there is a plugin to process tool paths that I don't know about.)

I wonder if @Jeff Prince will step in here. It sounds like he has more experience than I do. However, we just bought an iPad Pro to try this out. (we are remodeling a museum exhibit and wanted to get measurements of the space. (pro tip: Costco has a 90-day return policy, whereas other retailers have 30 days.) What we found was that the iPad built a fairly sparse point cloud. We could easily get the point cloud into VW and view it. It rounded the corners at the floor and ceiling a bit. If we used a clip cube, and focused on eye level, it did an OK job. (+/- an inch or so when compared to a laser measurement) You do have to build new geometry yourself, using the point cloud as a guide. There is software out there that will create geometry from point clouds (Geomagic design X is one in the machine-design world) and there are some apps for the iPad that will do this too, but none of them are much more helpful than the raw point cloud - in my experience, so far... These apps will create meshes and not VW objects like Slabs and Walls. In my opinion, the "Value Add" of buying a $1000 tablet to scan the room is that it makes it much easier to check for the small details. Outlets, fire annunciators, and soffits don't always get placed or built as expected. So the scan data can be used to check for things that are easy to miss, or hard to verify. In the end, a laser measure and good notes are the better tools for reconstructing an existing space and the scan data is great an an audit too to verify if you got everything you need. I imagine there are techniques to get better results from the iPad but in the end, I think it is not the right tool if you are serious about reconstructing a space primarily from point cloud data. I'll attach a sample of a point cloud we got from the iPad of the lobby of a building. Lobby.zip

I'd say there are a variety of communication tools at our disposal. Whether it is a rendering or a series of construction details, they are all visual communications of a designer's intent. The audience with which the designer is communicating may suggest which tools they use. To my understanding* (and for those who don't know), converting geometry to triangles(one kind of tesselation) is primarily about Normals. A Normal is a vector, perpendicular to a surface. Computer graphics need to know what direction a surface is facing to model how light would interact with the surface. Since a (mathematically efficient) surface can have infinite Normals, we don't want our computer to have to calculate infinite solutions. Computer graphics must sample the surfaces and create an approximation of the curves. Since a triangle is a planar surface with the fewest vertices, it is the most efficient way to convert a surface with an infinite number of Normals into something the computer can calculate. There is some newer technology called "Gaussian Splatting" where the computer "throws" an array of points at a geometric surface to calculate the normals without converting the geometry to triangles first. It is being used in some real-time rendering applications. So, to say Blender "dumbs down" NURBS to a mesh is a bit like saying compiling software to "1's" and "0's" is dumbing down the code. Here is a secret, every shaded (OpenGL) view in VW is a visual display of polygons, regardless of the underlying geometry to compute/store the object. Unless we are working exclusively in top/plan or wireframe, all of our fancy ACIS, CGS, NURBS, or Subdivision geometry is converted to polygons for better onscreen display. I'm 98% sure this is true of Hidden Line views as well. Vectorworks uses code built into the graphics engines of our computers to control the display of those polygons to make them look smooth. So really, all of our section views, details, elevations, etc. are just "pretty pictures" of the underlying geometry - acting as intermediate phases to the ultimate construction of our ideas. Oh, and while Gcode (used in running everything from $100 3D printers to the most sophisticated milling machines) has some capacity for arcs it also simplifies almost everything down to triangles, and ultimately, straight lines. It's just that the code that converts Surfaces to GCode is optimized for that task and will subsample geometry based on the capabilities of the machine being used. So, it is better to have that conversion done by the processing software that outputs GCode than to give it something that has already been converted to triangles. *I'm only a Nerd at heart, I am not a software engineer so my understanding is likely flawed on many levels. I am, however, fascinated by all of this magic that gives us these incredible tools to bring our ideas to life. So, I love learning as much as I can about it all and, when appropriate, sharing my understanding with others. Feel free to correct my explanations if I have misrepresented anything.

Wow, I just tested optimizing a PDF by only flattening Transparency. It cut the PDF from 150 Mb to 50Mb, I think need a Vectorscript to set all VPs to solid fill!

I assume you don't have a copy of Acrobat - there are a few ways to reduce file size in Acrobat. What version of VW are you using? If you have something other than VW Fundamentals,(Designer, Architect, Landscape, Spotlight) you can Publish your set to PDF and in the "Publish to..." Options. Select the box to rasterize the PDF at 150 DPI. That should effectively do what you are asking Preview to do, but all in one step. Even in Fundamentals (I believe) you can export a PDF as a raster file and set the resolution lower.

I'm not an expert in Indirect/Global Illumination. However, as I understand it, Raytracing (basic RenderWorks) follows a ray from each pixel of the final image back to the light sources. When you add Indirect/Global Illumination, the software casts "rays" from each light source out into the scene and blends those outputs into the Raytracing results. The Quality settings for Indirect/Global Illumination adjust things like the total number of cast rays, the number of bounces, reflection, and refraction accuracy(collectively "Samples")to calculate. If you have a ton of lights, each light will cast fewer rays. Lights far from a surface will have those samples spread far apart, so you can get blotchy results or Samples that don't blend well into the raytracing. I assume that the position of the camera affects how many samples it's getting too. ( ticking the "Use Emitter" on lights gives you more access to control of these settings - but I don't use that much. ) I think this is what you are seeing with the large grey area with the circular edge - It looks like an unblended Sample to me. If using the same scene, with the same settings, you get different results, then I would lean harder on memory or software bugs. Re-reading your post, it seems like that is the case. Closing and reopening the file clears it up (for a while) Again, just guessing, but it feels like the rendering is running out of memory before it finishes the blending and just craps out what it can. Any chance you can share a file? I'd be curious how you have your lights and materials set up.

Just a few guesses It looks like you are having sampling or graphics memory issues. The first three images suggest the sampling quality of Indirect or environmental lighting is too low. the last image suggests the memory problem where Samples are getting dropped. I don't know enough about the M1 Mac Mini to understand its specs but it should be more than up to the task. A) try upping sample quality for longer render times B) make sure anything else that is using the graphics card is shut down rendering, ( you tube, Chrome browser, - Sometime I even power down my second monitor to give my graphics card more VRAM to work with.) Bart

I see that an updated VW viewer is on the list for the next update. I'm looking forward to that. The current VW Viewer is a super reduced version of VW so the user interface is geared towards people who know VW. It is terrible for people who don't do CAD or design work. VGX files shared via the VW Cloud seem to work fairly well but the graphics or pretty poor. Also, I have found that the navigation is a bit clunky too, it is easy for novices to get lost in the 3D space. Frankly our clients react best to the Panospheres and a VW "Presentation." it's not a "Fly-through" but It gives them a great sense of the space with a very easy interface. Example(https://cloud.vectorworks.net/presentations/11ee540be8324e318f890a222fb0a2eb/196542/?next=/11ee540be8324e318f890a222fb0a2eb/)

@Pat Stanford I had my PC specs in parentheses, the part after the "-" above from " Win 11...SSD". Testing again here (Win 11, 64Gb RAM, 12th Gen i7 Processor, NVMe boot drive, and evo 870 SSD) I also had parens around my WiFi specs. (802.11ac...) I wish I saved the exact text but I edited it a bit after it wouldn't work. I found the thread about 403 errors in another forum and tried deleting the "()" and it worked.

I just had this issue with parens I was not trying to insert code tags just using parenthesis in a sentence. I kept getting the 403 Forbidden page. Finally removed all the parens and the post went through fine.

Weird, I can't seem tp post a reply without a 403 Forbidden error, Trying a 3rd time.: I have a pretty fast system -Win 11, 64Gb RAM, 12th Gen i7 Processor, NVMe boot drive, and evo870 SSD for projects I'd say my 1Gb file saves in 5-10 sec and a commit takes 15-20 sec to process. Sometimes we get a little lag when the other person sends a commit and VW needs a minute to sync the files before we can check out another element. I connect to VCS with a Wifi, 802.11ac, and a 600Mb fiber internet service.