Importing Vector Designs into 3D Coat via Blender

[Jax_Cavalera]

This is a quick tutorial I have put together to help other graphic designers using 3D Coat, bring their logo work from their design program into 3D Coat.

 

The advantage of doing this is that you can produce really nice logo proofs for your clients that will provide a higher chance of getting your professional designs selected by clients within the online graphic design industry.

 

The industry is very cut-throat and any step up you can get could be the one you needed to win more bids on jobs.

 

Currently, there are a few bugs regarding the importing of these Vector designs into 3D Coat so it would be really nice if those could be smoothed out.  It would also be really great if in the future there was direct support for importing SVG, AI or EPS file formats.  For me personally CDR file format compatibility would be ideal though I'm fine exporting to more standard formats.

 

It would be really neat if the front surfaces were constructed using larger polygons such as ngons so that they provided a smooth reflection though if the shader issues are fixed they will look great either way.

 

Anyways, to those graphic designers out there that have wanted to do this, here is a video tutorial.. hope you are able to follow what I do and any questions, please feel free to ask below and I'll do my best to assist.

 

The video will take a bit to upload due to our limited internet speed here in Australia.  At the time of posting this, the video is at 15% uploaded.

 

[digman]

It is the shader that is causing that look. Choose the Shader in the picture, It is a PicMat Depth Shader.

Since in Blender you were able apply smoothing hopefully the edges and rounded sections will be clean,  You might have to bump up the smoothing in Blender to get them to be clean in 3DC.

You will just have to test to see if they merge clean in Surface Mode.

Resample to 3 to 4 million polygons to keep the sharper edges and remove the imported triangle structure. Also you can now choose different shaders.  Again test for results as this is a method you are exploring.

Continue your work if all goes well.

 

Sadly at the moment the script is not functioning in Linux Blender 2.68A...

Edited August 16, 2013 by digman

[Jax_Cavalera]

Hey Digi,

Thanks for looking into this further, I will give the tests you mention a go and see if it improves the quality or performance of the imports and render.  Hopefully this workflow can be established as a good work around and perhaps once we isolate exactly where the problems are being generated, we can make a clear feature request/bug report.

[Pix Jigsaw]

I'm not sure your approach is the most efficient way to go.  Have you considered converting your logo to a black and white png and directly importing the image using 3D Coat's Import Image as Mesh feature?  Here's what I did with your logo.

 

1.  I did a screen capture of your logo from you video to get the image.

 

2.  Brought it into Photoshop

3.  Up-rezzed the file. 

4.  Inverted the image (white becomes the object in 3DC) 

5.  Applied Gaussian Blur at about 1.

 6.  Saved as a png. 

 

Then I brought it into 3DC using Import Image as Mesh.

 

In the Import Image as Mesh dialog, I used the logo png as the top and bottom stencil.  I left the thickness at default. 

 

When the object showed up on my screen I Adjusted the thickness with the transform tool and pressed enter.   It comes in to surface mode.  I Switched to voxel room and used standard 3DC tools to make a few variations.

 

Here is the resullt.  It only took a about10 minutes.  I just used standard 3D Coat shaders.

 

 

 

Give this method a try and see what you think.

 

Here is a little doodle I did using a vector drawing program on an Ipad to make the basic shape.  I then imported a jpeg of it into 3DC.  The process was pretty much the same as for the logo, described above.

 

 

 

 

[Jax_Cavalera]

Hi Jigsaw,

Thank you for providing an extra method to be used.  I'm sure that will be helpful for a lot of people who are not as concerned with using this for professional graphic design work.  To answer your question though, yes I did try the method you describe to import the design in.

 

There is a critical problem to consider which deterred me from using this method as hinted above and explained below.

 

The design is vector from the start which means it is able to be scaled to any size without loss of quality on edges and is also the most accurate and sharp representation of the design possible to achieve.  The method you describe will work fine for a bit of fun but for anything professional, you will want to maintain the vector format's quality and integrity.

 

I agree that the other method you show will produce good enough results for the casual usage in say a game, but for professional logo presentations you would definitely want to have the exact pin-point sharp edges and smooth curve lines transported from the vector image to the 3D extrusion.

 

It's something that I saw happen quite often during my more avid graphic design years in the online industry.  A designer would produce their client's design using a method that saw minimal quality reduction and then another designer would use software that was able to produce identical conversion from vector to polygons.  The designs would often be very similar and yet the design that looked sharper in the 3D previews often came out in front.

 

The online graphic design industry is a highly competitive field and any slight advantages you can find to get a foot up over the competition will definitely be worth the effort.

 

I agree that the process I demonstrate is not the most efficient as it would be far more practical to simply import your SVG or EPS or AI or CDR file right into 3D Coat and then perform a simple extrusion with crisp results.  It has been mentioned that the shaders I used in 3D Coat have been the cause of the poor representation due to the wireframe showing all curves as they were seen correctly displayed in Blender.

 

It's unfortunate that the shaders I used would typically be the most popular for logo presentations and the preferred options to have working correctly in this situation.

Edited August 19, 2013 by Jax_Cavalera

[Pix Jigsaw]

The design is vector from the start which means it is able to be scaled to any size without loss of quality on edges and is also the most accurate and sharp representation of the design possible to achieve.  The method you describe will work fine for a bit of fun but for anything professional, you will want to maintain the vector format's quality and integrity.

 

 

Jax

 

You realize that the Blender importer script is approximating the curves in your logo with multiple flat faces, right?  Look at 3:56 and 7:58 in your video.  That's not maintaining a vector format's quality and integrity.  Your object file output approximates smooth curves with line segments and flat faces.  It's the magic of smooth shading camouflages the roughness.  Take a close look at one of your curves in Blender.  How many line segments is it using to round the corner?

 

The image import method is capable of generating higher quality if you start with a quality stencil.  Don't get trapped in a print mindset and think that if you are using a non-vector image as a starting point, the end product is going to result in unacceptable quality for professional work.  The image import stencil method is producing a mid-to-high poly representation of the logo.  In fact, you could bake it down using displacement onto your low poly obj file if you want to smooth out its roughness.  

[Jax_Cavalera]

The way it appears to function is that it is actually maintaining the vector nodes and curve information and then it starts off with a base smoothing value of 1.  As you increase this value, it increases the polygon count so that it can better fill out the vector curves which produces a far superior result with more precise mapping of the vector curves to the point where the human eye would not be able to distinguish between the vector original and the polygon duplicate at extreme close-up inspection.

 

To this extent I am lead to believe that your analysis of how this import script within blender works to be in-accurate.  It's not working on the same principles at all when compared to the import of a raster file format within 3D Coat.

 

The process you describe is doing as you explain within 3D coat for a raster image where there are no solid vector curves for the shape to map within the confines of and hence produces a "smoother version" of the original raster image.

 

Vector file formats such as EPS, SVG, AI and CDR are also able to be exported as DXF files which of cause are 3D Cad file formats.  The only difference between a standard mesh and one obtained via a vector file is that the vector shapes when closed are essentially N-Gons ... the script then simply needs to determine how many subdivisions (smoothing) are required to break the N-Gons down so that they form polygons which can then be better extruded and handled in 3D modelling software as non-planes.  If N-Gons were able to contain curved lines between vertices (nodes) then this would allow for instant importation, however this is not the case and that is why it needs to perform the subdivision process to represent the curve.

 

The key point to note is that during the entire process, the image vector information is not discarded and is instead used for pin-point mapping and conversion.

 

As far as time required to perform this sharper conversion from vector design to polygon plane goes, it is questionable that the process takes longer to perform.  It does require an extra program as the import of vector within 3D Coat is currently not a supported feature though actual time-wise is quite swift and painless.

 

I hope that this better explains the advantages of using the Vector designs with this process.  Of cause as neither of us were responsible for writing the import script it is still conjecture as to which explanation fits the process taking place in both situations better.

"