Author Archive

CheckMate Pro v2 Specification: Poles

Friday, August 16th, 2013 by

One of the most common questions we get about the CheckMate Pro v2 standard is about poles. A pole is a set of edges coming into a single vertex, the way edges are arranged around a vertex at the top of a sphere primitive.


A pole at the tip of a sphere primitive.

In CheckMate Pro v2, we do not allow poles with 6 or more edges on curved surfaces. The reason for this requirement is that such poles can cause problems in renderings. The renderer interprets the pole as a sharp point, causing a break in highlights and textures. Such rendering problems are usually evident only from certain angles, so you might not see why we make this requirement. However, we want CheckMate Pro v2 models to work perfectly for customers no matter which angle they choose for a rendering.

If your 3D model has poles with six or more edges on a curved surface or in an area that could have better topology, there are several easy ways to change the geometry to meet the specification. These changes won’t affect the overall shape of the geometry, but will avoid any possible problems with poles.

Solution: Reroute the edges

If your topology has poles on a surface, you can usually find ways to re-route the edges to have better edge flow. Sometimes this will require more edges, but that’s okay. If a guarantee of no artifacts from poles means the poly count is a little higher, customers don’t mind. They also like to be able to select a series or “loop” of edges with tools within their 3D software.

Six-sided poles on a curved surface

Six-sided pole on a surface

No 6-sided poles and more ease of selecting edge loops around the holes.

No 6-sided poles and more ease of selecting edge loops around the holes.

There are often several ways to reroute edges. The example above is just one. Considering the ease of selecting edge loops can often point the way to ways to reroute your edge loops.

When you submit your 3D model for CheckMate Pro v2, our inspectors will assist you in improving your geometry by pointing out areas that should be fixed, and suggesting alternate ways to route your edges.


The CheckMate Difference – July Sales

Thursday, August 8th, 2013 by

Sales of CheckMate models continue to be on the rise, for one simple reason: customers love them! CheckMate models make up just over 5% of our 3D model catalog, but account for 23% of sales revenue.

CheckMate 3D Models Sales

With our new Feature Graph category system, we have a greater need than ever for  CheckMate certified models. We want at least one in each of our 14,000+ categories. Submit to CheckMate and join the wave!


Siggraph 2013: TurboSquid and CGSociety

Thursday, July 25th, 2013 by

On July 24, TurboSquid got together with CGSociety at Siggraph to present a panel called “Teaching CG and VFX Online”. Four instructors talked about the challenges they’ve overcome using online training to teach computer graphics, and then the panel opened up a discussion with the 40-odd attendees.

Michele Bousquet talked about how TurboSquid uses YouTube’s Closed Caption option when creating CheckMate training videos for artists. TurboSquid creates an accurate English language version of the audio so our many non-English speaking artists can use the auto-translate feature to read subtitles in their own language.

Presenters Michele Bousquet, David Luong, Bryan Wynia, Ara Kermanikian

Presenters Michele Bousquet, David Luong, Bryan Wynia, Ara Kermanikian

Other panelists talked about the importance of frequent submission and review of students’ work, daily online contact with students to give feedback, and the ability to approximate live training with webinars.

The panel was organized by Kirsty Parkin at CGWorkshops, the CGSociety arm that offers online workshops in a variety of VFX subjects.

Presenters with Kirsty Parkin and Andrew Plumer of CGSociety

Presenters with Kirsty Parkin and Andrew Plumer of CGSociety

Day One of Siggraph 2013 with TurboSquid

Tuesday, July 23rd, 2013 by

Anaheim Convention Center

Greetings from sunny Anaheim, CA, where the 40th annual SIGGRAPH conference is being held this year. I arrived on Sunday just in time to go to some early sessions and see a few of the exhibits.

Going early gave me time to review the Posters display and play “spot the TurboSquid model”. This year, Andrey Kravchenko’s Masha won the contest hands down.


One of my favorite features of Siggraph is the bookstore. The emphasis this year is on non-platform specific art and technology books, including this suite from Disney.


Dear Santa, I know it’s only July but…

A favorite among early attendees is the Technical Papers Fast Forward, a two-hour roller coaster ride through all the papers being presented at Siggraph. Each presenter gets exactly 30 seconds to preview his or her paper and entice attendees to come to the full presentation later in the week. The Fast Forward is one of the most popular early events at Siggraph, attracting over 2000 attendees this year. Many presenters prepare amusing and informative videos and even dress up as zombies in an effort to attract attendees to their papers.


Crowd at the Technical Papers Fast Forward

The Emerging Technologies exhibition literally got a smile out of me when I visited the Incendiary Reflection presentation. This group from the University of Tokyo posits that emotions can be influenced by our own facial expressions. The image below left shows my actual facial expression, captured by a pinhole camera and displayed on a mirror-like surface in a frame. The image below right appeared gradually a moment later, with my own facial features lightly displaced to form a gentle smile. While it might sound a little creepy, the result actually did make me smile. Future usage could include placement in clothing retailers, where the sight of yourself smiling in that retro 70’s lapel suit might actually convince you to buy it.


 The next few days will be a whirlwind of courses, papers, panels, and of course the vendor exhibition. On Wednesday, I’ll be on a panel for the Birds of a Feather meeting “Teaching CG and VFX Online” in Room 202B, 11am-12pm. Stop by if you’re in the neighborhood!

CheckMate Pro v2 Refinements

Saturday, July 6th, 2013 by

Since we released CheckMate Pro v2 on June 15, many artists have risen to the challenge and improved their 3D models to meet the updated specification. We’ve also heard from some artists with questions about how certain 3D models could possibly fit the new requirements.

We take these questions seriously. Every week, the CheckMate team reviews these concerns and looks for ways to refine the specification to achieve the goal of “better 3D models on TurboSquid” and make it possible for models in every category to be certified. Based on your questions, we’ve added refinements to the specification for certain types of objects.

We are also working on training videos and articles to help you meet the requirements, and to give our inspectors tools for helping you meet them.

Specification Refinements

Because of your feedback, we’ve added these points to the specification:

One-sided planes do not have to be subdividable.

One-sided planes representing leaves do not have to be subdividable.

  • Edge flow and subdividability are not necessary on small, insignificant objects. Screws, bolts, rivets, wires, and other pieces that are very small in comparison with the overall model size can be created with any poly modeling method.
  • Closeup wireframes are required. If the topology of detailed areas isn’t easy to see in the full view of the model, you must provide closeup wireframes of those detailed areas. This will apply to the majority of CheckMate Pro v2 models.
  • One-sided thin objects with opacity maps do not have to be subdividable. For example, in a 3D model of a tree, if the leaves are small planes with texture/opacity maps applied, these planes do not have to be subdividable. Note that if the same leaves are created as boxes with opacity maps, this rules does not apply for CheckMate Pro v2, as these are not one-sided objects. If the boxes are thick, the sides of the object will be invisible, which is not good quality. If the boxes are shaped to fit the leaf shape, they should be subdividable. Note that this point applies only to objects which are very thin in real life and which can be reasonably be represented with a one-sided object, such as leaves, decals, paper, etc.
  • The model cannot include openings (borders) in the geometry that cause parts of the model to become see-through. An example might be a slice or crack where two parts of a hard surface come together, as with a cell phone or motorcycle. If we can zoom in and look through the crack to see the universe beyond, this is a fail. This is a common problem with subdividable models that don’t have sufficient holding edges where two parts meet.
  • The percentage of triangles (three-sided faces) is not a deciding point on realtime models. Realtime models have to have the most efficient geometry possible, so understandably there are a high percentage of triangles.

Guidelines for Meeting the Pro v2 Specification

These next points are not in the specification, but are just guidelines for artists.

  • If you’re having trouble making the edge flow work on your model, consider breaking it into separate objects. Look at how the real-world object is made. If the real-world object is constructed of separate pieces, then you should probably model it that way, too. In years past, we learned to model objects as all one piece, and how exciting it was to learn techniques that made this possible! But this approach is not always the best way to make 3D models in 2013.
  • Don’t mix subdividable objects and realtime objects in the same model. Decide on one objective for your entire model, and go that way only.
  • 3D text almost always has to be edited before it can meet the Pro v2 requirements. Whether you edit the underlying shape or the 3D model itself, you can (and should) make better geometry out of the default text your 3D program produces. Add chamfered edges so the customer can subdivide the text. If you think the text doesn’t need to be subdividable because the customer will never render a close-up, then consider using a texture, decal, or normal map instead.
  • Try for the best topology possible. Our inspectors are trained to look for this point specifically: “Is there any obvious way to make the topology better?” If so, the inspector will advise you on how to do this. Our goal is to make your models so good that customers won’t be able to resist them.

    CheckMate Pro v2 Edge Flow

    CheckMate Pro v2 topology

  • We expect that subdividable models will have more geometry than the minimum necessary to make the shape. There is no problem with adding extra edges to create good edge flow. On the other hand, excessive geometry is not allowed. Try for the minimum necessary for good edge flow. Our inspectors will help you achieve this.
  • If you really like modeling with the minimum polys necessary without regard for edge flow, then make realtime models.
  • If you have to subdivide the model more than 2-3 times to get a smooth rendering at 1200×1200, then you probably need more detail in the base mesh.

Subdividable vs. Realtime

Several artists have suggested that we split the Pro v2 specification into two completely separate specifications, one for subdividable and one for realtime. We have considered doing so, but there are still far more similarities between the two than there are differences. Both require clean geometry (no isolated vertices, etc.), real world scale, excellent textures, accurate product information, and so on. If, at some point in the future, we find that there’s far more divergence in the specifications for these two types of models, we will of course split the specification. But for now, we find that one specification with exceptions for realtime models works best.

I’ll have more answers to questions about architectural models and edge flow in future blog posts, and we’ll have more edge flow training videos soon. In the meantime, I hope this helps answer your questions about CheckMate Pro v2.



End User Event in Netherlands

Friday, June 28th, 2013 by

This week I’m attending the End User Event in Utrecht, NL, a conference for CG artists with a focus on 3ds Max. The conference is small (attendance this year reached an all-time high of 220) but unlike larger conferences, EUE gives you face time with your favorite experts.

Class topics at this year’s conference are largely focused on rigging and rendering. The classes, spread out over five rooms, include both a demo of new features in the mental ray renderer included with 3ds Max 2014, and a sneak peek at V-Ray 3.0 from Chaos Group. Both renderers are boasting new progressive scan tools that can drastically reduce render times.

Zap Anderson and Neil Hazzard of Autodesk demonstrate new mental ray features in 3ds Max 2014.

Zap Anderson and Neil Hazzard of Autodesk demo new mental ray features in 3ds Max 2014.

The conference rooms are located above a pub, and the schedule actually includes periods labeled “Networking/Drinks” for mingling and casual chats with presenters about everything from particle simulation to 3D printing. While this activity often involves swapping business cards, just as often it means you’re making a whole bunch of new friends.

Neil Hazzard (Autodesk) improves his darts game with coaching from Antonia Yordanova (Chaos Group)

Neil Hazzard (Autodesk) improves his darts game with coaching from Toni Yardanova (Chaos Group)

Jeff Mottle ( catches up with 3ds Max guru Ted Boardman.

Jeff Mottle ( catches up with 3ds Max guru Ted Boardman.

Michele Bousquet with Nils Norgren of Neoscape

Michele Bousquet with Nils Norgren of Neoscape

After hours party

After hours party

Terms of Use Privacy Policy Site Map © 2013 TurboSquid