Why do Staff picked models don't apply to 3D Warehouse 'rules'

Today’s Staff picked Models on the 3D Warehouse include models which are excessively large in poly count,materials and size (some of them 39MB).
It is frustrating when one say’s ‘keep the polgon count low’ and developed a special program for this purpose and the other one is Picking model’s like these.
please click link below for the program :

The “goodness” of models cannot be measured in one dimension, but it depends on the use case (there may be cases where a 39MB is not a sign of bad quality). A certification issued to a modeler (instead of to models) shouldn’t be understood that the modeler is not allowed to upload models that don’t fulfill some criteria. Also, the “staff picks” are a very wide category of anything curious, awesome or unique. For example I don’t think the Wall-E model is seriously useful for interior design.

I agree, any single selection of certified models can not satisfy all different criteria, so for end users it is more meaningful to select models in several categories (for interior design, as components, low poly, for rendering).

Maybe there should be an easily findable categorization (like the catalog browser is for individual manufacturers) for all models by application field/audience?

Do you have any specific examples that you could link to?

I don’t have any specific examples : they are all over the place :smile:
I am more curieus as of how the SketchUp team is seeing this, they blog about ‘great’ Models :

( not good or bad) But staff is picking bad examples IMO.


The 3DW Content Developer program, along with the blog post and resources it links to, were created primarily to help with our ongoing efforts to nurture building product manufacturers who are interested in publishing SketchUp models of their products to 3D Warehouse, and who want/need help understanding what it means to do that in a manner that best supports the needs of the folks who will ultimately end up using those models. We published and publicized the info more broadly since, anyone who creates any model of any thing might also want for their model to be seen as being as ‘great’ as possible.

While we appreciate and value the contributions of manufacturers (or the content developers working on their behalf) and have an understanding about the contexts surrounding the use of those models in projects, we also know that not all models that are uploaded to 3DW are shared or used for the same reasons. The staff picks section of the home page attempts to celebrate a broader spectrum of diversity and focuses less on the more stringent definition of greatness outlined across some of our documentation on the subject.

I appreciate your feedback about the quality of the staff pick models and pointing out the somewhat mixed nature of the message it sends when compared to other efforts. I think there’s more we could do to qualitatively vet featured models and will make sure to pass the message along to colleagues who look after the featured content for 3DW.

I fully agree with Mike here,

I have this kind of models in my projects:

  1. As simple as possible in order to convey an idea and with enough details that it can be built. They are as light as possible to the model itself - these correspond more or less to the description you follow.

  2. Very High poly models that imported into skechup or Thea render, prepared with high definition textures (around 4096pxx4096px) and turned into Thea proxies.

  3. Thea proxies that are placeholders for the #2 models and are even simpler than the #1 models, sometimes only the autogenerated proxy boxes.

All of these models are very good and fit 3 different perspectives of “goodness”. Usually I only use the 3D Warehouse to search for models fitting the #2 description and models fitting the #3 description would also be useful if they would come along #2.

From the 3, I think you can guess that the only ones completely useless would be the box proxies…


1 Like

Thanks for your reply Mike,
I know reflected my own perspective in posting this topic. As a trainer I spent a great deal of my time explaining to students that by uploading large models from the 3D warehouse the files get bloated and leggy, but most of the students consider this the responsibility of the manufacturer, and sometimes they point the finger at SketchUp, they don’t wanna spent time in creating their own models. It is difficult to explain that, by making your own models, you would learn faster and eventually, you will gain some
benefits. The diversity of users all use SketchUp for other reasons,which make it difficult for the team to focus, I understand that, after all : 3D for everyone!!
Again , thank you for your reply , Mike
Jack (actually:)

1 Like

I also teach people to model light, as I’m teaching them to model architectural projects. And I tell them to download models from the warehouse as long as they are NEVER downloaded into the main model. That tends to produce mess…

Finally I teach them how to fix models. That is needed…

The rule of thumb is Big and Heavy models for Rendering. Small and lightfor project management, if possible modelled fast by oneself.

This basic frame of mind has helped me navigate 3d warehouse when I need to, and has had great results in my actual architectural models.

1 Like

I’ve run experiments, where I’ve asked people to fix THE SIMPLEST of things, like zoom extents before you upload, and scaling properly, thinking that if these experiments went well we could automate the checklist. The response rate was not high enough for us to consider semi-automating at this time. @MikeTadros et al have other ideas to pursue, so it’s not as if we’re short on direction, it just didn’t rise to the level of doing it now.

Barry, I do understand you wanting to uniformize presentation at least, however I think you’re giving us a bad service by doing so.

One of the most basic filters we have in 3D Warehouse is preciselly the bad presentation of models.

Scaling is another story, you cannot know automatically what should be the correct scale for your upload.

Were you guys thinking of a checklist that would pop up when a user clicked the upload model button?


“Hey, is your model at the correct scale?”

I imagine someone with less sketchup knowledge would get scared and not submit it. Probably that would be a good thing… :smiley:

1 Like

Yea, I wouldn’t try that for exactly your point: were it me on my first upload, I probably would not upload.

I tried leaving comments. People likely get emails when someone leaves a comment. Were it me, and asked in a nice way, I might fix a thing or two. I do that now, randomly, when people have bad thumbnails.

The saving grace is the 3D Model view, which allows me to view the model, zoom extents, and check it out before commenting, downloading, or generally seeing if I’m interested.

Final point: we aren’t demanding uniform presentation, and I think we’d want more community input before we’d even consider forcing people. We could easily zoom extents in our renderbots, but that would ruin the nice setup that some people do when they try to capture a nice 2D picture in a 3D model, and other cases…

There are really two types of models here. Those meant to be used as components in other models and those that are whole projects of their own. The key to getting a 3DWarehouse that is useful for pros is to be able to filter out ALL the latter. Being able to filter out the former could also be nice when you just want to explore other people’s projects for fun.

Here are some filtering ideas. Most could have a range slider where you can set both the maximum and minimum value you want to display. All could also have a checkbox to enable/disable the filter.

  • Bounding box (E.g. the diagonal length if you don’t want 3 different inputs)
  • File size: (In the component browser the maximum files size could default to e.g. 1 MB. For architectural work hardly any component larger than 1 MB is a good component)
  • Polygon count.
  • Texture file size (Absolute size in MB so you can distinguish heavy geometry from heavy imagery, maybe also texture size to whole model size ratio could be useful)
  • Good origin filter (Some way to filter out models with a badly placed origin, e.g. an origin outside the bounding box)

Another approach is to allow the user to tick a checkbox on upload that their model is meant to be used as a component and not a seen viewed of its own. Making this an opt-in rather than an opt-out would filter out many low quality hobby models. With this checkbox ticked a pre-flight guide could be shown where the user have to tick a number of checkboxes that their model fulfills the criteria for being a good component. Some of these checks could even be automated. Then people searching 3DWarehouse could chose to only show models that are marked as components.

Lastly, these features could very well be Pro only. With all the people publicly saying in the forum they are using Make for professional work I think it might be a good idea to add more features to Pro that are indispensable for professionals but not a big deal for hobbyist.

1 Like

Christina, I love the idea of extra features being available for Pro users. I wonder how it could be implemented, though. I use two or three different accounts for my 3DWH use, so signing in could be an issue.

Also, having higher-quality models be more prominently displayed for free and Pro users could help promote good component practices for all. Many of the free users are creating the bulk of the content, so it could motivate them to meet the higher criteria filters.

Until a programming solution is decided upon, though, what about a fast and easy initial solution? Instead of a pre-upload checkbox list, maybe a comfortable first step would be a bulletpoint reminder list at the bottom of the upload page. It might help train users to think about those points that perhaps they had never thought about. It could include a link to a Help Desk article about good 3DWH component practices. After all, whoever did the lego man the size of a mountain (it’s in there) probably didn’t even know how to scale a model precisely, and perhaps had never thought about scale. So a bulletpoint list along with help desk article links to the specific techniques might be helpful. It would be an easily implemented first step without a lot of programming or cost. Also, it’s a way to get good training out there where everyone will see it repeatedly, which addresses the root cause of these problems.


This topic was automatically closed 91 days after the last reply. New replies are no longer allowed.