Home 3D Art Showcase & Critiques

Class Project - Quality Assurance

Hi everyone!

I am in a class at my University and as part of a project we need to have a little quality assurance "testing" done. Before you read farther you should know that you will NOT in any way be personally cited for your opinions. Our goal is to get an estimate of the general consensus of our work. It will be cited as "Industry professionals said ..." or the "3D graphics community rated..." and only averages will be used.

The test:

What we need from you is a rating on a scale from 10 to 0. You will be rating how accurately each of 4 models' geometry represents the picture below. A ten on this scale would represent a model that is visually congruent to the original house. A zero would represent a model that is different to the point of no longer being the same house as the original picture, with five being the median.

The house:

This is the original picture given to each modeler. This is the image each model needs to be compared to.
AbMt0.jpg


Your response:

Please rate EACH of the 4 models on the 10 to 0 scale. Also below your response please let us know which of these categories you belong to, so that we have a way in which to generalize our results.

  • University level expertise
  • Industry professional (1-10 years)
  • Industry professional (10+ years)
  • Personal/Freelance Experience (1-10 years)
  • Other


If you have any questions please feel free to ask.

Replies

  • Wilder4690
  • Wilder4690
  • Wilder4690
  • Wilder4690
  • Vailias
    Offline / Send Message
    Vailias polycounter lvl 18
    Do you have any criteria other than geometric accuracy?
    Such as polycount or modularity?
  • pinkbox
    B - 7.5 (needs bottom part of the pillars plus stair wideness from c to get closer to 10)

    EDIT - the more i look at B i decide the roof shape and edge lip part isnt as good as C either. so i drop Bs rating to 7.5.

    C - 7

    A and D 1-2ish
  • Wilder4690
    Vailias wrote: »
    Do you have any criteria other than geometric accuracy?
    Such as polycount or modularity?

    The overall concept for this project is determining which modeling process better suits creating models for architectural visualization for real time. The target game engine being UDK. So, in a sense, polycount and modularity play a rather late role in the criteria. But with a home as small as the example, could you do it in a modular fashion or would it be faster to just create the whole house?

    pinkbox wrote: »
    B - 7.5 (needs bottom part of the pillars plus stair wideness from c to get closer to 10)

    EDIT - the more i look at B i decide the roof shape and edge lip part isnt as good as C either. so i drop Bs rating to 7.5.

    C - 7

    A and D 1-2ish

    Thank you for your response!

    If anyone else can help us out, it would be greatly appreciated. All we ask is you rate the homes for how close you think the modeling was to the original and if you think the model is suitable for real time.

    Thank you, in advance, for your help.
  • trenoIX
    A - 7
    B - 8
    C - 9
    D - 5

    Experience: Other
  • Shoy333
    A - 5
    B - 7
    C - 9
    D - 4

    Experience: Industry Experience 1.5 years
  • VelvetElvis
    Offline / Send Message
    VelvetElvis polycounter lvl 12
    A-6
    B-8
    C-6
    D-4

    Experience, 8 years.
  • konstruct
    Offline / Send Message
    konstruct polycounter lvl 18
    A - 2
    B - 3
    C - 5
    D - 1

    Experience: 7 years

    Wilder4690 wrote: »
    The overall concept for this project is determining which modeling process better suits creating models for architectural visualization for real time.

    Its cool you guys are trying to pick apart which method is best suited for a particular job. Something to keep in mind however, is context. Is this meant for a hero piece? Its this an asset to be used in an open world environment?
    How close does the viewer get? should we be able to run around inside of it? How do you plan to texture this? how important is pixel density?

    When I gave you guys ratings, all I was looking at was the attention to spacial relationships. While C looks the best in that category, It would be the worst for say for example, a top down, RTS game. Context is key.
  • IdiotDrummerGuy
    A: 4
    B: 7
    C: 6
    D: 2

    experience: industry pro ~1.5 years

    I agree with what konstruct said about context; it all depends on how players are going to interact with this building.
  • Wilder4690
    Thank you for your ratings konstruct, VelvetElvis, IdiotDrummerGuy, and Shoy333.
    Its cool you guys are trying to pick apart which method is best suited for a particular job. Something to keep in mind however, is context. Is this meant for a hero piece? Its this an asset to be used in an open world environment?
    How close does the viewer get? should we be able to run around inside of it? How do you plan to texture this? how important is pixel density?

    When I gave you guys ratings, all I was looking at was the attention to spacial relationships. While C looks the best in that category, It would be the worst for say for example, a top down, RTS game. Context is key.

    Thank very much for the further insight, konstruct. Your concern for context is definitely a very big issue that we have not probably spent enough time assessing. The context in this case would be for an exterior (currently) virtual flyby of the house. Our hope however was to get essentially neutral ratings that focus on solely on the faithful recreation of geometry, but this in not necessarily possible. What is "close enough" for one situation would not be close enough for another. We will be sure to mention that there is some possible bias due to the context of the model in our report.

    Thanks again everyone.
  • oobersli
    Offline / Send Message
    oobersli polycounter lvl 17
    A:4
    B:6
    C:5
    D:4

    exp: 7 yrs

    none of them seem to have their measurements right and have taken their own creative liberties on objects that are easily seen and created. Rating them rather low on the fact that it doesn't seem any of them tried to find a constant object on there to get a reference of measurement. like using a window/door/fence post and finding that average width/length/height and using that as a ruler/guide for the rest of the building. :\
Sign In or Register to comment.