Show Us Your Dawn Renders!

Discussion in 'It's About the Art!' started by HiveWire3D, Nov 28, 2015.

  1. Michele

    Michele Member

    Thanks! And yeah, that's EArkham's ZWorld Broken Road. The first one.
     
  2. Ken1171

    Ken1171 Distinguished Contributing Artist

    Wow, it's December already! ^___^

    Xmas2018_1024.jpg
     
    Gisela, Me195, Rae134 and 1 other person like this.
  3. Miss B

    Miss B Drawing Life 1 Pixel at a Time CV-BEE

  4. Me195

    Me195 Extraordinary

    :roflmao: Nice. Great characters.
     
  5. Me195

    Me195 Extraordinary

  6. Me195

    Me195 Extraordinary

    Thank you. I used to exaggerate facial expressions from that distance until I saw a few of your renders. I learned that small alterations make for better expressions even at that distance.
     
    Hornet3d likes this.
  7. Hornet3d

    Hornet3d Distinguished


    The trick someone gave me some time ago (I wish I could remember who so I could give them the credit) is have an expression close to what you are looking for and then run an animation over 30 frames with the exaggerated expression at frame thirty. Now you can step through the frames until you get just waht you are looking for. It works a treat.
     
    Me195 likes this.
  8. Lyne

    Lyne Dances with Bees HW Honey Bear

  9. Hornet3d

    Hornet3d Distinguished

  10. Ken1171

    Ken1171 Distinguished Contributing Artist

    Excuse me for geeking out, but I've got some interesting results from mixing up Van Gogh's "Starry Night" painting with my render of Mia Felis using a "Convolutional Neural Network" (CNN). The artificial intelligence analyzed Van Gogh's painting using its Tensor AI cores, extracted the art style details (took hours!), and then transferred them to my Poser render. It has basically reinterpreted my render. The result would also transfer the Van Gogh's blue/yellow palette, but I chose to preserve my own.

    So this is the "Vangoghized" Mia Felis result that came out of the neural network. You can actually see the brush strokes. Isn't computer science cool?

    StarryMiaFelis_1200.jpg
     
    Last edited: Dec 13, 2018
    RobZhena, Rae134, Miss B and 2 others like this.
  11. Miss B

    Miss B Drawing Life 1 Pixel at a Time CV-BEE

    I sometimes like to run finished renders through some of Photoshop's filters to come up with something different. I did that once with one of my Bryce renders, and I wound up with a watercolor version of the render.

    Not sure what filter(s) I'd use to get the Van Gough effect though. Very interesting to say the least.
     
  12. Ken1171

    Ken1171 Distinguished Contributing Artist

    I think AI-based filters will be the next big thing. :)
     
  13. Miss B

    Miss B Drawing Life 1 Pixel at a Time CV-BEE

  14. Michele

    Michele Member

    Never heard of that; the effects are nice!

    I tend to stick to Topaz Clean but should be braver about trying new postwork effects.
     
  15. Me195

    Me195 Extraordinary

  16. Ken1171

    Ken1171 Distinguished Contributing Artist

    Hehe she looks like she knows her way around knifes! ^^

    That's understandable. Convolutional neural networks are a specialized field of AI, which is a specialized field of computer science. Image convolution became quite popular among programmers in the late 1990s, when people started to use it to "morph" one image into another. It became so popular that Michael Jackson used it in one of his music videos to morph people of different nationalities and ethnicities one into the next, which was impressive back in the day. It was also used in feature movies to make people transform into other creatures. However, that was just simple matrix interpolation kind of image convolution using control points, to make one image match features from the other as they morph.

    Convolutional neural networks use AI to perform the process, where it can extract several different "layers" of information out of it. In this particular case, it extracts the art style from the artist, and applies it to another image. This is MUCH more complex and computationally expensive than simple matrix interpolation kind of convolution. As a comparison, matrix interpolation convolution can execute in almost real-time, while neural networks take long to compute a solution, and the results can be quite subjective and unpredictable. But that's what I used to "Vangoghize" my Poser render, using Van Gogh's Starry Night painting as the source for the art style. This cannot be done with my home PC, so I used Google's Tensor AI cores web service to calculate the process for me. I think the results are amazing (geekly speaking). The idea that AI can extract the artistic style from one painting, and apply it to any other image is intriguing to me. :)
     
    Me195 likes this.
  17. Hornet3d

    Hornet3d Distinguished

    Well I am not quite sure where 2018 went but as another Christmas and New Year is fast approaching it seems the right time to reflect on the year. With that is mind I would like to thank everyone here for everything, for the Hivewire team for providing the content I have enjoyed using so much and for everyone else here at the forum that has made it such a great place to visit. I have recieved so much support and help for which I am so very grateful. More than that I have had the pleasure in sharing your art and sometimes, the ups and downs of life.

    My sincere thanks to you all and my best wishes and prayers that you will have a great holiday, irrespective of your beliefs. May your god be with you.

    Christmas Card HW.jpg
     
  18. mininessie

    mininessie Dances with Bees Contributing Artist

    Waiting for him

    [​IMG]
     
    Gisela, Janet, Me195 and 3 others like this.
  19. Ken1171

    Ken1171 Distinguished Contributing Artist

    Lovely! I have a suggestion - reduce the eye surface specular so they eyes show properly. :)
     
    mininessie likes this.
  20. mininessie

    mininessie Dances with Bees Contributing Artist

    thanks Ken...but i made that on purpose! ;)
     

Share This Page