My Project: Brash Lonergan adventures - Warning AI Discussion Possible

1910111214

Comments

  • DartanbeckDartanbeck Posts: 21,937
    edited December 2022

    Yeah... Diomede. Always travelling.

    I used to travel all the time as a young adult. Been to 36 states. One day I should put my Rosie and little dog Chicha in the car and find out where Diomede is travelling to and meet up!

    Chicha is our new Podenko (look that one up) - she's the shorter, Dacshund version of Podenko.

    She likes to travel by foot... Everywhere! We just recently moved not too far from where we were. Chicha and I have a really nice two block walk to The Lake (Michigan) where there are picnic tables and benches, rabbits, trees and all kinds of things (like rabbits) for her to smell! She'd Love to show you some wonderful sites! You know where Algoma is, right? ;)

     

    Post edited by Dartanbeck on
  • DartanbeckDartanbeck Posts: 21,937

    Bring Brash and Moxie! We'll have an adventure!

    IMG_20220916_101252_4.jpg
    395 x 466 - 76K
  • DiomedeDiomede Posts: 15,388
    edited December 2022

    I would love to travel even more.  I have a lot of Amtrak frequent user points from before Covid but have been afraid to use them.  Next big travel event for me will be to family for Christmas, but intend to stay long enough and have enough room that it will be worth bringing my main computer, and not just my laptop.

    - Last year, I made some headway devoting more time to storytelling rather than content production, but then (a) Genesis 9 release has me thinking about content creation again, and (b) a couple of non-Brash, non-Moxie story ideas popped in my head.  One story idea is even an animation, not a comic or illustrated book.

    (Lowered voices.  If inflation keeps up and gets worse, I may have to take the Arkansas bar exam and un-retire.  Sue somebody!  Ugh.  I planned pretty well in ways I could, but 2/3s of my retirement is not indexed.  Keep that between us.  I have made significant improvement in all aspects of 3D since I left Washington, DC, at least I think so.  It would suck to not be able to spend as much time on 3D now that I am actually learning)

    Post edited by Diomede on
  • DartanbeckDartanbeck Posts: 21,937

    If it comes to that, do you think you could sue a few for me too? LOL

  • DartanbeckDartanbeck Posts: 21,937

    I don't know why, but I always thought that The Adventures of Brash and Moxie was going to eventually be animated.

  • DiomedeDiomede Posts: 15,388

    Dartanbeck said:

    I don't know why, but I always thought that The Adventures of Brash and Moxie was going to eventually be animated.

    I do think I will animate Brash and Moxie for one of my entries for Carrara Challenge 65: Turning Point.  If I do, which is still an 'if,' I think I will use some GMIC filters.  I genuinely dislike the idea of using a photo-real look for Brash and Moxie.  Probably some deep psychological explanation.

  • DiomedeDiomede Posts: 15,388

    Dartanbeck said:

    I don't know why, but I always thought that The Adventures of Brash and Moxie was going to eventually be animated.

    I do think I will animate Brash and Moxie for one of my entries for Carrara Challenge 65: Turning Point.  If I do, which is still an 'if,' I think I will use some GMIC filters.  I genuinely dislike the idea of using a photo-real look for Brash and Moxie.  Probably some deep psychological explanation.

  • DiomedeDiomede Posts: 15,388
    edited November 2023

    Brash Lonergan in...

    Brash Lonergan #32 - Attack of the AI 'Bots

    I created the following image using my Brash Lonergan model in Carrara, generating a render, and then processing in Stable Diffusion.  I will post the workflow in subsequent posts.

     

     

    bb08 brash in scifi.png
    512 x 768 - 446K
    Post edited by Diomede on
  • DiomedeDiomede Posts: 15,388
    edited November 2023

    Workflow for above image

    (1) I had modeled my Brash mesh

    (2) I attached some clothes and hair I created for him in Carrara.  

    (3) I rendered out the full pose and loaded it in the controlnet (openpose) channel to detect the stick figure pose

    (4) In a second Controlnet tab, I loaded the Carrara again and chose the 'canny' function to detect the outlines

    (5)  Here are the prompts.

    Positive

    high quality photo of man running, (1boy) 35 years old, excited, blonde hair, wearing ornate militaristic science fiction ((uniform)), long blue pants, long sleeve red jacket, leather boots, in front of spaceship in landing bay, depth of field, in photorealistic style

    Negative

    low quality, bad hands, mangled fingers, nsfw, nudity

    Post edited by Diomede on
  • Bunyip02Bunyip02 Posts: 9,226

    Diomede said:

    Brash Lonergan in...

    Brash Lonergan #32 - Attack of the AI 'Bots

    I created the following image using my Brash Lonergan model in Carrara, generating a render, and then processing in Stable Diffusion.  I will post the workflow in subsequent posts.

    Go Brash !!!

  • DiomedeDiomede Posts: 15,388
    edited November 2023

    Thanks, @Bunyip.

    But we all know that I prefer non-photo for Brash.  So I changed the Positive prompt to

    comic panel of man running, (1boy) 35 years old, excited, blonde hair, wearing ornate militaristic science fiction ((uniform)), long blue pants, long sleeve red jacket, leather boots, in front of spaceship in landing bay, in 1960s comic style of Dan DeCarlo 

     

    bb10 comic style.png
    512 x 768 - 567K
    Post edited by Diomede on
  • StezzaStezza Posts: 8,731

    wow! yes

  • wsterdanwsterdan Posts: 3,041

    Diomede said:

    Thanks, @Bunyip.

    But we all know that I prefer non-photo for Brash.  So I changed the Positive prompt to

    comic panel of man running, (1boy) 35 years old, excited, blonde hair, wearing ornate militaristic science fiction ((uniform)), long blue pants, long sleeve red jacket, leather boots, in front of spaceship in landing bay, in 1960s comic style of Dan DeCarlo 

     

    Sweet! *This* is the kind of stuff I love to see. 

  • DiomedeDiomede Posts: 15,388
    edited November 2023

    Thanks, Walt. 

    Warning - Rant. 

    Thinking about the Dan Decarlo style has triggered the following musings about the pros and cons of mixing the current versions of AI in my workflow.

    Pro.  Supposedly, there is a way to create a 'style' for Stable Diffusion. The ideal would be for me to create character presets from my custom Brash and Moxie models so that I have consistent characters.  Some solutions include embeddings, face swaps, and LORAs.  And then be able to generate new images of my hero and heroine in settings and using various props consistent with the look of my cast.  One solution includes the ability to train and save a 'style' from a set of images.  For personal fulfillment, I'd like the characters and the style to be things I had a hand in.  A week in, and I already see the potential to combine 3D programs with Stable Diffusion in the way I am thinking of.  Carrara in particular is amazing because of its Swiss army knife set of modeling, terrain, plant, sky, hair, rigging, rendering, and multipass tools.  Blender is in the same category.  One could combine Daz Studio canvases, Hexagon, and Bryce in the same way.

    Yay!!!  - I think I see a great tool in AI to enhance my creativity, not a substitute for my creativity.

    Con.  Ethics, rights, and law.  I am concerned that when my prompt included 'in the 1960s comic style of Dan Decarlo' that the database included Dan Decarlo images obtained under the premise that usage would be limited to research and education.  While my current use is consistent with that premise (I am using it to educate myself - truly), I am concerned that much of the use of stable diffusion violates the premise.  

    Anyway, that was just a rant.  There is a whole thread on similar issues in the Commons.  https://www.daz3d.com/forums/discussion/650861/is-ai-killing-the-3d-star/p1

     

    Post edited by Diomede on
  • OK pure facts and logic

    • a checkpoint or safe tensors file averages less than 10GB
    • it was trained on the LAION5 database which used 512x512 thumbnails scraped from the web
    • exactly how much actual data from any one artist do you think that file holds?
    • a file with all the wikipedia text information on every artist scraped would be bigger I hazard to guess

    is this fair use?

    so far at least a court of law has not ruled against it but given the artists mounting a class action an opportunity to reword their claim

    we wait and learn

  • DiomedeDiomede Posts: 15,388

    My remarks are based upon articles like the following.

    https://apnews.com/article/getty-images-artificial-intelligence-ai-image-generator-stable-diffusion-a98eeaaeb2bf13c5e8874ceb6a8ce196

    I have not done enough research to be confident in my grasp of the facts.  Am always willing to listen.  Happy to be directed to additional sources of information.  yes

  • we will have to wait and see what judges decide

  • grinch2901grinch2901 Posts: 1,247
    edited December 2023

    Deleted since my comments weren't really related to Carrara, I was using DAZ Studio to try some of these ideas.  Long story short: really interesting, probably give me some reasons to start looking at buying more content here. 

     

    Post edited by grinch2901 on
  • DiomedeDiomede Posts: 15,388

    grinch2901 said:

    Deleted since my comments weren't really related to Carrara, I was using DAZ Studio to try some of these ideas.  Long story short: really interesting, probably give me some reasons to start looking at buying more content here. 

    I'd be interested in how you are applying similar ideas in Daz Studio.  Here is a thread in which I used very old Poser figures along with simple objects modeled in Carrara.  If this is smilar to what you are doing with Studio, I would love to see it.  

    https://www.daz3d.com/forums/discussion/659331/p4-dork-is-out-of-the-graveyard-carrara-to-sd-workflow-warning-ai-discussion/p1

    There is a more general thread in the Art Studio Forum.

     

  • DiomedeDiomede Posts: 15,388
    edited May 30

    Gathering workflow reminders.

     

     

    Musings on Learning 3D LORA and Controlnet.png
    1024 x 1376 - 1017K
    Post edited by Diomede on
  • wsterdanwsterdan Posts: 3,041
    edited May 31

    Cool stuff!

    I've been playing with Openart.ai for the last little while and they're constantly improving their consistency and range of styles. If you haven't yet, you might want to watch some of their recent videos showing off some of their different engines, and their editing abilities. 

    As well, they do a pretty good job of easy training for characters, objects and styles.

    Post edited by wsterdan on
  • DiomedeDiomede Posts: 15,388

    Thanks!  Will take a look.

  • HeadwaxHeadwax Posts: 10,385

    all hurtling along in bounds and leaps 

  • DiomedeDiomede Posts: 15,388
    edited June 2

    Here are two separate examples of using the same low res Carrara scene and multipass renders (depth pass and coverage pass) to create a power plant on an scifi alien planet.  I didn't directly use the main render (beauty pass) in this example.  I used SD1.5, interface was ComfyUI, with Controlnet preprocessing the depth or 'scribble' 3D.  The good folks at Digital Art Live (a Daz3D vendor of tutorials) have been exploring ComfyUI and I am picking up some interesting ideas.

    3D to ComfyUI Controlnet Alien Facility.png
    1024 x 1376 - 1M
    Post edited by Diomede on
  • DiomedeDiomede Posts: 15,388

    Headwax said:

    all hurtling along in bounds and leaps 

    Thanks.  Variation of everthing should move as fast as appropriate, but no faster.

    Great to see you around the forums.  yes

  • HeadwaxHeadwax Posts: 10,385

    Diomede said:

    Headwax said:

    all hurtling along in bounds and leaps 

    Thanks.  Variation of everthing should move as fast as appropriate, but no faster.

    Great to see you around the forums.  yes

    Here a lot, wat hing, have trouble logging in on old devices !

  • Bunyip02Bunyip02 Posts: 9,226

    Headwax said:

    Diomede said:

    Headwax said:

    all hurtling along in bounds and leaps 

    Thanks.  Variation of everthing should move as fast as appropriate, but no faster.

    Great to see you around the forums.  yes

    Here a lot, wat hing, have trouble logging in on old devices !

    wat hing - dreaded two left thumbs typo I take it ?

  • StezzaStezza Posts: 8,731

    My guess is watching 

    it's a PITA trying to browse the forums on an iPad 

  • Bunyip02Bunyip02 Posts: 9,226
    edited June 6

    Stezza said:

    My guess is watching 

    it's a PITA trying to browse the forums on an iPad 

    Had already guessed it was missing a "c", don't have an iPad, sounds like I am not missing out when it comes to browsing the forums !!!

    Post edited by Bunyip02 on
Sign In or Register to comment.