AI Visualizations-Thougthful discussions

[Luis M Ruiz]

Al image generators, they can create stunning visuals from just a text description or geometry. But are they tools for artistic expression or a copyright nightmare? Let's discuss the potential of Al art, its limitations, and the ethical considerations. What do you think the future holds for this technology?

[bcd]

Every time I experiment with it I can't get over the feeling that I'm stealing.

[BartH]

There is a lot to talk about here. 

 

AI is likely the next Smartphone, it is already ubiquitous, we wont be able to live without it in 5 years.
And, as consumers, users, and creators, we have to make some demands. 

 

We have to have transparency on how the AI models are trained.

We will have mechanisms to monetize the use of AI and we must be required share royalties with the sources that helped train the model.

 

Then, in the spirit of @Jeff Prince, make the machines do the mundane. Please don't democratize creativity. 

 

 

 

[VIRTUALENVIRONS]

15 hours ago, BartHays said:

There is a lot to talk about here. 

I agree with this statement.  The problem is no one is talking.

 

@Luis M Ruiz. Love your work, but I get the feeling that no one (including Vectorworks) understands where this technology fits into Vectorworks.  It/you produce some of the most stunning images imaginable, but there is no link back to Vectorworks workflow.  

 

The Submarine Image was incredible, but it was made from a very simplistic Subdivision object.  In a real project, the client wants to see the real thing can be made, printed, animated, etc.

 

I have an idea that might go a long way towards understanding what the Visualizer can do.  Below are three images.  

First image - the movie image

Second image - What it really looks like

Third Image - The Vectorworks model.

 

If you are interested, I would give you the Vectorworks model to see if you can reproduce or enhance the look of the second image, but only as it is modelled.

 

This is where I see the AI Visualizer adding value.

 

 

 

 

[BartH]

@VIRTUALENVIRONS It's hard to argue that no one is talking about AI.
 

You are correct that AI can produce stunning images. And, I am all for that - for exploring ideas or quickly investigating alternate solutions to a design prompt.

Also, it is correct to say that the link back to a VW workflow is absent. 

I would say these same two things of Pinterest, or a simple Google Image search.

The AI visualizer, however, has the potential of being more connected to an Initial VW concept image.

 

But let me clear this up, AI is not rendering, It does not reference the Model geometry.

 

Using the example above. AI has no idea what this submarine really looks like. It can only approximate it based on other images tagged with "submarine" or other terms used in the prompt.  Secondarily, the AI Visualizer use a secondary image called a ControlNet to help guide the image generation. It basically tells the AI where to put edges. 

Here is a ControlNet for your middle image: 

 

You see, its not reading the geometry of the file, It is looking for edges in the starting image. (It would be better if you had the a shaded view of the model on a white background). From there is it is comparing the terms in the prompt, with millions of other images with similar terms(tokens) and then generating a new image, (guided by the ControlNet). The creativity slider, changes how much bias to give to the tokens and the ControlNet as it generates the images. 
 

So, as to where AI might add value, there is (currently) no way to reproduce "photorealistic" renderings of an, as modeled submarine, in AI. The AI is always guessing as to what "Submarine" is and filling in details based on other images. You can use a variety of controls on top of an image of the Sub (IP adapters and other types of ControlNets) but it is not rendering the model. None of these generated images would be useful to show a client as "rendering of the product".
 

However, they might be useful in terms of getting the feel of a marketing strategy. So I had a little fun.
note: these images were not generated with the AI Visualizer)

 

You can see it is not using the modeled geometry, but rater a snapshot of the model as a reference to influence Image generation.

 

(hmm. I wonder what that signature is all about? )

 

And finally, to you point, the Image to 3D model versions of AI are really in their infancy. So, there is no way to get the products of the AI visualizer back in to a VW model without just putting in the work. 

 

anyway, I hope that helps.

 

Bart

[VIRTUALENVIRONS]

11 hours ago, BartHays said:

anyway, I hope that helps.

HI Bart,

Thanks very much for this and beautiful images.  I do not know how the Visualizer works and although your post was very well put together, I probably never will.  I have 2024, but not service select.

 

I see your point on the model being better on a white background.  I think that is why your image is so accurate.  This model is a Russian Typhoon Submarine, so I am assuming the Visualizer was able to get very accurate data.

 

So, to keep this going, have you used it yet in a client based situation?  Either in conjunction with Vectorworks or on its own?   I am interested in where it fits into designers daily lives.

 

Also, I wonder why it put in the two lights as if it was powered by jet engines?

 

Once again excellent explanation. and imagery.

 

Paul

 

 

[Luis M Ruiz]

With Vectorworks' AI Visualizer being so new (launched about a month ago), it's difficult to say definitively how it's integrating into professional workflows. Like many users, I'm currently in the phase of exploration – testing its features and uncovering its strengths and weaknesses. This will ultimately help me determine its value in my own practice. Open to learn from others too.

[BartH]

@VIRTUALENVIRONS

 

Thank you for the kind words,

 

To answer your question. I have used AI only very minimally for any professional work. With all of the ethical concerns about how the AI models were built, and who gave consent for their work to be used, and who is (or isn't) getting compensated for the use of copyrighted images, I can't ethically justify getting compensated myself for AI generated content.
 

Here is the one exception so far. For an exhibit design, we were proposing a wall of portraits, in the end they would be real photos of people from the local community. For the concept sketch I generated several portraits of random (not real) people to get the idea across. 


 

 

In one other case, I used AI to generate a landscape scene with specific elements that I could take to a professional mural painter and say "Here is a reference for what we are looking for. Can you make something like this, but bigger, better and in your artistic style?"

 

Admittedly, I did both of these with some trepidation. I'm still not 100% confident even these uses are without some unintended consequences on the Creative community. 

 

Like @Luis M Ruiz I'm in the exploratory phases. But I have broad concerns about the legality and ethics of AI.

I am also convinced that it is here to stay. So, I want to be an knowledgeable advocate for Creatives as best as I can.  

 

 

[VIRTUALENVIRONS]

Does the size of the image (within reason) affect how the Visualizer will respond?

[BartH]

6 hours ago, VIRTUALENVIRONS said:

Does the size of the image (within reason) affect how the Visualizer will respond?

 No, (and kind of yes) 

The AI model used by the AI Visualizer ( I believe it was stated to be SDXL) is trained on 1024 x 1024 pixel images. 

It can work well with images that are not square as long as they have a similar total number of pixels (1365 x 768)

The AI Visualizer will resample the input image down to something it can work with before processing. 
 

If you use a more sophisticated Generative AI tool, you can adjust the thresholds of how the ControlNet converts an image to lines.
And, you can have that conversion happen on higher-resolution images before it gets downsampled to the working image. so you can kind of tweak the level of detail that comes out of the control net to be more or less strict with your starting image. 

 

For the black and white image of your sub above, I had to lower the thresholds to get the detail in the rear of the sub.

Now, if we used a Hidden Line drawing of the sub, instead of a rendering, you would have a much more precise ControlNet to work with. 

Can you post that same view as a Hidden Line drawing?  

 

 

[VIRTUALENVIRONS]

1 hour ago, BartHays said:

Can you post that same view as a Hidden Line drawing?  

Hi Bart,

Yes I can, but tomorrow, thank you for asking.

 

I have something I would Like you or anyone else to play with.   In my career, I have modelled four massive site models, one on another planet.

On earth, my site models are all historical.  Below is one of them.  This is the First Fort Henry, Kingston ON Canada.

 

There is a fort there now and they call it "old Fort Henry, but it is actually new Fort Henry.  The one in the video predates the current 'old Fort Henry".    @Luis M Ruiz did some interesting work on a test video of mine, but the visualizer got the timelines mix up between WW1 and WW2.  But the visuals were great.
 

It includes Historical Architecture, defensive earthworks, cannons, etc.

The prompts might be:

Fort Henry French Design, Kingston ON.

Predates current Old Fort Henry Kingston ON.

Perhaps war of independence.

French fort designs 1700

 

 

 

 

 

[VIRTUALENVIRONS]

20 hours ago, BartHays said:

Can you post that same view as a Hidden Line drawing?  

[BartH]

4 hours ago, VIRTUALENVIRONS said:

 

OK, I spent a little more time but I have some real work to get done. 

Here are just a couple more images of the Typhoon, I probably generated 50 - 60 to try and get something  I could pass off as a rendering. I don't think I even got close. the AI is just not seeing the detail in the reference images and accurately conveying that to the generated images. 

I am sure I could get better results with a few hours of work, but why not just do it right in C4D?? 
 

[VIRTUALENVIRONS]

11 minutes ago, BartHays said:

but why not just do it right in C4D?? 

My sentiments exactly for the real thing, but for the abstract the Visualizer is compelling.   It does appear that the Visualizer is aimed at the Artistic world..

 

Good to know you are a C4D person.  It is a great combination.

 

What is your particular area of work or interest? 

[Claes Lundstrom]

On the subject of AI, I played around a bit with AI in Photoshop. Generating full images generates very mixed results. Some ok, some terrible. It does however better when adding elements to the picture, and for extending a picture beyond the borders of the picture. In this example, I started with the rendering to the left, added some motion waves about the boat and a coast line far behind. In the right image, I extended the image on all sides to get more of an overview. The waves probably took 5-10 attempts to get the effect I wanted (Photoshop generates suggestion in batches of three). The coast line was ok within one try. The same applied to the extensions, though Photoshop seems to prefer extending in one direction at the time. 

 

 

[VIRTUALENVIRONS]

Very nice images..  Is that boat one of your design Claes?

 

The lake is very calm, always nice, but have you ever tried to get a wavy or rough water scenario?

 

 

[Claes Lundstrom]

2 hours ago, VIRTUALENVIRONS said:

Very nice images..  Is that boat one of your design Claes?

 

The lake is very calm, always nice, but have you ever tried to get a wavy or rough water scenario?

 

 

It's based on a 1968 Riva Aquarama, specifically based on Lamborghini founder Ferruccio Lamborghini's private Riva powered by twin Lamborghini V12 engines. The textures are UV-mapped from a series of pictures of the real boat. Here is a link for those interested; 

 

 

 

 

On product renderings of boats, you typically use comparatively flat water as too much drama on the surrounding tends to distract more than it adds to the picture. I guess the message is that it's something pleasant, not something dangerous. 

[Matt Overton]

Surely, any thoughtful discussion of AI now must include recent events of ChatGPT literally stealing someone's voice.

 

I'd think they bet the farm and have lost with the hubris of such a move. 

[Jeff Prince]

On 5/16/2024 at 8:18 PM, BartHays said:

Please don't democratize creativity. 

 

Isn't it interesting that the majority of the conversation here is about the "How do I get the tool to do _____ " instead of the philosophical questions @Luis M Ruiz posed?

 

That gives you a window into the answer to this AI debate....

Creativity won't be democratized, rather the imitation of creativity will be commoditized...

...and the consumer won't care.

Why should they when most of our fellow professionals just want to use this stuff to make up for their own inability to create?

Architecture is headed the way of Graphic Design and will largely be a race to the bottom based on price and building codes.

 

Only the wealthy and powerful will be patrons of true artistic pursuits.  Maybe some of us will find opportunity as garden hermits on the estates we once designed.

 

Craftsmanship and the trades are already dying a slow and undignified death, unless you can make a YouTube channel showing someone making trinkets.  And while the uniformed will point to Etsy and the 'rebirth' of farmer's markets as examples to the contrary, the reality is most of that stuff is just drop shipped by Alibaba of delivered to the artisanal cheesemaker's house by Sysco.  Social media is destroying communication.  The world is once again generally swinging towards authoritarianism.  It feels like the people are becoming increasingly self absorbed and trading the 'greater good' for 'greed is good' once again.

 

Maybe AI can work on those things too and become our ambivalent overlord.  More likely it will proclaim "you are bugs" and act accordingly.  That's why AI is so good at producing dystopian images, it's preparing us for the spiritual and physical eventuality of the human race.  I'm for one am not going out like Kent Brockman.

 

Too heavy?

 

Here's what the AI Visualizer produced with the following prompt:

 

Creativity will be commoditized by AI

 

 

 

It's interesting how AI images can sometimes remind me of Omni magazine covers from back in the day.

It's gotta be tough being a sci-fi artist right about now. 

 

and then...

"when the AI creativity party ends with unicorns and Flying Saucers"

 

and finally....

Creativity Factory

 

And people thought Brave New World was scary...

what happens when AI is minding the hatchery?

 

 

 

Edited May 23, 2024 by Jeff Prince