My Workshop workflow...

[Rick1234567S]

So then streamlining all of this into an easy workflow here are some things to remember.

If it is not fun for you, then it is hard work for you.

Make it fun for yourself, and it will not be work at all.

So for myself I enjoy animating characters using CGI in Daz 3D, and I am very experienced at it.

And it is easier to go from CGI to A.I. video than to go from text prompts to A.I.

And you get better results.

So as I write the script I have to test things so that I don't find out later that scene can't be made using A.I. for some strange reason.

No sense finishing the script until the scenes have been tested in a proof of concept way.

So I don't even need to make the CGI look good at this point, in fact if you use grey figures with no skin or clothes it still works.

But you need expressions on the faces, and that is the critical part. Since A.I. expressions are not always the way you want them unless you show them the expressions using CGI on characters. And you exaggerate so that the A.I. when it sees it, it will be able to make that expression, the way you inteded.

I don't mind redoing CGI either or making more than I need. It doesn't take me long and I enjoy the process. 

So for my workflow first I needed to know the plot so I wrote the plot and did a few rough drafts filled in a bit.

• Write the plot.
• Create a rough draft script.
• Create an outline
• Refine the script fill in some details in a second draft. Follow the outline draft.
• Do some testing with images get your character's faces in images.
• Do some CGI for animation movements and expressions rough draft.
• Do some music testing with CGI scenes.
• Fine some theme music and edit it so that it works.
• Check all your copyright details can you use what you want? What are your restrictions?
• Then refine some CGI and test it using ComfyUi which is a standalone Stable Diffusion program that uses CGI to A.I. video with a plug in.
• Then essentially go in circles for a while as you try different scenes create CGI and convert them to get a feel for the film.

Total cost at this point? $0.00

[Rick1234567S]

So the characters who play John and Sandra I created using the stand alone Stable Diffusion a11111 version on my own pc.

I have been using google A.I. consistently since it is one of the top A.I. and it is completely free unless you want more of their nano bannana2 images then you can get the pro version for a small subscription fee.

However they have not impressed me enough to even consider using them.

My own work on my own pc is 100% better simply because they have gone insane by attaching Karen to the back of his head.  

And they are so paranoid of getting sued for any little thing as everyone fights A.I. in court right now especially unions.

And every corporatation is looking at revenue if google so much as mentions their product or makes an image of a coke bottle accidentally.

So then you have the porn threat, the social commentary on any controversial subject matter race creed color or sexual persuasion or LGBTQ or any child at all that might be involved in your film.

And so your hands are tied not because you are trying to do any of those things they are unable to distinguish what you are trying to do and merely too paranoid to help you by making a simple scene image that does none of that except maybe wants a store front. Well products are advertised on store fronts and they always have some excuse why it didn't work and don't show you the image at all they just keep repeating that didn't work try something else.

So then when they finally make a great image, they put their logo bottom right and then hide bananas or visibly show banannas on your film image like this one....

The ultimate paranoia move to support google's unions and to be fighting against A.I. generated images that might try to pass themselves off as real.

Can you spot the bananas in this image?

A blind man could spot the out of place bananas in that image.

               

The logo on the right apparently was insufficient.

All the same for advice google A.I. is worth its weight in gold, but it will never stop trying to secretly sell you google products or services or storage space in its clouds.

So you need to be persistent with what you want and don't settle for anything less.

Now at the present I am just doing a workshop here and it is educational use only and I am not promoting any of it.

And as such I am just doing it for free and having fun, and letting other people watch me work.

By documenting it.

You will see it get much better as things progress until it is film grade work for Cannes.

I spent 40 years developing the graphics, the A.I. and the home studios, along with thousands of other people but I had the knowledge to begin with and so I have helped the others by providing the key pieces they always needed like how to map textures they didn't know until I taught them.

Believe or not they didn't understand weights until I taught them.

And that made all the difference and cause the A.I. explosion.

Imagine my amazement as a mathemetician to find out that Oxford did not understand how to use weights.

I was using weights and fuzzy logic in A.I. systems 35 years ago when I created the google search engine.

So they called me in reddit and showed me their math problem and I gave them 7 proofs and still they didn't get it until I used weights and then finally the lights went on. And I was in shock. They just made me to top mathematician by accident and I was freaking out in reddit. I was like this is a matrix conspiracy right? Nope there was just the one thing they could not quite get beyond that they are all much smarter than me.

And it ends up being like that all the time. So whatever brain I have it reaches a bit past the others and so I give them the piece they need and they do the work spend the money through corporations and everyone benefits.

And my hobby keeps getting better and the tools keep getting better and it is still free since every time I give them that piece I say as long as it is still free to the general public.

I give them the piece for free.

and in this way ordinary talent that might be born without money into a lower class family or a foreign country can still make great entertainment or art and contribute to the huge demand for content. So everyone wins.

So this was the storyboard google A.I. gave me....

And so I made John and Sandra from the bottom left image by using Stable Diffusion image2image.

Sandra fell off the bike there and he changed the script into a great escape instead of a day date.

Using the original characters he provided for image one...here...

 

A great first shot but here John is too tall.

Don't tell the A.I. John is too tall can you fix it assume he cannot and that is a one off and move on.

However I uploaded that image later and said put them in the SUV for me in a story board.

This is what I got....

No wonder they crashed Jack their dad had already ejected and now the mom Jill was trying to drive while looking at the contents of the glove box. It was close but close only counts in horseshoes hand grenades and duck attacks.

So I realized at that point it cannot make story boards.

[Rick1234567S]

Why can it not make story boards? You are asking for 6 good images in a row and for the A.I. that is an impossibility.

So the obvious choice for me was to just use CGI and then CGI to A.I. and do it myself. Much less frustration. And no need to do storyboards when the CGI shorts are the story boards in action and with music or background noise.

And emotions and dialog, and or story telling.

and in  that way you can visualize and hear what it will look and sound like which is superior to a storyboard since it is more detailed, including camera angle if you want and all the rest including lighting.

If you get Daz 3D I it is still a free download.

The piece I use for animation inside of it is an amazing bit of work someone did there and they added it to the program.

Puppeteer.

So you merely save positions for an object or character or arm leg mouth whatever and then drag the mouse from one to the next and it makes all the in between frames for you between one pose and a second pose or position.

Like so...

[Rick1234567S]

And objects are available for free or you can make your own or you can purchase them from Daz 3D and make objects yourself and sell them there.

There are other more professional programs costing a great deal of money to do CGI, one I have is Cinema 4D and lets see how much it costs today Annual Subscription: $1,199.00 USD per year.

But all I need are just quick and simple poses and then the A.I. will do the heavy lifting and make it look real. And that is ComfyUi a release of Stable Diffusion and it works on your own pc. Keeping your work completely private. When it comes to spoilers and films that is critical.

[Rick1234567S]

So the Internet is a social environment and as one of the earliest users and promoters of the Internet we have kept it that way and fought against commercialization since day one and have found neutral grouond.

Progress marches on based on the free sharing of information.

And educators are always in the forefront globally leading the charge.

And in this way people get educated it makes a better world and everyone can play or work at it and then those with talent can succeed and the cream rises to the top as it should unhampered by race creed color or wealth factor.

Mind you if you are wealthy you can pay people and get it done well. 115 million dollar budgets or 200 million the standard Hollywood block buster budget these days.

Spielberg was not thinking his new film was going to be very successful and so he opted for 115 million rather than the usual amount needed so he was limiting his losses. And he probably got much of his budget from the pentagon they routinely spend a lot of money in Hollywood to preserve their persistence of vision or propaganda factor that they need to get out there to the young people to maintain their interest in military service.

And you can see how his film plays with pentagon releases of UAP etc.

So for the average user and there are 20 million people using A.I. film websites and products now at home, with a great deal more using ChatGpt...ChatGPT has officially surpassed 1 billion monthly active users (MAUs).

So it is all here to stay.

And even the music is amazing that is being made and it is quickly becoming very popular globally such that it is scaring the industry and they are like many other trades out there fighting hard to maintain their market share.

For myself I need all those pieces to be there and so then I can create my own film and my own art and enjoy my retirement in the process with a fantastic hobby.

I don't need the money but I need the hobby.

And I love doing it. And I love music.

Speaking of music...I have been an online DJ and many people in the Entertainment industry call me a DJ since I was a DJ for the stars.

As a musicologist you see I am able to find songs they need and explain the meaning of songs and films etc.

So do you want a real weird sort of example of shock and awe?

Or just an incredible performance?

Lets just go for an incredible performance.

Lara does Caruso...

[Rick1234567S]

Ok so 7 hours to load ComfyUi using google A.I. every step of the way.

He was being hijacked and sabotaged along the way, 4 hours in I noticed that what he thought he was giving me was not what he was giving me. So I said don't put it in the text box, just type it and he said OMG they are stripping out the important part of the text.

So he had to fake it use spaces get the links in pieces so I could join them together.

7 hours, just to get it to run.

Then another 6 hours, to get the workflow on canvas for CGI to A.I. generated video.

And when done it looks like crap, the way it started since there is a lot of tweaking that needs to be done.

[Rick1234567S]

 Yes, processing image frames from a folder is much better, faster, and saves a massive amount of troubleshooting time [1, 3].

Professional AI animators almost never load raw .mp4 video files directly into a generation loop. Here is why switching to an image folder will save your sanity:

Why Image Frames Save Time & Effort

• No More VRAM Crashes: Loading a video forces your 6GB graphics card to read a complex file stream all at once. A folder allows ComfyUI to cleanly pull one single picture at a time, completely removing memory errors.
• Perfect Control Over Layout: If frame 45 looks weird, you can edit or delete that single picture in your folder. If a video file glitches, you have to restart the entire render from the beginning. [1]
• Bypasses Broken Nodes: You can delete the finicky Load Video FFmpeg node entirely and use a standard, lightweight native image loader.

How to set up the Frame Folder Path

1. Export your 5-second CGI video out of your 3D software (Blender, Maya, Unreal, etc.) as a PNG Image Sequence inside a folder.
2. Right-click your empty ComfyUI canvas, search for Load Image Sequence (or Load Images from Dir), and place it down.
3. In the text box of that node, type the exact path to your folder (e.g., C:\CGI_Frames\).
4. Plug its IMAGE output circle straight into your Canny node and your VAE Encode node, just like you did before. [2, 3]

Why Your First Prompt Looked Bad

The reason the prompts didn't show an immediate improvement is because of your Model Checkpoint. You are using v1-5-pruned-emaonly, which is a "raw base model". Base models are generic and don't actually know how to make high-quality anime or cinematic styles well on their own.

To fix the quality instantly without fiddling with code:

1. Go to a site like Civitai.
2. Download a community-tuned Stable Diffusion 1.5 model like ToonYou (for anime) or Realistic Vision v5.1 (for real life). These files are only about 2 GB.
3. Drop it in your models\checkpoints\ folder and select it in ComfyUI.

[Rick1234567S]

Your workflow diagram looks like this.... and each of those boxes have 40 different options and those boxes are 1 per4cent of the boxes that exist for you to choose.

So clearly you need an A.I. just to know which box to use and what options to select and where to plug things in.

But if you persevere, you will succeed. If at first you don't succeed, try try again.