The video can be watched in 4K
Stable Diffusion is The Most Feature Rich & Professional Generative AI Toolkit
It is Open Source and expands in it's depth and capabilities incredibly fast
Compared to Midjourney, Dall-E 3 and other easy to use prompting oriented platforms,
Stable Diffusion can be controlled very precisely by a large selection of different input methods besides text prompting - And it can also be fine tuned (trained) specifically on image + concept (text pairing) datasets.
- It's hard to master, but incredibly powerful
I use my own high resolution "renderer" fine tuned model, that has been trained to be concept agnostic, i.e. it does not have any specific conceptual ideas such as a particular person, object or style to be called in prompting - Hence the logic is that it's a renderer helping me to do highresolution work between photographic work and Fine Art.
The process of making Ultra High Resolution images is complex and highly iterative.
It involves reprocessing the same image many times over, guiding it into the directions I choose, and thus making it far less random than any purely prompt based engines & workflows.
Often Photoshop is a valuable companion to clean out issues, change the design, alterate the image in all the traditional "photoshopping" ways when something very specific is desired and the element of randomness of the AI just won't cut it and prompting isn't enough.
Compared to Midjourney, Dall-E 3 and other easy to use prompting oriented platforms,
Stable Diffusion can be controlled very precisely by a large selection of different input methods besides text prompting - And it can also be fine tuned (trained) specifically on image + concept (text pairing) datasets.
- It's hard to master, but incredibly powerful
I use my own high resolution "renderer" fine tuned model, that has been trained to be concept agnostic, i.e. it does not have any specific conceptual ideas such as a particular person, object or style to be called in prompting - Hence the logic is that it's a renderer helping me to do highresolution work between photographic work and Fine Art.
The process of making Ultra High Resolution images is complex and highly iterative.
It involves reprocessing the same image many times over, guiding it into the directions I choose, and thus making it far less random than any purely prompt based engines & workflows.
Often Photoshop is a valuable companion to clean out issues, change the design, alterate the image in all the traditional "photoshopping" ways when something very specific is desired and the element of randomness of the AI just won't cut it and prompting isn't enough.
Most up to date Ad agencies have been excited about
Generative AI for many good reasons,
But so far have had to sit back and wait to use
But so far have had to sit back and wait to use
the technology anywhere outisde social media campaigns.
This is no longer an issue as AI images can be
This is no longer an issue as AI images can be
fully developed into 100-200+ Megapixel file sizes!
How Agencies Can Benefit From GenAi:
Higher client turnover rate thanks to faster delivery times
Less complexity in production, less back and forth, less time wasted
Less complexity in production, less back and forth, less time wasted
Faster production cycle; foregoing casting, complex licensing negotiations, location scouting & permits, styling, studio & gear rentals, etc.
Being able to confirm with the client that images are just what the client is looking for at early-production stage
Smaller commitments in the formentioned areas means
more flexibility and less risk in being bold in terms of vision
Anything We Can Imagine
