How To Use AI In Your Photobashing Process
ฝัง
- เผยแพร่เมื่อ 2 มิ.ย. 2022
- AI (Artificial Intelligence) artwork has been a huge subject of experimentation and debate these last few months. Whether this is a fad or the future, who knows. But if we are going to use AI, my goal is have the AI assist me rather than replace me, I want to find that balance between my own artistic ideas / style and that of the algorithm. And while AI continues to evolve, and so workflows will as well, I wanted to share with you my current attempts at finding that balance. So this tutorial will discuss the methods I've been using to incorporate AI artwork into some of my paintings.
For more info and other lessons, please visit www.neilblevins.com/art_lesson... - บันเทิง
![เกิดใหม่ทั้งทีก็เป็นสไลม์ไปซะแล้ว ซีซั่น 3 - ตอนที่ 59 [ซับไทย]](http://i.ytimg.com/vi/GmW4a8prWUQ/mqdefault.jpg)
Great video! Long time fan BTW. This is exactly how I see this tech being used by professionals as a 'force multiplier' vs. a 'job killer'
Thanks Jason! And yup, that's my hope, can we have the good without the bad? I hope so.
Doesn't matter if the ai becomes good enough to do EVERYTHING without you.
Great philosophy and approach to the tech! Human creativity remains a wonderful thing and can be elevated by a computer assist! Striking the balance is the art learned over time!
Thanks Kevin, glad you enjoyed watching my process. Having a lot of fun with it, and figure out other ways to use the tools!
Try out “knolling”. As a prompt. It many times can give you an entire sheet of little parts of whatever you are “knolling”
And this is exacly how the AI should be used! Great video
Thanks, glad you enjoyed it!
I personally (as a concept artist) feel very intrigued - this would be great for Initial sketches, it would offer more detail and refinement in the same or quicker time... It could then be used to build it out of 3d if needed... Or a base to paint up a concept in photoshop. Very Intrigued by how it would handle environments
Thanks for sharing your approach - it makes perfect sense. Using AI to kitbash or experiment very quickly is great - I have also often seen combinations which I would not have tried, or seen anywhere else before. This is very refreshing compared to finding refs on Google or elsewhere.
Yup. In some ways its far better than photobashing elements, because you can basically guarantee the element is unique. I can't tell you the number of times I've seen the same google image show up as part of different concept paintings :)
Thank you very much, Neil! I was one of the people who asked.
I remember! :) Hope this deeper explanation helps!
Well said Neil! Short and succinct. I’m just diving in myself and learning how to craft useful prompts while also curating the results with a critical eye.
Thanks Steve! And glad you're having fun experimenting too!
Thanks for sharing some insights into your process and your perspective on the technology. I've been really excited about how iterative/collaborative the technology can be so far. I'm having a lot of fun just playing and creating. I definitely have some new ideas to try after watching this. Cheers! Keep creating awesome and inspiring work :)
Thanks Stephen, and glad you're having a lot of fun experimenting with the tech!
So great. I've had fun using midjourney images as bases for extrapolating into larger stories / compositions. Really powerful tool that helps me quickly shake up my patterns and default imagination.
Awesome, glad you've been having fun experimenting too!
thank you for explaining your process and how you utilize AI in your workflow! It was very insightful and I think it's really eye-opening the way you summed up your philosophy on such matters. Can't wait to check out more of your work!
Thanks, glad you found it insightful!
Thanks for sharing such deep insights into your creative and technical processes
No prollem Karan, glad you found it interesting!
I needed this video, since my new boss wants us to photobash ai art based oj original artwork
Super fascinating! Thanks for sharing!!
Thanks, glad you found it interesting!
thanks for sharing. very good approach.
Thanks Christian!
Brilliant video. Thanks.
Thanks Zeno!
Thank you so much, you really inspired me for using this technique for my next project
Much respect for you sir
Awesome, glad you found it inspiring!
Super! Thank you for uploading this as I just started the AI tool integration journey myself. Cheers.
Cool, hope you find some inspiration from it, and report back if you come up with some different workflows!
@@ArtOfSoulburn Roger that and will do. One idea first on my list to try is to use 3D forms as a mask generator that will by used as a visual constraint for the AI to work within, essentially a human controlled silhouette that can have depth.
Great video, found you through twitter and this is very well made.
Thanks, glad you enjoyed it!
Some good info in here, but missing the meat of what I was looking for. I wanted to hear a specific play by play of how to upload the image to be found by the AI, and how to add it to a text prompt. That seems unlcear.
So when I released this video, we were not allowed to discuss specifics about how to do that because it was still in closed beta. So the video is a little vague in that area at the request of the software company. But now that things are starting to open up, hopefully we can start discussing more specifics publicly. In the meantime, if you're on the beta, go to the discord and ask about using image prompts and someone there can give you the basics. Hope that helps!
Thanks very much for sharing your thoughts! As I also mainly use photobashing, like you said AI helps a lot in generating our own photobash elments and skips tedious process like cropping photo, editing, photo painting etc. But what I find the most annoying thing about social media art community in general, they find photobash is cheating and AI art is fake without even wanting to learn more about the benefits. It's kinda sickening really. Sorry for the rants btw 😅
The cheating debate is an annoying one. Whether something is cheating or not is all about what the purpose of your art is. If the purpose of a piece of art is to show how well you paint a tree, and you use photobashing, then that is cheating, since your painting doesn't show the ability that backs up your purpose. If your goal, like for example in doing concept art, is to produce an idea that will later be built in 3d for a videogame or film, then no technique is cheating. And of course lying about how you did something is also cheating, but that's kind of a special case. Anyways, people can always say they don't like the look of photobashing, that's a matter of taste, but as long as you're upfront with your techniques, nothing about AI or photobashing is cheating IMO.
thanks Neil! How do you upload your own stuff to midjourney? I thought it was just the /imagine command in discord?
Check the instructions section in the discord, there's lots of other features including image prompts. If you have more questions, ask on the discord itself, they have asked us not to discuss specific features in detail publicly yet.
Again, very inspiring video - may I ask, which AI software do you use, or is that a secret?
Thanks Reinhold. These images were done with midjourney, but you could get similar results with Disco Diffusion, and likely Dall-e2.
I’m interested in photobashing, but I want to use the AI after I have photobashed multiple images together. For example, let’s say I take multiple public domain images, and I use different parts from them to create a new image. I might take the subject from one image, the background from another image, and some other elements from different images that I want to add to the background. I photobash these different image parts together, and create a new composition from them. But I have photobashed the different parts together in a choppy way, and the results don’t really look refined. The important things for me to preserve here are the contents of the photo, as well as the composition that I have chosen by putting these different images together. So, my question is, would I be able to take an image that’s photobashed in a choppy not-so-neat way, and make it look more natural, so that the different elements appear like they belong together in the same image? Is there a way to use AI to improve the look of photobashed images?
I was actually photobashing different images together because I wanted to create blog photos & mockups that tell a certain story, and have the composition that I was looking for. I wasn’t able to find the exact images that I wanted, so I was wondering if I could create my own images using a combination of photobashing & AI.
I think it would be interesting to see a video on how to modify photobashed images using AI to make it look like the different parts naturally belong together in the same scene (by fixing up the lighting & blending everything together better into the scene), and preventing things from looking like each of the parts are individual cutouts that came from different images. 🤔
I want to merge everything within the photobashed image into something more cohesive, and get rid of any messiness left behind from the photobashing process.
I wonder if this can be done, and what’s the best way to accomplish it with the help of AI.
So I have another tutorial here: th-cam.com/video/iGh_DWPoIB0/w-d-xo.html that focuses on making variations of a preexisting image. It may be worth experimenting with, I suspect it'll do what you're looking for.
Honestly I feel in the future of premium entertainment will be high quality human AI colab work to expedite and increase output. Currently The industry just is killing its self to produce enough. Fortunately I really don't see high end story telling and art combined with good novel ideas and concepts to be something the AI will be able to do 100% on its own. At the end of the day and for a probably foreseeable future the AI will still need a prompt so it while able to extremely enhance your ideas into a medium you have no personal skill with you still need the ability to come up with the prompt. I see these tools only strengthening and increasing output quality of everyone not creating more people interested in being creative. As well as I'm not sure if it will kill art jobs either. With this new tool I feel it will actually increase demand of tv shows and movies and this will empower and reinvigorate the movie, streaming, and gaming industries. Current budgets for shows, movies and games are so high the big companies wont take uncalculated risks i.e. releasing more artistic works and stories. This tech decreases budgets and will allow small studios to break back into the fray. This will effect multiple industries and bring back those mid tier companies like in video games we use to have alot more mid tier developers not full triple but not Indie either. That mid ground in many creative industries died due to the increase in costs for quality art and CGI. This lessens that work load and lets creative people reorganize with a healthy middle available once again is my prediction.
I'm all for AI speeding up the "boring" parts of the process, so I agree about a future where we use AI to produce stuff faster. However, if we produce more entertainment content, I'm not sure this will lead to more jobs necessarily. The demand for TV shows, movies and games is already pretty high, and I'm not sure producing better ones will increase that demand, there's only so many hours in the day. There's certainly a research paper in this somewhere, but I think we've been relying on the increase in population to increase demand, and stuff like cell phones on us all the time have also increased demand for entertainment, but I wonder if we've hit the max here and there's any more new money to be added to the industry, or if the new game is just to change what part of the industry gets what percentage of a static pie of cash. That said, if more mid sized projects get a bigger slice of the entertainment pie instead of it all going to 5 large companies, I'd certainly love that, I want new IPs and not just the same stuff repackaged again and again. We'll see how this all goes, but yes, lots of great discussions to be had on how to use the tech to try and make life better for us all.
@@ArtOfSoulburn see but for instance facebook and other online small tier streamers are fighting like crazy for more content and so are streaming services. It a hungry era. Your right their arnt more hours in the day thats true but by unbind entertainment to time slops we have freed up the ability to inbibe far more and with less acceptance and rewatching the same thing over and over. I believe Facebook said their target goals are around 4k hours of uploaded content watch time per day as a minimum to just keep up with its current user base so they dont run out of content. Ai can allow even these lowest tiers to improve greatly. Plus the decrease in costs then also creates a new lower entey point for smaller budget projects. Youll see a loss in jobs numbers at the top is my prediction but more creators over all will be able to survive working more for themselves createing a more sustainable indie space for people. I can see a time when a kid with a great Idea and a hamd full of friends could make a viral movie with effects on par with current day cgi in a few years. These smaller creators with access to tools that powerful will still be outmatched by current day industry professionals working for themself doing the same kinda thing.
Also even in your static pie analogy not really a believer personally in that theory more smaller companies existing and taking a larger percentage means less executive bloat over all in most cases meaning more artists over all would be able to survive off that pie vs the top guzzling most of the money away not to the artists but to the people running it from the top down.
@@Matt-st1tt I'm not sure the ability to make good content faster will stop the executive bloat. I mean, those videos need to be seen somewhere and marketed somewhere, and regardless of whether you're a big company or an indie artist, those channels are controlled by executive bloat. For example, the number of people making a living wage off youtube is actually quite small, apparently you need millions of youtube views to make even a few thousand dollars. Whereas the CEO of youtube is apparently worth $765 million. And I'm sure you've seen articles about the amount of money (or lack thereof) musicians make from spotify, but their executives are doing fine. Anyways, I'm certainly no expert on marketing and all the economic ins and outs of the entertainment industry, and even if I was, I suspect this tech will change those rules quickly, so the best we can do is explore the tech and find ways to let it help us express ourselves and hope that in the end there's still a way to make a living and also find some fullfillment. And if it somehow grows the industry or allow more small creators to make a living, I'd be more than happy :)
Hi how do you feed image in midjourney? I thought it only takes /imagine commnd.. p.s hue fan by the way
Thanks Sajeel! So midjourney doesn't want us discussing specific features and how they are implemented outside the beta, so go to the discord and ask and someone will tell you the command!
Ai not going to replace artist. But it would take their easy income.