How to : วิธีใช้ AI วาดรูปอลังการด้วย MidJourney | ขั้นตอน + คลังภาพฟรี | Riety
How to : วิธีใช้ AI วาดรูปอลังการด้วย MidJourney | ขั้นตอน + คลังภาพฟรี | Riety
ใคร ๆ ก็วาดรูปโคตรอลังการได้ เพียงแค่พิมพ์คำสั่งไม่กี่คำ บอกวิธีการใช้งาน MidJourney ละเอียดตั้งแต่เริ่มจนจบทุกขั้นตอน รวมถึงคลังคำศัพท์และรูปภาพสำหรับนำมาใช้ทำภาพศิลปะแนวต่าง ๆ และทำความรู้จัก AI สายวาดรูปตัวอื่น ๆ อย่าง WOMBO / Dall-E / Deep Dream Generator
Line Official Account ID: @rietyofficial
Instagram: https://www.instagram.com/riety.pahn/
Facebook: https://www.facebook.com/rietypahn
#Riety #midjourney #aiart #digitalart
Content
0 -> Recently there are artworks generated by AI
3.766 -> It's a self-learning computer program that can create artworks by itself
8.671 -> I'm going to tell you about it
11.214 -> with detailed step-by-step process
14.639 -> And I won't only tell you about the steps, but also
16.836 -> the sources for vocabulary and images that we can use
21.583 -> as an input for the AI to create our desired artworks
25.671 -> Let's start with the question: What is AI-Generated Art?
28.501 -> Some people may not know it
30.019 -> AI-Generated Art is
31.701 -> AI with ability to learn various vocabulary and images
35.958 -> and keep those in its database
37.814 -> until it becomes smart enough to generate artworks itself
41.106 -> To compare, it's like a toddler
44.161 -> Before a toddler can speak to their parents, they need to learn words first
47.932 -> like dog, hungry, eat, jelly, something like that
50.527 -> Afterwards, the toddler can tell their mom that they want to eat a dog-shaped jelly
52.971 -> Like form these words into a sentence
54.835 -> AI works like that as well
56.701 -> And since we know how AI works,
59.214 -> we, as humans who're working with AI, have to know how to talk with it
63.614 -> Which words or which commands we should use
67.048 -> that will make the AI understand what we want
69.443 -> so that we get the works that meet our needs most
72.557 -> I'll start by telling a history of AI first
76.715 -> A recent AI that becomes viral now is called Midjourney
80.487 -> because it's an AI that can generate very beautiful artworks like these
84.186 -> Anyway, Midjourney isn't the first AI that can generate artworks
88.087 -> Previously there were Dall E, Wombo, and Deep Dream Generator
92.93 -> The name Deep Dream Generator came from
94.703 -> a goal to develop this AI for people to illustrate their dreams for others to see
100.615 -> which is a very romantic concept
102.544 -> And actually even before that there a program we can play doodle with Google
106.958 -> I see a bread
108.457 -> Oh no
109.127 -> It's steak
110.044 -> It's steak
112.471 -> I haven't drawn anything
113.889 -> only 2 circles
117.729 -> Okay, let's just say I earned a point
119.2 -> The reason I talked about these programs is because
121.772 -> Midjourney is developed using codes that have been used in these previous programs
126.872 -> So, if we understand how it learns,
130.501 -> we'll be able to use it more
132.529 -> The AI we'll be using mainly today is Midjourney
135.629 -> Let's start from the first step
137.829 -> Midjourney is a program in Discord
142.586 -> So the first thing you have to do if you don't have an account yet is to sign up for Discord
147.157 -> If you already have it, you can join my Discord channel Hotel de Mentía
152.129 -> There, we talk about NFT, art, and AI-Generative Art, so something nerdy like this
158.329 -> Now we have almost eight thousand members
159.887 -> You can join too
161.371 -> Once you have a Discord account, open a web browser
165.277 -> It can be Google Chrome or others
166.548 -> but I'd recommend Google Chrome. Why? You'll understand more towards the end of this clip
170.093 -> Type www.midjourney.com
173.623 -> and you'll see a website looking like this
176.371 -> Then, sign in with Discord
178.73 -> Once you've joined using Discord, you'll see the first page of Midjourney
183.911 -> It's a home page, AKA your gallery
188.049 -> where your artworks will appear
189.886 -> And here's the Community Post section where you'll find others' artworks that were previously done
195.757 -> This page is very important and useful for us
199.844 -> because if you click on these images,
204.657 -> you'll see a message bubble
207.715 -> which is called prompt
209.978 -> Prompt is a command or word we use to talk with this program
214.044 -> so that it generates an image
216.201 -> In case you still haven't come up with your own wording,
219.489 -> you can copy these prompts and use them to create your own artworks
223.534 -> so that you understand more
225.157 -> I'm not telling you to just copy them, but you can try them to improve your understanding of this AI
230.271 -> Then, go to Discord
231.644 -> and search for a room called Newbie
236.101 -> It can be any room with this name which is for new comers
238.24 -> And you'll see a box below where you can type
241.4 -> In this box, type a command /imagine
245.101 -> Simply type /imagine
246.886 -> And there'll be a prompt for you to type in the box
250.422 -> The prompt is like a blank canvas for you in the digital world
254.679 -> You can type anything
257.082 -> and the AI will find a similar result to generate an artwork for you
263.186 -> You can type just one word. I used to try typing 'riety'
266.972 -> Actually there's also a paid version of Midjourney
269.671 -> but you can try using for free at first
271.472 -> And if you're watching this tutorial, it can help you save a lot of money
275.134 -> You can fully understand its usage before making a decision whether you'd like to pay for it or not
280.486 -> For the first image, I input "wedding" + "two female asian"
284.773 -> "in the middle of Ring of Fire with sunflowers and a puppy"
288.839 -> Put very mismatched things together
291.199 -> which are "wedding" and "Ring of Fire"
293.798 -> And the results are these images
296.301 -> This is very interesting
297.531 -> We'll see 2 women in the picture
300.526 -> engaging in loving activities
302.901 -> because I input "wedding" right
305.08 -> I also got Asian costumes in different styles
308.306 -> in Thai, Indian, and other styles
312.957 -> so this is Asia in the AI's image library
315.315 -> and "in the middle of the Ring of Fire," I got images in red tones
318.314 -> and "with sunflowers," I actually got sunflowers
321.014 -> We'll see that it's pretty intelligent and it can mash the images well
325.457 -> You might have gotten a better understanding of how it works from seeing this image
328.386 -> Now, if there's any image we particularly like,
331.214 -> you can see there's U1, U2, V2, V3, etc. under the image
336.386 -> If you click the button, you'll get the image in a larger size
341.317 -> For example, if I like the top left image, it's the U1
344.971 -> The top right is U2
347.134 -> The bottom left is U3, and the last image is U4
350.307 -> You can click on the image you want to enlarge
353.872 -> Or if you like the first image
356.158 -> and want it but want to have more variation,
359.887 -> you can click V1
360.843 -> and the AI will generate another 4 images based on the first one
365.7 -> I'd recommend you to find the version you like before clicking enlarge
370.25 -> because the cost is calculated from the amount of time spent on the site
374.633 -> so the more you click enlarge, the faster the free trial is used up
378.049 -> We need to make the most of the free trial
380.282 -> For the second image, I input "emperor penguin" + "in a suit" + "running in the swimming pool"
385 -> + "lots of details"
386.958 -> The phrase "lots of details" that I put at the ending is copied from other people
391.286 -> because I saw people used it and there were really lots of details in the images
395.929 -> We can also put "4K," "8K," or the desired image ratio
401.414 -> Like 16:9 or 1:1
404.879 -> When writing a prompt,
406.858 -> don't think you're writing grammatically correct English
410.666 -> There's no need for "The Red Riding Hood, walking in the park"
414.787 -> There's no need for a sentence like that
415.901 -> We can just use Red Riding Hood + wolf + park
419.768 -> which makes it easier for the AI to understand
422 -> My "penguin in a suit" is kind of cute
425.349 -> but it's just emperor penguin and not much else
427.743 -> Next, let's try "realistic" + "Mark Ryden"
431.587 -> He's an artist I like
433.44 -> + "rococo" + "portrait" + "girl" + "white" + "satin dress" + "in a tower"
437.275 -> After inputting these words, I got an image that
440.058 -> very resembles the style of my favorite artist but not totally alike
443.814 -> This is because Midjourney uses free vocabulary and image sources
449.788 -> which are legal and can be used without copyright infringement
452.801 -> So you won't get pictures from artists that are still alive
456.401 -> but you may get Van Gogh's pictures or other aritsts
458.701 -> who passed away long ago and their works become a public domain
463.729 -> And this gets more fun the more I use it
465.714 -> I try weirder combinations
468.115 -> like piano floating in the sky
470.677 -> Red Riding Hood's cottage. The last man on earth
474.475 -> You can imagine anything. You can input anything and it will always show results
478.34 -> Lastly, I input my name and my friend's name
480.906 -> and got something like this as a result
482.192 -> I think it's very cool, almost like an album cover
486.071 -> Amateur artists can simply use this tool to make album covers
490.127 -> Back to what I've said earlier about the way we talk with AI
493.814 -> To do it, we need to have a word bank
496.009 -> A word bank I'd recommend is a website called AestheticFandom.com
504.586 -> This website gathers beautiful items from different fandoms or interest groups
509.514 -> For example, cottagecore
511.031 -> Those who like cottages, fields, butterflies, flowers
514.715 -> witches, fairies,
516.377 -> can visit the cottagecore site
518.299 -> Or if you're interested in bees,
521.2 -> you can search for Beecore
523.141 -> and there'll be images of honey and other things
525.672 -> In case you don't have any ideas or don't like anything in particular,
529.014 -> you can scroll through the fandoms and see what they have to offer
533.5 -> Another source I recommend to search for words and images is Pinterest
538 -> Some might think that Pinterest is a website everyone knows,
540.686 -> but these images appear in Pinterest because of tags
544.457 -> You can try exploring the tags in Pinterest and see what words they use
547.361 -> and you can use them too
548.778 -> Anyway, if you've tried all this and want to use this program,
552.457 -> but don't want to spend that much money
555.394 -> there's a software that helps make images clearer without the need to use Midjourney
560.843 -> For example, Topazlap which can help enlarge images
564.214 -> but it's not free
566.358 -> And there's another site called colab.research.google.com
570.303 -> where you can make images clearer
572.458 -> So you can do it for fun
573.286 -> But if possible, let's subscribe because they've made a lot of efforts creating this program for us
577.598 -> Some people might use AI-generated pictures as a base
581.786 -> like cooking rice, and then we want to make our own food
584.844 -> We can save these images and process them in Photoshop
588.358 -> so that we can do more edits or paintings
591.313 -> But if Photoshop's too expensive because currently Adobe charges a lot,
596.13 -> you can try Affinity Photo where you need to pay only once
600.913 -> or Gimp, a free program that can manipulate photos
604.944 -> This one's also good
606.286 -> In case you like what you did,
608.491 -> you like AI-generated art,
611.058 -> are there any other websites?
612.415 -> The answer is yes
613.005 -> There're websites that were created before Midjourney
617.329 -> like DALL E and Deep Dream Generator
620.705 -> Deep Dream Generator is one of the first and it's pretty straightforward
624.586 -> It'll ask you to select images you like, choose artistic style you like
628.478 -> and it will generate an image accordingly
631.357 -> I selected my self-portrait and Van Gogh
635.186 -> and I got my portrait in Van Gogh's style which is pretty scary
638.257 -> Another fun website is Wombo
641.679 -> Wombo is also available in App Store and Google play store
646.257 -> To create an artwork, Wombo will ask us to enter a prompt which we can input anything
652.573 -> but there'll also be artistic styles for us to choose
655.517 -> like realistic, throw-back, or Ghibli
658.529 -> We can apply the words we use in Wombo in Midjourney as well
664.475 -> because as I said these programs are developed over time
667.2 -> In Wombo, let's say I like Arcane a lot and I select Arcane image
671.6 -> Then, I can upload my own images
676.294 -> whether it's an artwork I did
679.571 -> and I want something to spice it up
682.643 -> or any other pictures
683.651 -> Or we can use suggested images
686.271 -> The coolest feature which is the reason why I recommended to use Chrome is that
690.4 -> you can connect your wallet with Wombo and use the NFT in your wallet
696.358 -> but it's currently only available for MetaMask
698.358 -> but my NFT collection is in Fantom Wallet
700.729 -> so I can't do it
702.154 -> After we're done generating artwork from Wombo, there'll be an option to download the file or print it
707.243 -> Once we download the file or select to print,
710.801 -> the print will be delivered to your doorstep and we'll get a picture on the wall just like what we wanted
713.6 -> And we can do just everything
715.887 -> So that's for images in general,
717.8 -> but if you want to create a portrait or a picture of a person,
721.83 -> I'd recommend you to read the tips in a website called betchashesews.com
726.547 -> This website is good at creating AI-generate portrait images
732.271 -> For the detailed step-by-step process that I've been telling you,
735.215 -> much of the information is from slaphappylarry.com
741.729 -> You can take a look there. I read the tutorials from this website when I was practicing
745.658 -> After watching this, many people in artistic field might feel worried
750.601 -> about their future career after having been practicing painting for years but now the AI can do exact same thing
755.714 -> As for me, personally I think AI can never replace humans
760.403 -> In what way?
761.693 -> It's true that AI gains advantage in terms of speed and quality of work,
765.894 -> but humans possess skills that computers don't
769.115 -> which are communicative and storytelling skills
772.329 -> We can tell stories
774.121 -> For example, you can make an artwork that tells stories
776.372 -> about each prompt you input
778.557 -> and why did you input it. There's a meaning behind that
782.104 -> Also, when you've generated an artwork
784.179 -> and you use it to continue your work,
785.972 -> it can be useful when you can't come up with new ideas,
789.058 -> or maybe you might have an image in mind but don't know how to begin
793.029 -> I think using this tool is very useful
796.796 -> It's something younger me would really want
799.414 -> because sometimes we know what we want to draw but we don't have ideas
802.843 -> and we can't find any reference of what we want to draw
805.986 -> So, I like to think that it's a very useful tool
809.072 -> more than to think it will replace humans
812.171 -> Moving forward, artists may need to focus more on storytelling
816.044 -> or on explaining the process of the work
819.886 -> For general public, I think this is one of the best thing
823.144 -> because I always believe that art is for everyone
826.518 -> and everyone should be able to make artworks
828.667 -> There's no right or wrong. There's no beautiful or ugly
831.029 -> If you enjoy it and do it, it's enough
833.986 -> And these tools allow us to really create artworks
839.586 -> If one day you get married and you want wedding pictures in your own style,
842.915 -> you can do this by yourself and then continue painting
847.057 -> So, I think it's very useful
849.857 -> and although Midjourney can create very beautiful artworks,
853.357 -> I believe this is an early stage of AI generative art
857.401 -> I believe it will be even more sophisticated than this in the future
861.771 -> Lastly, the value in artworks
865.386 -> depends on how much value we place on them
867.43 -> as well as the equal amount of value receivers of the message and people who purchase our artworks place on them
872.614 -> If two people are communicating in the same way,
874.824 -> this value will never fall
876.871 -> Artists from former generations in Renaissance era
880.524 -> had to mix the oil paints from various minerals by themselves
884.344 -> which results in different shades of colors from different artists
887.415 -> But since there are commercially available paints and equipment,
891.472 -> every one can start making artworks faster and more convenient
895.058 -> It doesn't make the artworks we do with the store-bought paints
899.587 -> renonate ourselves any less or have lower value
902.886 -> This also applies to generative art
904.245 -> Eventually it will become a tool that everyone can access,
908.6 -> and has a chance to further develop
911.901 -> So today is a brief introduction to AI-generative art
916.314 -> That's it for me now
917.322 -> Hope everyone enjoys using it
919.465 -> If you really enjoy Midjourney, please subscribe to it
922.915 -> I believe that such a good program must be very difficult to develop
927.543 -> And I hope you're happy making art
930.349 -> Bye bye
933.271 -> Because I'm enjoying AI-generative art a lot,
936.871 -> next clip we'll have a look at whether the making of the AI-generated artworks
939.915 -> really only takes a split second
942.371 -> If I draw what AI does,
944.414 -> how long will it take me?
946.5 -> Please follow my next clip
948.297 -> If you like it, please click like and subscribe
950.585 -> Goodbye
Source: https://www.youtube.com/watch?v=DQsqTL8U8Pc