AI繪圖一次就上手!深入實作教學 Midjourney niji journey ChatGPT v5
Summary
TLDR本视频向观众介绍了AI绘图机器人Midjourney的使用方法,从基础到深入,详细讲解了如何通过文本提示生成图像,并结合ChatGPT提升创作效率。视频中还探讨了Midjourney的主要功能、商业应用场景,以及如何通过不同设置优化图像生成结果。同时,对Midjourney的付费计划和创建私人创作空间的方法进行了说明,并介绍了其动画专精版本niji journey,以及一些免费的替代软件。
Takeaways
- 🤖 AI绘画机器人Midjourney可以将文本输入转换为图片,与文本处理AI ChatGPT互补。
- 🎨 Midjourney的主要应用领域包括平面设计、游戏艺术设计、电影或动画概念艺术以及逼真摄影。
- 🖌️ Midjourney的操作过程类似于编写程序,通过输入关键词或编写程序来生成图像。
- 🌐 Midjourney的官方网站提供了一个展示区域和一个生成区域,用户可以查看自己和他人的作品。
- 🔧 Midjourney的设置界面允许用户选择不同的创作模式和风格值,以优化生成的图像。
- 🔄 Midjourney支持与ChatGPT结合使用,利用ChatGPT生成高质量的提示词(prompts)来指导图像生成。
- 🎮 游戏公司可以使用Midjourney生成游戏概念图,与ChatGPT编写的游戏程序相结合,加速游戏开发过程。
- 🖼️ 用户可以通过简单的指令,即使没有艺术天赋,也可以绘制出脑中的想法。
- 📸 Midjourney的V5模型擅长制作超现实的照片,而V4模型更适合创造氛围感。
- 💰 Midjourney提供不同的付费计划,包括基础版、商业版和专业版,以满足不同用户的需求。
- 🔒 Pro计划提供隐私功能,允许用户设置生成的图片为私有,保护创作内容的隐私性。
Q & A
Midjourney是什么?
-Midjourney是一个人工智能绘图机器人,能够将文本提示转换为图像,广泛应用于图形设计、游戏艺术设计、电影概念艺术等领域。
Midjourney与ChatGPT有什么不同?
-Midjourney专注于图像处理,可以将文本提示转换成图片,而ChatGPT则专注于文本处理,两者可以相互补充,共同创造出丰富的内容。
Midjourney的主要特点有哪些?
-Midjourney的主要特点包括:1. 能够根据文本提示生成图像;2. 应用于多种商业用途如图形设计、游戏艺术设计等;3. 与ChatGPT结合使用可以提高创作效率。
如何使用Midjourney生成图像?
-使用Midjourney生成图像需要输入文本提示,这些提示包括主题、背景、风格、媒介、颜色和氛围等元素。用户可以通过调整这些元素来控制最终生成的图像。
Midjourney的创作模式有哪些?
-Midjourney的创作模式包括MJ版本4和版本5等,不同版本适合生成不同类型的图像,例如V5擅长制作超现实照片,而V4更适合创造氛围感。
如何设置Midjourney的参数?
-Midjourney的参数设置包括创作模式的选择、风格值的调整等。通过输入特定的命令,比如--v 5,可以调整AI生成图像的模型和风格。
Midjourney的收费计划有哪些?
-Midjourney的收费计划分为每月10美元、30美元和60美元三种,不同计划提供不同的图片生成数量和速度,以及一些额外的功能,如隐私设置和同时进行的任务数量。
如何创建自己的私人创作空间?
-在Midjourney中,用户可以通过创建自己的服务器或者通过私人消息与Midjourney Bot交流来创建私人创作空间,以便私密地生成和存储图像。
niji journey是什么?
-niji journey是Midjourney的姊妹产品,专门用于生成日本动漫风格的图像。它接受中文输入,使得中文用户也能方便地创作动漫风格的图像。
除了Midjourney,还有哪些类似的AI绘图工具?
-除了Midjourney,还有其他一些AI绘图工具,如DALL-E、Bing Image Creator等,它们同样可以根据文本提示生成图像,但可能在功能和风格上有所不同。
如何提高使用Midjourney的创作效率?
-可以通过与ChatGPT结合使用来提高Midjourney的创作效率。ChatGPT可以帮助生成高质量的文本提示,然后将这些提示用于Midjourney生成图像,从而加速整个创作过程。
Outlines
🤖 初识AI绘图机器人Midjourney
本段落介绍了AI绘图机器人Midjourney的基本概念和功能,以及它与文本处理AI机器人ChatGPT的互补性。Midjourney主要用于图像处理,可以将文本输入转换为图片,适用于图形设计、游戏艺术设计、电影概念艺术等领域。同时,介绍了Midjourney与ChatGPT结合使用的可能性和创造力,以及如何通过简单指令实现即使没有艺术天赋也能创作出脑海中的想法。
🎨 Midjourney的基本操作和设置
详细介绍了如何使用Midjourney生成图片,包括基本操作流程、必要的格式和内容输入,以及重要的Midjourney设置,如创作模式或图像比例等。通过实例解释了如何输入主题、背景、风格、媒介、色彩和氛围等元素来创建图像,并探讨了如何通过官方网站和Discord社区使用Midjourney,以及如何生成和优化图像结果。
🔧 Midjourney的高级设置和优化
深入探讨了Midjourney的高级设置,强调了创作模式选择的重要性,如MJ版本4和版本5,以及它们在生成超现实照片和氛围感方面的差异。讨论了Style值对AI创作发散程度的影响,并通过实例展示了不同Style值下生成的图像差异。此外,还介绍了如何使用ChatGPT来优化Midjourney的提示词,以及如何结合两者来加速图像生成过程。
💡 创意激发:结合ChatGPT和Midjourney
本段落讲述了如何利用ChatGPT作为一个强大的工具来帮助撰写Midjourney的提示词,通过角色扮演的方式让ChatGPT生成高质量的提示词。展示了如何通过ChatGPT生成不同主题和风格的提示词,并将它们用于Midjourney生成图像。通过实际操作案例,说明了如何根据反馈调整提示词,以及如何将Midjourney生成的图像用于网页设计、标志设计等商业用途。
🌐 Midjourney的付费计划和私有空间
介绍了Midjourney的付费计划,包括不同的月度和年度订阅选项,以及它们提供的功能和图片生成额度。讨论了Pro计划的附加功能,如匿名模式和更多的Fast Hour。说明了如何在Midjourney中创建私有服务器和DM,以及如何通过niji journey生成动漫风格的图片。最后,提供了一些免费的替代软件选项,供不愿意支付Midjourney订阅费用的用户尝试。
🚀 Midjourney的双胞胎兄弟:niji journey
介绍了Midjourney的动漫风格版本——niji journey,这是一个专门生成日本动漫风格图像的AI机器人。解释了niji journey的使用方式,包括加入Discord社区和与niji journey Bot互动。强调了niji journey接受中文输入,并展示了使用中文提示词生成的动漫风格图像示例。
📚 总结与展望
总结了Midjourney的主要使用方法和功能,包括如何使用文本提示生成图像、不同创作模式的特点、以及如何结合ChatGPT提高效率。讨论了Midjourney的付费计划和私有空间的创建,以及niji journey作为动漫风格生成工具的介绍。最后,提到了将会持续关注和分享有关Midjourney和有用AI工具的更新,鼓励观众关注频道和社交媒体以获取最新信息。
Mindmap
Keywords
💡AI绘图
💡Midjourney
💡ChatGPT
💡文本提示
💡创作模式
💡图像比例
💡风格和媒介
💡色彩和氛围
💡Discord社区
💡免费试用
💡付费计划
Highlights
介绍了AI绘图机器人Midjourney的基本概念和使用方法。
Midjourney与ChatGPT可以相互补充,一个处理文本,一个处理图像。
Midjourney的主要应用领域包括平面设计、游戏艺术设计、电影或动画概念艺术等。
Midjourney可以将输入的文本转换成图片,类似于编写程序。
Midjourney的使用涉及输入必要的格式和内容,以及重要的设置如创作模式或图片比例等。
Midjourney的官方网页可以直接在浏览器中使用,与Discord社区紧密相连。
Midjourney目前暂停了免费试用,具体恢复时间未定。
Midjourney的不同版本(如V4和V5)适用于生成不同类型的图像。
通过调整Style值,可以控制AI创作的变化程度。
结合ChatGPT可以快速生成高质量的Midjourney提示词。
Midjourney提供多种付费计划,包括不同的图片生成额度和使用权限。
Midjourney的Pro计划提供匿名功能和更多的Fast Hour。
介绍了如何创建私人创作空间以及与Midjourney Bot进行私聊的方法。
Niji Journey是Midjourney的姊妹版,专注于生成日本动漫风格的图片。
提供了一些Midjourney的替代软件,供用户免费试用。
Bing Image Creator是另一个可以生成图片的工具,但生成速度较慢。
鼓励用户通过Midjourney发挥创意,尝试不同领域的创作。
Transcripts
Do you want to learn about AI drawing
but don’t know where to start?
Today we will take you from the most basic to a
step-by-step in-depth understanding
of the most popular AI drawing robot,
Midjourney.
Our AI in-depth teaching video has come
to the second episode. We have already mentioned
how to use ChatGPT
, so let’s talk about Midjourney today
. These two robots can be said to complement each other. One
is for text processing and the other is for image processing
. So in addition to introducing Midjourney today,
we will also combine ChatGPT in the second half.
Using them together , these two AIs
can create a lot of sparks when they come together
, and they have a wide range of uses
. So today we will take
a look at what Midjourney is
and how to use it. What are the main features of Midjourney
? For some uses,
let’s briefly distinguish several relatively large commercial uses
, such as graphic design, game art design, film
, or animation concept art
, or realistic photography.
For example, some logos for commercial design
or website design and packaging design
are all fine Use Midjourney to design a draft first
, and you will find that there are some pictures on the screen
, which are very similar to English words.
I don’t know what language they are writing in
. This is because Midjourney
has no way to generate meaningful text content.
It can digest what we send.
These text commands , but it cannot
produce text content like ChatGPT
, but this part is also very easy to solve
, just use some third-party App or Photoshop
to repair it
, and then we will see the game art part
Many game companies are
now using AI to generate game concept maps
. What you see on the screen is, for example,
character design or game environment design as an example
. If you use ChatGPT
to write game programs,
this is the process of generating games.
It will really speed up a lot. In
addition, this Midjourney can also generate
, for example, animation-style
or movie-style pictures
. For example, the pictures on the screen are pictures generated with the keywords of the four words movie picture
. Of course, there are also pictures in it. Adding some other elements
will produce such an effect
. Then use Midjourney
to generate some creative pictures.
For example, Victorians swipe their mobile phones on the side of the road
, or monkeys make trouble in the office, which
are more dramatic scenes
. So from From now on, even if you don’t have artistic talent
, you can
still draw the ideas in your
mind
through simple instructions. Next
, we will show you
how to use Midjourney to generate pictures from the most basic operations
. When the AI robot gives instructions
, I will introduce some necessary formats
and the content we need to input.
In addition, we will talk about some
very important Midjourney settings , such as the creation mode or image ratio, etc.
These are very important. Important elements .
Of course, at the end, we will also introduce
how to combine ChatGPT
to make the pictures produced by Midjourney better
. First of all, we will talk about
the AI model of Midjourney.
What it actually does
is to convert the text we input into
pictures.
Rather than drawing, the whole process is more like entering keywords
or writing a program.
Like
this , the
Prompt field you see on the screen
contains the keywords I entered.
Midjourney will insert this field In the bit, what
I wrote is converted into a picture
. For example, the Prompts I gave it here
are monkeys dancing in the office.
The 1980 behind it is the year I specified. After sending it like this,
it
will produce the following picture
, then Let’s take another example
. In my Prompts field, I write
Victorians using smartphones.
It will produce the following picture for me
. However, it is impossible for AI to produce these simple words
immediately in
our minds. Ideal for perfect images So
today, in addition to the quick basic teaching, we
will also teach you
some (more) advanced usage skills of Midjourney
. Next
, we will see the official website of Midjourney
, which is related to ChatGPT is also a website. As long as
your mobile phone and computer use a web browser
, you can use it directly.
The
official website of Midjourney
will be connected to two very important places.
The first one is
the generation area that actually uses AI to make paintings. It’s the
button on the right , and the second one is the display area on the left
. Any pictures you generate with this AI
will be displayed here synchronously
. After we click in, the grid on the left is the display area on the official website just now
. You can see your works or other people's works in
it. You can enjoy everyone's achievements here
, just to compare with each other
. It's a bit like a gallery. As for the generation area that
you just clicked on the right side of the official website,
you will enter Now that there is a Discord community like this
, if you don’t know
what Discord is, let me explain it to you
a little bit
. It
is a platform that combines communication software and forums. One of the biggest features of Discord
is that there are a lot of robots on it.
Everyone You can interact
with these robots on this platform,
and you can even design your own robots on it.
So
the Midjourney AI we are going to talk about today is
actually a robot on Discord
. Here is a bad news for you.
That is, Midjourney
is currently suspending the free trial.
According to the official statement, it is because of being abused
. They have not mentioned
when they will resume this trial
. As for the price of Midjourney,
we will mention
it in more detail later. We will now Go back to
the part used by Midjourney.
When we joined Discord
, we saw a row of lists on the left .
This is called the channel list
. Before there was a free trial, these
people can only
create on these channels that start with Newbies . Normally you
You will see this block here, and you will
keep scrolling up
because everyone is creating this in the same place
. As for the paid version
, you can have your own space without being crowded with everyone.
As for how to have your own private space
We
will talk about
it later. To start drawing
, go to the input field below to start a conversation
with the robot
. When Midjourney wants to talk to the robot, you
need to type a slash in the input field , and then it will jump out like this Form Then
this form is the function menu of the robot.
The first option we see is the slash Imagine
, which is actually the AI drawing function we want to use. You can
directly click on this item, or you
can directly
press Enter on the computer. It is called into the input field
, and then after grabbing it into the input field,
the space behind the Prompt
We can start to input the content we want to create. What kind of things can we input
?
This is a huge knowledge. Now
there is even a profession like Prompt Engineer
, because behind the scenes, you have to start to understand
the AI that digests text. The method
, which text is placed first and
which text is placed last will
actually affect your final result
, so let’s take a look
at the general directions set by Prompt
. Let’s explain a little bit one by one.
First of all, you
The theme and background
of the theme must be placed at the forefront.
Suppose I want to design a new Disney princess,
I will describe its appearance
. Apart from the appearance, what kind of thing is
it doing , or what is it doing? The scene and
environment can describe this part
. Here we see that it is all in English
because it is input in Chinese
. In fact, Midjourney can’t understand it
. We will talk about another
Midjourney twin later. It is listening. Can understand Chinese,
but the method of use is exactly the same
, so you can listen to this part first ,
and then come to the second part is style and medium , which is to describe the style of this painting
or picture
. It can be one or
The names of multiple artists, manga artists,
directors or illustrators
, such as Disney, Hayao Miyazaki
, Van Gogh, Junji Ito, etc. can be filled in directly
, or you can also fill in a wider range of styles
, such as Impressionist paintings, abstract art, etc.
These nouns can be added.
Then you can also fill in the specified art medium .
What does the art medium refer to
? You want to generate a hand-painted illustration
or a graphic design drawing. Or
you can specify the photo, the most
interesting part is the part of the photo
, you can also give it the camera model or
the photographer ’s name as a reference
, and finally the color and light , or
the adjective of the atmosphere , such as the afternoon sun in the film lighting
Product lighting,
etc. Adjectives for atmosphere
, such as nostalgic,
1980s, futuristic,
etc. If you want to specify a color, you can write it
, but I think the color should be used with caution
because sometimes it is specified
The color may not appear where you want it to appear
. As for the three major items we mentioned above,
it is not necessary
to write about each of them
. It is just for your reference,
so let’s share it with you next.
For example, here I want to generate a
new Disney princess, I
just want to describe its appearance
, then I set the style in Spanish portrait
, and at the end I emphasized again that I want the style of this Disney Pixar
Then pay attention, each prompt
must
be separated by a comma and a space,
so we have not mentioned color or
atmosphere adjectives here,
just a simple example to demonstrate to everyone ,
except for Prompt Except for the keywords,
you will see this -- at the end
. These are some basic settings
-- ar 2:3 is actually setting the aspect ratio of the picture
. ar refers to the abbreviation of Aspect Ratio.
I will mention more settings of this kind later .
After writing it, let’s test it
out. Send out the prompt of this Disney princess .
That’s it. Then press Enter . It will
take a little time to generate a picture in Midjourney
, about In a minute it
will also show the percentage of the progress on the screen
and it will go from a blur like this on the left
and slowly start to show your theme
, let's say this is 10% progress
and this is 30% progress Then
when it is finished, there will be four pictures like this
. At this time, you may have a question
, that is, if you use the same prompt
to send it again, will
it generate exactly the same picture?
The answer is no
, it is
Midjourney. It is sent with exactly the same keywords
, it will not generate the same pictures repeatedly
, that is, every batch of commands you send out
will
produce different results
, then let’s take a look at the four pictures we
just generated,
you will find The styles of these four pictures are a little different
, and there are eight buttons below
to perform some actions on these four pictures
, and you will see that there are 1 to 4 here
, which respectively represent the above 1 2 3 4
Then let’s take a closer look at
what U and V mean
. U stands for Upscale
, which means that after you click on it, one of 1 to 4
will be
upgraded into a complete large picture . If
you like all four pictures If so,
you can also press from U1 to U4
to upgrade all of them
. The V part below stands for Variation
, that is, the change type
. After pressing it
, you can make
a picture that is very similar to 1 to 4 just now
, that is, the possible composition and color.
It will be the same , but
the details of the face or which accessory may be changed
, so let
’s demonstrate
it to everyone In this way , give
me a complete big picture , which
can also be regarded as our final product. If you want to download it
, you can click in and press the right
button to store it in our computer
. Then we Going back to the eight buttons just now
, suppose
I want to press the V1 below
to generate a variation of the first picture. What will happen? After I press V1
like this
, it will generate four pictures like
this. The first picture just now is very similar
, but the face or accessories are a little different. There
is a redo button on the far right of the eight buttons .
If you are not satisfied with the four pictures today
, you can click this to start again.
It will generate exactly the same Prompts
and then generate four completely different images
. As just now
, it is a complete process of using Midjourney
. Usually when we use Midjourney, we
just keep repeating this step
until it produces what we are satisfied with.
So far, as we just said, it
actually takes a little time when it is generated,
so you can also
send the exact same Prompts several times
at the beginning
(recommended 2-3 times) and then let it run by itself, assuming
we We have finished our works today
, let ’s take a look at the part of the gallery we just mentioned on the official website,
that is, all the completed works
will be placed in the gallery of the official website simultaneously,
let ’s take a look at this part
, there will actually be Divide one area
into your own works
, and the other area is everyone's works. Suppose you see other people's paintings in this gallery today
and
think they are very beautiful.
You can also put the cursor on this picture
and you will see where others used it.
You can copy some Prompts
and use the same Prompts to generate pictures
Next, let’s talk about some more advanced things
to optimize the results of images generated by Midjourney
. Midjourney actually has a lot of settings .
We won’t bring all the settings today
, but only talk about a few more important
ones. To call out the
interface of this setting,
we input the slash in the input field
, and then it will pop up the menu
. At this time, we have to go to Settings.
If you can’t see the Settings,
you can type after the slash. Press an S
and the Settings should appear.
Then we click on this item
to call it into our input field and
press Enter to send it out
. Then it will pop up this form
. This form
is actually the setting interface of Midjourney
. First of all Let’s take a look at the first two rows.
These two rows are in the creative mode
. You can only choose one of them at a time
. The ones we use most often are
MJ version 4 and MJ version 5.
Their abbreviations are V4 and V5 respectively
.
After choosing it, it is actually using different AI models to generate images
. The relatively new V5 model
is
a model that is commonly used by everyone.
It
is
best at making super-real photos
, and there are other things that are quite worthwhile. The thing mentioned is that
you don’t know. Before Midjourney released the V5 model last month, Midjourney was
actually not very good at drawing eyes and hands.
So after the release of V5, everyone
found that the ability to draw eyes and fingers of the V5 model has greatly improved. Progress
Although when I use it myself
, I think it sometimes draws a little weirdly
, but it is much better than before
. As for the V4 part,
it can’t make very realistic photos , but V4 is very suitable
It is used to create a sense of atmosphere . It has a little more personality than V5.
Here I would like to add a little bit about the working principle
of this
setting
. It is that the things we adjusted
in the setting interface just now
will be automatically added to the prompt you sent.
For example, When we pressed the MJ version 5
creation
mode
just now, when we sent the prompt
, the robot will automatically add
this --v 5 after our prompt
, or we can see that the robot will automatically add it for us in the setting interface.
What material is
the Suffix part that you see on the screen
. For example,
today I clicked the
button of MJ version 4 and the Style low below in the settings. After
pressing it,
these two options will appear automatically. After
that , when I send
When prompted, it will be automatically
presented in this way after my prompt, that is,
it will be automatically added like this
, and then we will mention other commonly used modes in the next row.
For example, the recently popular Niji mode
, Niji is It is specially used to draw anime.
For example , let’s say we switch to Niji version 5 now
, and then
we
send it out with exactly the same keyword as Disney’s Spanish Princess. You
will see if there is an ending here
, and it will automatically add a --niji 5 for us.
Then it will become Japanese anime style.
This effect
is very good, so the operation mode has a great influence.
That is, when you are creating the same theme, you
can also switch to different modes to try
it out. The results will be quite different.
Except for the creation In addition to the mode,
let’s
take a look. The other setting below is called the Style value.
This
value is a bit like controlling the degree of divergence of AI creation
. Then, if the Style value is lower,
it will
generate images
more conservatively according to the prompt you gave. Let’s take a look at an example.
For example,
let’s use the prompt of Disney’s Spanish Princess
just now to test for high Style values.
The difference between low and low
is that if the Style value on the right is increased
, it will not stick to the Disney characters, and some different styles
will start to appear
, even if it has almost completely escaped from the Disney style
. It seems that the background is also here.
It is the part of its free play
, because we
did not mention its background at all in the original Prompt.
What
kind of background is it , so
it can be freely played on the side with a high Style value,
like the part with a low Style value on the left
You can see that there is almost nothing in the background, but
it is
relatively simple
, and the characters are also relatively close
to
the setting of Disney Pixar we gave it. Now,
the basic application teaching of Midjourney has almost come to an end , and you may still have one. The big question
is that I don’t know where to start when writing prompts
. For example, if you want to make a logo
style, how to grasp it
or what words can be used
, so here we will mention
another
very powerful tool.
It’s ChatGPT,
so if you haven’t watched
the video we introduced ChatGPT in the last episode,
you can click the link in the upper left corner
to learn how ChatGPT works , so
when we don’t know which prompts to write,
we will You can turn to ChatGPT
to help us write some high-quality prompts.
It’s role-playing. Remember
, so next we will use two steps
to combine ChatGPT and Midjourney
. The first step is to let ChatGPT
write Midjourney’s prompts
and then Let's throw these prompts
into ChatGPT to generate pictures
. How does this actually work?
First, we open the ChatGPT page
. Now we are going to make a Midjourney
Prompt Generator (producer).
Let's open a new chat room
and give He appointed role-playing, and
this role
is to help us make
the Prompt that can be used by Midjourney.
You can adjust the script of this role-playing freely.
It doesn’t have to be like my script
. The picture here
is what I gave to the ChatGPT chat room
What is the first command here
? Let’s break down its structure
. Because ChatGPT
does not know how Midjourney works.
They are two completely separate and independent AI robots
, so here I am To explain it
, I am using an image-generated AI
called Midjourney
, and then I want you to act as a machine
that generates MidjourneyPrompt , and then the following sentence,
I will start to explain my rules
, that is, if I put a slash
in front of a topic today
At that time, you have to
help me generate a prompt that fits this theme
according to different situations
Then I wrote an example
, assuming that I input a slash and then a picture of a shoe product
. The following demonstration is that I wrote
a large list of prompts that can be used by Midjourney
, and the structure is the same as ours just now.
It is the theme and then add some atmosphere words. Then my style is Product Photography.
Finally, we can also add the camera model
or the way of shooting
, but this list
is just an example for ChatGPT.
Then the next ChatGPT
replied that he is ready
, so from now on Our
creative process can use ChatGPT to generate text prompts
and then paste them on the Midjourney page to generate
pictures. The examples you see next
, whether it is web design or realistic photos,
are all produced using this generator. Prompt
to do
it, then
I can start to use slashes to add a simple theme
and let it play the rest.
For example, I say (Chinese can be used)
I want to make a logo for a new company
and then ask for Apple
Then it gave me five sets of prompts
. If you find that it only
gives you one or three sets of prompts when you use it,
you can ask it yourself that
you want five sets of prompts
. Then we can start to choose the combination
we like more.
For example, if I like the first set of Prompt,
I will copy it
and paste it into Midjourney
to generate a picture
. After sending it out,
you can get something like the above screen.
Have you noticed
the logo of the child
? It is obvious that Midjourney can’t grasp the concept that Apple is a company.
So at this time
, we have to give feedback to ChatGPT
and tell ChatGPT that the picture
I
want is not a real apple. Next, ChatGPT apologized to me
, and then another five groups
of prompts were generated below that did not mention the word Apple.
At this time,
I will copy it and paste it into Midjourney
, and the real apple will not appear. The
resulting Logo is This is more normal
, so here is the part of the logo that is completed
. These are all made in V4 mode. Sometimes,
it
doesn’t matter which mode is generated
, it must look better
, so you can try different
styles. Have you found out here
that you have any problems
, you have to tell ChatGPT directly
, just like I just encountered
the problem that Midjourney can’t recognize the word Apple, so
I can directly complain to ChatGPT, so
let’s do the next example Let’s take a look at the webpage design , which is
the pictures generated by Midjourney
, which can be used as templates for designing webpages
, or as art settings, etc.
In the same way, we can first
talk about our theme with ChatGPT
and then let it provide some available prompts
, then continue. In this ChatGPT chat room
where the characters have been set just now, we will start with a slash first and
then add our theme. The theme, if you are used to it, you can
also input it in Chinese
and then ask him to output it in English.
It just
depends on how you compare
Get used to
the theme of this web design
. After I actually tested it myself,
I found that
if you enter
the word website in Midjourney’s Prompt,
there is a 99% chance that the generated picture
will show a real device
, which is like the lower left corner of the screen. It’s the same
as the picture of a computer or a mobile phone
, but we don’t need a picture of the device today, all
we want is a webpage
. Then I found out later that the word web page
must be used so that the picture of the real device will not appear
. So I am here
When talking about my
needs with ChatGPT,
I first specified that I want to use the two words web page to call the website,
then ChatGPT
will give me three sets of prompts,
and I can throw them all into Midjourney
to try and see . I can also switch between V4 and V5
to see which one is better,
and you
can choose the typesetting
or style you like more. Suppose I like the one in the lower right corner of V5
. I can press U4 to get a big picture.
Then this
By the way, the sketch of the solar system web page is complete
. Suppose we want to use this website sketch again today
to make a mobile app version of the picture
. It is the same style
, but it is the typesetting on the mobile phone
. How to let Midjourney refer to it
has already been done.
For this ready-made image,
here we will mention another way to use Midjourney
, which is
to add a JPG or PNG image file URL in front of your prompt,
so that Midjourney can refer to that image
and then generate similar images . So
here We
need two materials. The first is the URL of the web page sketch
, and the second is the prompt for the app version of the web page.
So first of all,
we
ask it to generate the prompt for the app design
on ChatGPT .
In fact, I just posted it directly on the question side.
I uploaded the original web version of the prompt to give him a reference
, and asked him to design a
prompt that uses similar characters but is an app version,
and then come to the second step.
We need to get the picture URL of the web page sketch
. Here , click
on the one that we have just generated.
Then in the lower left corner,
there is an open in browser.
After clicking in,
when the picture expands to the maximum state
, you can click its URL
and copy it
. Here, you must
also pay attention to the end.
If it
ends in png or jpg, after collecting these two materials
, we will go back to Midjourney to generate the picture
. In the field of prompt
, we will first paste the URL of the reference picture
, then press a space , and then paste the app version. The prompt
is copied from ChatGPT like this,
and then
we package it up and send it out
. Here are the sketches of the two app versions that I think the final results are not bad
. They are the same as the original web version. The design language
or color tone are the same.
Those pictures
In fact, it can be used in the stage of design proposals. It
’s pretty easy to use
. Let’s move on to the next example.
We ’re going to generate realistic photos
. In fact,
the v5 mode will perform best in this part
, like on the screen .
The pictures and the pictures you will see next are all
generated using the v5 mode with ChatGPT,
so
at the beginning
, you can pre-book and tell what kind of photos
you want to take
, for example, street photography
or movie stills You can talk to ChatGPT.
For example,
for the first one , let’s make a movie still with an artistic conception.
First
, ask the ChatGPT character
to feature an
old lady
, and then add some keywords. Adding oil and vinegar
is to make this artistic conception more complete
. Now it is for the sake of cleanliness of the picture
. Otherwise, there will be
two sets of different prompts
below
. In the end, we will choose a set of prompts that we are satisfied with
and throw them into Midjourney.
The four pictures on the left will be generated , which are similar in style
, and if you look closely, they all seem to be the same person
, is n’t it very cinematic ? Then we
can
upgrade the one we like to the big picture on the right
Now suppose our characters and scenes remain unchanged
and directly become street photography
. In the same way, we first use ChatGPT to generate a prompt.
Pay attention here.
I ask him to freely use the camera model and photographer
. The photographer
will refer to his style
and generate similar photography. Style That
’s not to say that this picture was really taken by that photographer
. This is
the prompt that ChatGPT gave birth to freely.
Because each set of prompts
has a combination of different camera models and photographers
, it ’s pasted in Midjourney.
There will be relatively large differences in the results and
the composition of the pictures is also different
. Using this method of not specifying the camera and photographer ,
you can have a lot of ways to play
, not only the camera and photographer, but also
other types of pictures
. You can also use ChatGPT to play freely
For example, film directors, illustrators
, or cartoonists, etc.
We just took a solo photo. Now
let’s try to take
a street photo of a group of Taiwanese teenagers
. Here we also invite ChatGPT
to freely use the camera and photographer
, and then I will Pick a group
and post it to Midjourney
to generate pictures,
then it generates the group of pictures on the left
, which is quite interesting, but
I don’t know why the clothes I wear
just have a sense of age.
So I went back to ChatGPT
and asked it to adjust their clothing,
that is, to ask them to wear some more fashionable clothing, so
that it became the picture on the right,
and what everyone saw was more modern. The effect is also
quite good, but there is a small disadvantage,
that is,
the sight of his characters in the v5 mode
sometimes still drifts randomly.
For
example, when we see the picture on the right
, the boy on the far left just
doesn’t know where Where to look and
look carefully
In fact, the girl on the second from the right is not looking at the camera
, so
now Midjourney
has such a small shortcoming, let me tell you first , after watching the characters,
let’s
take a look at the natural landscape photos
, you can specify the location yourself
Or just tell him which side of the natural
landscape you want to see. This part
is a little regretful. The resolution of the
pictures generated by Midjourney v5 is not very good
, but if you need it, you
can actually use it through a third party. app
to enhance its resolution.
As for the landscape part
, we first let ChatGPT
retouch our prompts
and then ask it to generate a landscape photo of Taiwan with mountains and seas
. In Midjourney
, we used two models here.
Let’s generate the same prompts
. Let’s see which side you like better
.
Although the v4 side is very beautiful
, if you look closely, you will feel that it has a little fake feeling
. It seems that the filter cover is too heavy. Is it just a repair?
The picture has been edited too much
and it is quite suitable for the feeling of using it as a tablecloth
. Then the v5 on the right
is really like a real photo, as
if it was taken in a secret place in Taiwan
. The last two pictures are both V5 (slip of the tongue)
is that if these two pictures are put on the Internet
and you don’t tell
me, I will really think that he was taken by a human being . Finally, let’s try to generate a natural landscape without specifying a location.
That is,
I directly use a very
A general word, tell
him that I want a natural landscape like
this. Here you can see that these four pictures
are the results of v4
, which is a bit more dramatic,
but the effect is really full marks
, that is, each picture has a story. I feel
that the next four pictures
are made with the v5 model
, and they are really similar to the photos taken
. Of course, the prompts used in these four pictures are different
, so
using the combination of ChatGPT and Midjourney,
there is
really room for development. It's very big , I'm afraid we don't have any creativity
. So here
we
have almost finished the explanation of
the main usage of Midjourney (wait for more content).
Let's review
some of the content of Midjourney we just arrived.
To use Midjourney to generate pictures
, we need to use text The way of input is to
let him convert the prompts we provide into pictures
. The prompts
are a bit like the ingredients for our cooking. They can be divided into three types of content
that can be input
. The first is theme and background,
style and medium
, and the third is color. And the atmosphere
, these materials are not necessarily as many as possible
, but everyone has to keep trying to find out how to combine them
to suit your appetite
. Some people will also advocate fewer prompts
to give Midjourney a greater space to play
, and some people prefer more. Prompts
gives this picture more details
, so the vegetables and radishes have their own preferences. Then the next one is the setting interface of Midjourney
we mentioned just now
. The most important thing is the switching of the creative mode.
The style he is good at in each mode
is also It’s not the same.
It depends on
whether you want
, for example, a super-realistic style
or an anime style
. You can switch between different modes to see
which one has the best effect
. Finally, we also mentioned that it can be used with ChatGPT
to make it fast. Retouch your prompts
and then you can also communicate with
it
what is the final result you want to present in Midjourney
, then this ChatGPT
can speed up the whole process of image generation
, then we will talk about the paid part
, let’s look at it first Take a look at its monthly payment plan
Now you can see that there are mainly three plans on the screen
. The first one is 10 US dollars, which
is about NT$300. This is a monthly price
. After you subscribe
, you can use about 200 pictures a month. Here we also need to talk about
the way Midjourney calculates the quota, which is
actually the time you use to calculate the pictures
, so it doesn’t have to be exactly 200 pictures.
For example , if your pictures today are some relatively simple
pictures with relatively small files.
You can generate more than 200 pictures
, so I suggest that everyone
can start playing with this 10 dollars first
, and the second is that it is commercially available
, that is, Midjourney, except for the free trial,
all pictures can be used For
commercial use, let’s see the $30 a month part
. Its biggest feature is that it does not limit how many pictures
you can generate a month
, but it has a proviso
, which is the 15-hour fast generation time
, which is Midjourney. There is a very special mechanism
called Fast hours and Relax hours
. Fast hour is of course faster to produce the pictures you need.
When you produce more than 15 hours of generation time, it
will become Relax hour
, which will become slower to generate. The picture you want looks like this
, but 15 hours is actually quite a lot
, and if the 15 hours are not enough, you
can also buy more like this
. The last Pro plan is 60 dollars .
The biggest difference lies in
this plan. There is an incognito function
, that is, you can set the pictures you generate to be private
, so that only you can see your own works , because
didn’t we say that
Midjourney’s final works will be placed in the official gallery, as
long as you have your user ID? In fact,
you can find your work , but
using this incognito mode , you can set your work to be private
. For example, if you generate relatively confidential content
, you may need such a function.
For example,
if you are in a game creation company
If you use Midjourney , you must use this function
. In addition, the benefits of other Pro plans
, for example, its Fast Hour has more
, and there is a 30-hour quota
. If it is not enough, you can also purchase it
. The other one I didn’t mention just now
That is, the first two schemes can run up to three tasks at the same time, that
is, generate three pictures
at the same time. The Pro scheme can run up to 12 groups of tasks at the same time. So
in general, the biggest difference of Pro
is that it can be set to be private
and Fast hour can be changed.
There are more than 12 groups of tasks performed at the same time
. Next, let’s look at the part of the annual payment plan
. It is basically
a 20% discount for each plan,
but you have to pay for the whole year at once
. Then come again We talked about the part of private space
, so how
do you create your own private creation space on Midjourney?
The first step is to create a server
. There is a plus sign on the left side,
which is to create your own server.
Then here, Join a server,
you can create a server
. After you create it
, let’s go back to Midjourney’s server first , then
click on the Member list in the upper right corner, find Midjourney Bot,
and then there
’s Add to Server here.
On the server side,
we just
choose the server we just created
. After adding it
, we can directly
start generating the pictures we want in the input field of our channel,
and then come to the second way.
That is, we can directly
find
this Midjourney Bot and
start talking with it directly in the following Message,
and it will become a private message in DM
. In the private message, you
can directly start generating pictures
. This is the second type Method
Next, we will mention another
twin of Midjourney called niji journey
. This niji journey
is actually an AI robot specializing in animation. You can type any text on it
and
it will help you generate
Japanese anime-style pictures
. And the most important thing is It accepts Chinese input,
as you can see on the screen here
, I just used the prompts of the
Spanish Disney princess
to input it ,
and these four pictures will appear
, and the effect is very good
, it is really like Disney Pixar
Let ’s introduce
the difference between it and Midjourney
, that is,
after you type niji journey on the Internet, you will enter this website
, and then you scroll down, and
he will ask you to join another Discord community
. The way to join is exactly the same.
It will enter the server
, and you can start to generate pictures here , and then
you can also directly
talk to this niji journey Bot.
The method of use is almost exactly the same
, only the largest The difference is that it can be used in Chinese
, and the pictures it produces are also Japanese anime style
. Finally, let’s talk about the alternative software for free trial,
that is, if you are really reluctant
to spend the 10 dollars today to play and watch
Midjourney
You can also try it from
other alternative third-party websites. The usage of these two websites can be said to be
similar to Midjourney , that is,
they will have a prompt field
, and you can enter the keyword topic
you want. Then
come the third alternative software
, Bing Image Creator
, you can go to the Bing website
, you can find
this Image Creator,
and the same thing is,
after entering the prompt, it will help you generate pictures
, but after I actually tested it myself
Well, I think Bing’s generation speed is really a bit slow
. Well, the above is our
complete Midjourney teaching today.
I hope you all have fun
and this software is addictive
, because I really think the pictures it generates
are true. It’s so beautiful
, and each one is so beautiful
, and I want to try it in every field.
In fact, I think it’s more fun to use ChatGPT
. If there are any
updates related to Midjourney or useful AI tools,
I will also Make a video and upload it to our channel
, so if you don’t want to miss it
, remember to subscribe to our channel
and press the little ding bell to receive notifications . I
will also post some latest news and life-related content on Instagram
if you are interested.
Go follow up. Well
, we're almost done with today's video
. See you in the next video
. Bye bye
Browse More Related Video
How To Generate Images With ChatGPT (Create AI Art with Chat GPT)
Fooocus全功能詳解,跟Midjourney一樣簡易還免費??
Write Prompts like THIS for Success in Midjourney V6
Screens & 2D Graphics: Crash Course Computer Science #23
ExcaliAI Enhanced: More Visual Thinking Power
Unreal Engine 5 RPG Tutorial Series - #2: Locomotion - Blendspace, Crouching and Procedural Leaning!
5.0 / 5 (0 votes)