Transcripts
1. Introduction: Hello and welcome to the
exciting world of me, Tony, My name is Philip states, and I will guide you
through this course. I have designed this course to provide you with the
knowledge and skills you need to create impressive
images with Mitch journey, this course will
help you to elevate your understanding of mature
early and prompt creation. We will start with the basics explaining terms like prompt
engineering and prompts. Then I will show you how to connect Mitchell and
with this quote and use it privately so that
you can start prompting. Later. I will show you how to
create your first prompts. And then we will dive
into prompt improvements so that you will be able to
create exactly those results. You have ambition in your hat, because only when
you can truly bring your ideas to life
with me, Tony, can you really start utilizing the tool and make the
most out of Machete? So I will show you and
you will learn how to incorporate the style of
famous artists into prompts. Use camera perspectives
and camera settings to copy the work of
famous photographers. And of course, it will
also learn how to evoke specific moods
with emotions. Of course, you can watch
this course passively, but that is likely not the most effective way to truly acquire new knowledge. So that's also the
reason why I'm including some practical
examples you can try to use and tried to apply. So that's acquiring
the knowledge you will learn in that
course becomes easier. This way, you should be
able to learn faster. You can also complete
this course at your own pace and revisit
the lessons when Amber, you have questions or
need clarification. By the way, feel free to
reach out to me at anytime. Well, that's my introduction. So let's not delay any further and dive right into
the first lesson. See you there.
2. What are prompts?: What is the prompt
image journey? Well, a prompt that's a short tax command
you are using as a user when working
with the AI tools like midtone or chat GBT, e.g. we have here a short text, and if we entered here
in the Discord server, the mutually bot
will interpret it and create an image out
of debt tax command. In this example,
we have a chaplain is cartography with
a Japanese sword. It's figurative simplicity
in the Yucca art. And here you see the results. You see what Mitch journey did with the prompt I have
entered previously. But how exactly does this work? How can we actually
interpret this text and generate images out
of that tax command? For doing so, the imaginary part is dividing the words and phrases in a prompt into smaller pieces that
are known as tokens. And these tokens are
then compared to the parts training data and then used to create the image. So median is trained
with some billions of pictures and images
as training stator. And then when you're entering
a prompt attacks commands, it tries to create
the most probable, the most likely tokens
for that command. If you wish to improve the quality of your
generated images, the first thing you
should do is to improve the quality
of your prompts. To add more details that can be interpreted
correctly by the midtone bots. So excluding words that are not necessary for the desired
results and include words that helps
the midtone part to understand what you
want to have better. E.g. you could start with including a certain
style or art style, like we did here with occur art, which is a certain
Japanese art style. We could also now add some colors you want
to have in the image. We could make it realistic
for photography, or an oil painting or an
impressionistic painting. You see there are thousands
of options and I will try to show you as many as
possible in this course. But of course, that's just, there will just be an example of all the possibilities you have. Your prompts, you
are telling me, Johnnie Walker, to generate. The more detailed your
prompts are dead, that mutually knows
what it should do. So the key is to be able
to prompt correctly. In the end, when you're
doing prompt engineering, you have to do a lot
of trial and error. You have to try a lot of different Prompt
Configurations to get to the best
possible results. So now you have entered your first prompt
and midtone will now start generating
for preview images. That's the so-called grid mode. Now you could either
say Mitch, Any, do another variation, or
you could say midtone. Please upscale one of
the four options here. You can do that by clicking
on one of the foyer. Use here on the bar, U1, U2, U3, U4. And you is standing for
upscaling of the image. If I wish to e.g. upscale the second image, this one here, I
have to click on U2. I've already done that. And here's the upscale result. In conclusion or as a summary, it's important for you
as a user to be really precise with your
instructions to Swift instructions you
are giving to maturity. The more precise they are, the better your results will be. Additionally to tax commands, they are also parameters you
can use the midtone, e.g. here we have two. V5 or S, 500 or 916 kills. Nine toes are all parameters that give additional
information to meet you. And if e.g. v5, that's standing for the fifth
version of midtone. But of course you could
also use older versions of material like V4 or free. That could make sense sometimes because all the versions are a little bit more
creative in some ways, ER, That's the aspect ratio. So this parameter is simply telling Mitchell and
as the name suggests, which aspect ratio to use. And KOs, here's key has nine. That's a parameter
that tells me Tony, how creative or how unusual
the results can be. The higher this number is, the more creative will be
the output by my journey. With parameters,
you can influence the quality to creativity,
the aspect ratio, and a lot of many
more things in which, in a nutshell, this is what this course will focus on most. So how to communicate more efficiently with
maturity and how to use parameters immature need to create even better results. This course will
teach you a lot, but of course in the end, you have to put all
the knowledge into practice as well to become
better at prompting. And yeah, you will not become a really good project engineer with just knowing that theory. No, you have to use missionary, you have to generate pictures, images on a daily basis to
improve your skill set. Of course, I know that
that's the issue. You should not do that alone
without any instruction. That that's why I've
created a course to make it easier for you to get into the prompt
engineering image journey to help you to have
an easiest art, my journey, and to not waste too much money in the
beginning because you are exhausting your subscription by generating images that are
not really convincing, that are not really that
what you wanted to create. I hope with this
course, I've created the best possible instruction
for newbies immaturity. So if you're ready, let's move to the
next video lesson.
3. What is prompt engineering?: What is primed engineering? Well, that term refers
to the art of creating precise and efficient
prompts for a i2 is like CBT or midtone. Good prompt engineer should
be able to create prompts that lead to reliable
results in which journey or other UI tools. That's a skill that is needed more and more
in today's world. You see you can easily create
stunning art immaturity, but it's more difficult to
want to create image journey. If you have a clear picture
in mind and then create that picture with
maturity, you know, the artist of the future might be somebody who is
willing a brush, but maybe somebody
who can translate creative ideas into prompts that result in the desired outputs. For most people, that might
sound like something out of a science fiction or Black Mirror episode or
something like that. But in reality, it's a chart that's already paying
money to a lot of people. Use it to do engineering
for clients, to create customized
images, image journey, and others upload the best prompts platforms like
prom flat or prompt base or even add c to earn a passive income by
selling prompts. And that's also one of the
reasons why I've created this course to help you
become better in prompting. Either because you need prompt, you need prompting
for your chart. Because you are a
designer or an artist or a content creator who
needs to create images or because you wish to find a way to earn passive
income in the future. And as it looks at a moment, earning a passive income if Mitt Romney is
definitely an option.
4. Midjourney Subscription - What are your options?: What are the costs of maturity? In which of the three available
plans should you choose? As you probably know, there is a basis plan, then there is a standard plan. At first, there's a protoplanet. In the following minutes,
I'm going to show you what is included in those plants and which of
them is best suited for you, as you knew, for newbies, Mitchell and it is free. So as a newbie, you can use
around 20 to 2053 pictures, but then you have to decide
which plan to choose. That's because 25
pictures are simply not enough to really
work with maturity. Another issue is that
a free plan that you cannot commercially use
the generated images. And of course, this
is always an issue when you want to sell them on all the platforms or
monetize seminar two ways. Those are all reasons why
you will sooner or later, probably sooner switch to one of the commercial
plans of maturity. And now the question simply is, which one is the best? As we see here, the
first option is the basic plan that costs
eight bucks per month. But of course, just if
we pay it annually. So it's eight times 12, which is if I calculate
it correctly now, $96 for testing me, Tony, I would recommend to
use the monthly billing. So in that case, it's
a little bit more expansive, $2 more expensive. So instead of eight box, you will pay ten
bucks per month. But what's included
in this plan? First? It includes around
200 pictures per month, which is around ten pictures per dollar you are
paying to this tool, of course is not a lot. When you're using maturity, you probably will generate
a lot of pictures. Most of them, most likely, most of them we will not use for commercial purposes
or for any purposes. At least this time gives
you the ability to use Search Android images
for commercial purposes. You can access the
mammal gallery and your pictures will
be displayed there. You can even buy more credits and you can have free prompts, a free shops at the same time. So you can generate free time. You can generate free
images at the same time, which will speed up your work with me turn insignificantly. That's the most basic plan
for commercial users. And honestly, I believe
that for most users, it's not enough, it's
not included enough. I think for most,
the standard plan would be a better choice. This is $130 per month plan. If you pay monthly, the annual billing is less, it's $24 per month. I think it's a really
good choice for you. Let me show you here. You have 15 h of
first-generation. What does this mean? What fast generation means that when you enter into prompt, the images will be generated
between some seconds. In fact, with this gland is no limit for the number of
images you could generate. Because after using all your
first-generation hours, you can still generate images. With relaxed generation. The only problem now is that it takes a lot of time
to generate images. The time, of course, depends on how many people use the
tool at the same time. So it can easily take several minutes before you
can generate a single image. And of course this way, working with maturity
is a bit tedious. Still in theory, you
could almost create an infinite amount of images
similar to the basic plan. In a standard plan, you can have free image charts
at the same time. So you can create lot
of images very quickly. But the biggest
difference is that you have an unlimited
amount of images. You could generate a
theory and you have about 15 h of fast generation. Really believe that for
most users of midtone, this standard plan is
the very best option you can choose. And the 15 h of first-generation should be
enough for most purposes. Of course, if you
need more fast hours, you can still buy more
credits for that. But of course, Danny depends
how many more hours you need because you could also
change to the next plan. The problem that is causing $48 per month if you're
paying annually. If you're getting
method built, Dan, it's even 60 bucks per month, That's a bit expensive. Anything it would make
sense if you're really using the tool for
commercial purposes. And if you're really a
power user of the tool and needing that many
hours of fast generation, you have twice as many hours of first-generation available, but you're also
paying twice as much. So it's actually not the
best deal you could get. However, if you're such
a midtone power user, it's still the best
deal you can get. And for some users, 30 h might not be enough. It would even need more hours, but there's no other plans. So in that case, they would need to buy my hours more first-generation
hours manually. So what do I recommend? Well, as a newbie, I think the best
thing you can do is you have taken a basic plan, paid ten bucks a
month for one month, and then play around with the tool will
generate 200 images. Like this tool after tour. After 200 mm, you
should know dads, then you can upgrade
to the standard plan. Don't stick to the basic plan. I think it's not enough
for your purposes, but yeah, I think for you probably
the center plan is enough. If you're really power-user
using mature and as a team, then you can of course, think about upgrading
to the Pro plan.
5. Why are good AI prompts important?: Before we continue
with this course, let's ask the question, why is it actually
important to write good prompts for tools like
Mitch journey or chat GBT. Well, using tools
like meta-learning is not for free as we've
learned previously. If you're new user
in my journey, you have 23 images and then you have to upgrade
to a paid account. In the basic plan, you pay about around
ten bucks per month. So that means per back you can
generate about ten images. So that's still cheaper
than most stock pictures. For what you get. It's not a lot of money. But of course, the more
efficient you use your prompts, the more money you can save. And there's always
a component that is maybe more important than money. It's time because generating
images takes some time. And if you need to iterate
your prompt several times, you quickly need
one or more hours to generate a single picture. And of course, you want to
streamline that process. You want to become more
efficient and be able to generate images
within some minutes. And in that case, it really makes sense for you
to use mature need not go to the next
stock image platform. In a nutshell, becoming
more efficient in prompting immaturity
will save you some money, but it will of course,
save you a lot of time. And that is the main
reason why you should learn that skill and why you should become a good prompter. Because it's also a skill that will be needed more and
more in the future. As I stated previously, the artists of the future
is not somebody who is willing a brush for painting pictures of doing
something like that. And then he is somebody who, she is somebody who is able to communicate with
AI tools efficiently.
6. Connectin Midjourney with Discord: How can we connect
Mitchell and with Discord and start
creating images? If you have already connected your midterm and
account with this code, you don't need to watch
this video and you can, of course, chapter
the next video. But if you haven't done it yet, I will show you how to do
it in the next minutes. First, if you
haven't yet created a Discord account with now because you will need it later. On the missionary website, we now click on China better. That will connect Mitch
journey with discord. So now we have to
click and enter. So sorry that is now
the chairman version. Just follow the
steps, the same steps and the English version
and you should be fine. Now, they will verify if we are really a
human, not robots. So that's the most
complicated part. I hope he believes
me. Let's see. And yeah, it looks as
if that was successful. And you see the sailing
boat in the left sidebar, that's the journey that
meet Chinese server. We are connected
and you already see the first images generated
by Michel new users. If you have never used
this quote before, that might look very,
very confusing. So a little explanation. So here on the left sidebar, you have different chat rooms
or rooms. You can enter. In the middle you see the feet, whether user-generated
content is shown, but also information by
the midtone admin team. Here, e.g. are the rules you should follow when
using Mitch journey. I recommend reading it. I don't think it makes sense now to go through them one-by-one. The most should be
self explanatory. So you shouldn't be
rude to other people or to the admins. And you should not create new images or bloody
violent images. Getting started
here in this room, you have a short guide
that is showing you how to start generating
images in mutually. But if you do want
to read the guide, you can simply continue
watching this video. Because I'm giving you a tutorial important
for beginners. If you're a newbie, you should start using
the new channels, like new in 29, 59, et cetera. Because therefore new
users of maturity. And yet they like sandbox where you can
try out the features, relate to have a look at
a more advanced channels. Let's click here on
the first new channel. And you'll see it's
like a chat feed where users are generating
images all the time. Chances are that
you are surprised, like I was when I first
started using midtone, because I was surprised
that it's actually, everything's in public and
I am generating images and everybody could see the images and download the
images I'm creating. But before you are now stopping using mature and due
to privacy issues, let me tell you one thing. You can later when you have a subscription and paying when
you're paying for Mitchel, you can late inviting me 20 parts to a private
discord channel. And that channel you can, yeah, create images in privacy. But as a beginner who is
using Michal Ronnie in the free version where you have about 20 images you
can generate for free. Yeah, use the new BE
channel and use it as a sandbox where you just
learn the basics of maturity. And if you'd like to tool, then upgrade to one of the paid subscriptions and then you can start
creating images. So how do you actually start trading pictures,
immaturity, easy. You just go to the chatbox
and now you enter a slash. And now you enter, imagine. That's command here. You can select it from the list, or you can also of course, type it into the chatbox. And now there's a black
box opening with prompt. And here you are going to input the prompt
you want to create. If I'm now entering, Enter and what's happening? What's happening? Midtone is giving me an
error message and why? Well, easy because I've used all my 23 images
in that account. Of course now I will
switch to my paid account. And that account over show you how to create images
in mitch journey. This account was just now
used for showing you how to connect my journey with Discord
for the very beginning. I think the only
important thing for you to note is that we've slashed. Imagine you can start prompting. So, yeah, feel free
to use that command and try out several
prompts on your own. Just be aware that you only have about 20 to 25 prompts
available in your free account, then you have to switch
to a paid account. If you want to
continue this course. Of course I recommend to upgrade to one of
the paid accounts. You can do it on a monthly basis so that you don't need to pay more than $100 at once and would
recommend first testing. It's Mitch journey of course,
with the free account, then next stage with one of the paid accounts on a monthly basis and
then later when you're sure about the tool and if you really liked the tool and
get used to using the tool, of course, Dan, you can switch
to an annual subscription, which is a bit cheaper than,
than monthly subscription. How do you upgrade to a paid subscription
in meta-learning? Well, after we got
the error message, the imaginary part, send
us a message here you see the left sidebar,
maturity bonds. And if we click on that symbol, we see the message is, it's set to us here. Yeah, you're free. Trial
has come to an end. Please subscribe. Or we can do now is type in
slash, subscribe and Enter. And now we can click
on the page here. And that will lead us to
the subscription page on the website. And here we see the plants
that are available to us. And we will have a deeper
look into the subscriptions, into the options you have now, in the next video.
7. Midjourney privacy and connecting a private discord server: When you're starting to
generate images in my journey, you will likely do that in one of the public
channels of midtone. So what are we seeing here? Well, all the
images generated by all the users that are
active in this channel. This is of course very confusing and annoying
because you're always in this feed seeing new images generated
every second, and it's difficult to
find your own creations. The first thing you
should do is creating your own private
server on Discord. Doing this, it's very easy. You click on Add a server, then on create my own. Next for me and my friends. And now we give it a name, let's say mid
journey test server. And now we'll click on Create. You could also add
an image here. We don't do it now. And
we have our server. If you want to
search for a server, you see in this left sidebar, we have the list of all
the service you are active in or you have created. Good. This was the first step. We have our test server. But at the moment we cannot
create images for mature on here because we need first to invite
the mutually bots. How can we do that? Well, we have to go back
to the midtone channel. And here we are searching
for a mutually part. So we are already
one of the channels. That's important. You have to go to the channels
and then in the user list, on the right side, you see all users and there
is the Mitchell need bot. So let's click on the
image on the bottom or weight maybe 1 s.
For clarification, you will have to be in one
of the public channels. So one of the
channels from China, 1234 till January 20, user list of those channels. You see the Mitanni, but
you add it to a server. Then you select the server. Here, the maturity test
server you've created before. We confirm that can be checked. Now, we authorize it. Now. We need to confirm that we
are humans and not robots. Today that's not so easy. And capture. What do we need to do? We need to select bicycles. Let's see if I can
do that correctly. Check it. Yeah, we are confirmed
to be humans. Now, we can return to
the mutually tests over. And we see here the
mutually part wasn't. Why did successfully. Let's test this with a prompt. Imagine prompt. I want to have a picture of
a midtone, but let's see. Yeah, he's starting to
work and that's it. So we have our private
midtone discord channel and we're the only user. But of course you could invite other users as well if you want. But there's still
one privacy issue, because when the midtone Bart has finished
generating this image, the image will be shared to the public image
Journey gallery. Of course, there's
some cases where you don't want this to happen. You could go to the settings. So let's enter settings. And we entered, we have
two oxygen stealth mode. In this melt, the images
are not shared publicly. Let's do that. Let's
activate itself mode. Oh, what's happening? Stealth mode is not included in my plan. Why is that the case? Well, I have the standards plan, but not the Pro plan, and I would need to upgrade
to the sixth dollar per month program to
have the self mode. We see midtone privacy
is a matter of money. Is it worth it? Well, it depends on you. But if you don't want
to pay the money, they're still an
option you have. Because you could
delete images again. You do that by
adding a reaction. And you click on x or you
type in the X in the bar. And when you do that, the image is deleted
immediately. If you want to keep the image, you can always download it to your computer before you
delete it from my journey. This way, you can make
sure that you're the only one with access
to the image. Because when you delete
the image from this code, it will be deleted from
the gallery as well. This is one option for privacy. My journey without paying
60 bucks per month. I hope this was helpful for you. And the biggest takeaway
of this video is how to use me turning in your
private discord server. That is one thing you should do immediately if you
haven't done it yet. Because using Mitch journey in one of the public channels
is really annoying. And it's not fun using
micelle it this way. So if you haven't done
it yet, do it now.
8. Your first Midjourney prompt: How do we start
using Mitch journey? As you've seen when I've showed you how to connect
Meacham if discord, you can prompt something in
disco by slash, imagine. And then you input
your prompt into that black box here, e.g. what do we want to see? Let's say you want to generate Channel National Geographic
picture of a lion. So now we have to wait several
seconds till mid tone is generating for preview pictures that is called the grid mode. So now we are 93%, as you can see here in
the upper right corner. And normally it takes
about 1 min till the bot generates the four
pictures in the grid mode. Keep in mind, that
is the fast mode. So if you have used all the your fast hours
and meet journey, you can, in most plants still
use the relaxed mode, but that takes even longer. What do we see here? We have four pictures of a
lion, a male lion. And they all look
very realistic. What you keep in mind, what you should keep in mind here is that that's a very short prompt. And the more specific
the prompt is, the more customized
the resulting images will be the shore to the
prompt and more generic, the images will be. The topic of the
course, of course is to teach you how to customize your prompt so that
you can really generate the images
you want to generate. So in this grid mode now we have four images. And let's see. You want to see more
variations of the image where you have the nice
gray weather background and the lion is
walking towards us. We want to have more variations. What can we do simply? We simply click on
v4 because it's 123440 picture in
that grid mode. So let's click here. And now we're in
a new grid mode. We have now four variations
of that fourth picture. And we could continue that. In theory, we could
click on e.g. the first image. Of course the second as well. We could click on the third image or even
on the fourth image. But let's assume that
you actually liked the second image a lot and
you want to upskill it. How can we do it?
We click on u2. So the use, of course,
standing for upscaling. Now mature and is
upscaling the image. That's actually really
fast in the fifth version. And voila, we have the
apps, Kate learned picture. And if you want, you can now save it
by right-clicking. But of course it will also automatically saved in
your midtone gallery. There's also a privacy
mode where that image is not automatically transferred
to your image gallery. That how to set up that. That's something I
will show you later. And that's how you can generate
images, image journey. Of course, that's just
a basic introduction. There are thousands of
other things you can do to customize the
image even more. And that's of course topic for the next videos of
this online course.
9. The most important Midjourney commands: In the previous video, I've shown you how
to create prompts by inputting slash and imagine. But of course,
there are also tons of other commands
you could input. In this video, we will have
a look at those commands. So let's start with
the first one here. So the first ones that are the settings,
Let's click here. Let's enter. And now, what's happening? Now? You see tons of options, e.g. the Mitchell new
versions, 1,234.5, we see that a fifth version
is highlighted in green. So that indicates that
we're using it right now. But of course, you
could switch to older versions out and get maybe more creative results in January 5 version generates
the most realistic results, but some say that e.g. a. Third version is better
for creative results. Switching to an older
version can make sense if you're not trying to generate
photorealistic images. We will have a look at that
in one of the later videos, where I will compare the same prompt in
different versions. Besides the five versions, you see there are
some other versions of maturing as well, e.g. you have Nietzsche for,
and Nietzsche five. Those melts or interesting
for you if you want to try and read animate based images, we will have a look at that in one of the more advanced videos. There's also empty tests
and empty test photo, that mode for more
photo-realistic images. But now with the fifth version, I don't think that those Mozah actually really
important for you. In the third row,
we see that we have the base quality selected. Of course you could
change that as well. E.g. to have high-quality that could make sense if you
want to create prints. In most cases, the base
quality is sufficient. But of course sometimes you want to switch to high-quality. Remember in that case, you're also using more
of the fast hours. So producing images always in high-quality can
become Catholic quickly. Of course, you can also
switch to low-quality. In that case, you are
actually not needing so much time for
generating images. And that's a more efficient mode to generate a lot of images. For beginners, I think
it's a really good mode for testing images. If you want to play
around with the tool and not planning to
publish the results, that's probably the
best mode you can use because you're not really using the fast creation time of your subscription too much
style here you have low, medium, high, and very high. And that says we're indicates how creative your prompt will be interpreted by a mature cell with playing around
with that option, you can heavily influence, of course, the outcome
of your prompt. Of course, with some images that can make sense
if I does not, so much of which option
you use that's up to you. For most cases I
would say math or medium is the best option
and it will last row, we see that we are
in the public mode. That means that our
pictures or Publix, everything we generate
will be transferred to the proper Lake
Mead Shawnee gallery. Of course, we could switch
to the stealth mode. In that case, the pictures or images will not
be transferred, will lead to show
you how to use that. It can make sense if you e.g. want to monetize your images or your creations and sell them
on Etsy or prompt pays. You see that we have
activated remix mode. Or what does it mean? Well, if you remix an
image in that case, so if mode is activated, mutually will ask if you want to change to
prompt personally, I think it's a useful
when you're trying to create a better prompt and playing around with
different commands. Because in that case
it will always do some small minor changes to
the prompt before remixing. But it's not something you need. Have four start
prompting. Fast mode. You already know that
means that images will be generated with him
in most cases within 1 min, sometimes less. And that's the standard
mode of course. But if you want to save
some fast mode hours, you can of course, which toilets low relaxed mode, because in most
subscriptions you have unlimited relaxed
generation time. But of course, the
downside is that in relaxed generation time, creating images
takes a lot of time. It can really a nuisance. So I think in most cases, you will rather buy more credits than using the relaxed mode, too much data, the settings, that's how you are
adjusting the settings. But later, of course, I
will also show you how you can use parameters
within prompts. Because if those parameters, you can adjust or set the settings for a
specific prompt, e.g. you can say that a specific
prompt will be generated in the higher-quality or
in a different mode. So let's have a look
at the other options. So commands, you
can use your blend. What's planned? Well, we've planned, you can blend up to four
different images. E.g. we can, I could upload an image of me and the
second image of a landscape. And then if a blend both images, I should appear in
that landscape. That's of course also
something I will show you later in an advanced video. What else do we have? E.g. we have to ask, That's the kind of FAQ version. But to be honest, searching on Google always work better for me than asking something inside my journey. Then we also have info. Here you see some
information about your account, e.g.
watch subscription. I am having the setup secretion, the chop mode, which is fast, visibility mode public the
first time remaining or no, I have a lot of time
remaining because it just updated my lifetime
usage and relax usage. You'll see I've never used
relaxed usage so far. Well, that way the most important commands you
will use immaturity. So let's move to
the next lesson. Let's start generating
some images.
10. How can you improve your prompts?: So let's start prompting. I have now inserted the
prompt lion into midtone, and it now has generated the four pictures
in the grid mode. When our prompt is really short, like lie on Mitchell need actually doesn't
know what we really want. It could be a
watercolor picture. It could be an
realistic picture of it even could display now
a female lion cell, that means short prompts
normally result in for diverse options
in the grid mode, you start customizing
your prompts by adding more details. And that's now
something we will do. I've already prepared
some examples. So let's have a look. And let's start here. Let's assume that I want to have a photo-realistic
picture of alone. In that case, it helps to add a National Geographic
picture of alone. Because then you get that
high-quality picture there is often displayed in magazines like
National Geographic. We see Mitchell did a good job. We have four really realistic
looking pictures of loans. And yeah, but in some cases, you actually want to
do something else. E.g. you would like to have, like here, an origami
scotch off a lion. Of course, there are
endless options, are the options that
I've used, e.g. where a chocolate scotch of VLAN or a marble sculpture of line. You could also use the
skies of artists like here, we've used a Picasso painting
of a lounge in the prompt. And it generated images in
the style of Newcastle. Maybe another odd expert enough, so I'm not qualified
to see if that would be a convincing
the castle. So I'm happy to hear your
feedback if you're an expert. What's interesting
is that we even have the signature of the
castle here in the prompt. So at least with the
original signature. And now we are. Another
thing I did was having a Leonardo da Vinci
painting of a lion, which is looking like a
sketch of Leonardo da Vinci. For this example, I want to stay with the
National Geographic prompt and further develop
their realistic picture. And let's add some location. Let's put our loan
into a wide West Town. And you see how much Shawnee
interpreted that command. Of course, we can use any
location, e.g. the mass. But in this case, I don't think it looks so good. Maybe we should need we would need to
refine the problem a little bit to make it
blend better together. You can also change
the elements and put our line underwater,
diving underwater. So the problem now is National Geographic picture
of a loan diving underwater. And that's one of my
favorite prompt ideas to put animals underwater. It's always generating
really funny results. What else do we have? Yeah, we can actually
make our loan an astronaut and make
him wear a space. You, to be honest,
I don't like how that looks in the fifth
version of maturity. I played around with
that prompt little bit more and put the loan
in a spacesuit on mass. And to be honest, the astronauts looks good. It looks as if your mask
has already landed on Mars, but I don't see alone here. So that's a problem if
the fifth version of me, Tony, it's generating
really realistic images, but it's not the
good in generating creative images of things
that do not really exist. For creativity. It can pay off to switch to one of the older
versions for me, Tony, e.g. here is our lone astronaut in the Ford version of maturity. And yeah, actually it
looks much better. Cli on, in a space suit. It's not perfect,
but it's much closer to the idea I had
in the beginning. And here is another version, the liner space suit. And yeah, I like that. It's of course less realistic
going to fifth version, but it's much closer to the original idea I wanted to
create with Mitch journey. If you wish to make your
dreams become reality, it can make sense to switch back to one of
the older versions. And that's something
you should keep in mind when working
with Mitch journey. The older versions are much
better for creativity. And the new fifth version or
even the coming six version, probably Much better for realistic results that are
the basics of prompting. Now you should know how to create different styles
and mutually how to put your pictures or your
objects into locations. How to add e.g. clothes are different
scenarios to your idea. And that gives you a
lot of possibilities. So the best now
is to try it out. Think about some ideas
and try and practice. Put them in different places,
in different locations. And in the next
video, we will look how lighting can affect your images in a
very dramatic way.
11. Midjourney: Defining the aspect ratio: In this video,
we're talking about the aspect ratio in Midjourney. How you can change the
aspect ratio Midjourney. Because if we look at the
standard aspect ratio, for example, this prompt, we see it's one-to-one. That means V get squares
as aspect ratio. So we have now in the
grid mode for squares. And of course, in
many circumstances, we would like to have something different because squares are not the best format
for most pictures. For example, social
media posts need an different aspect ratio
than posters or videos. So you have, in most cases, you want to input the aspect
ratio with each prompts so that you can customize it to the needs of that very project. So the question is,
how can you do that? Well, I will show
you in this video. It's really easy and requires
not a lot of effort. Or you have to do is entering a parameter at the
end of the prompt. So let's show it here
with a practical example. I am going to copy
that prompt again. And now imagine enter. Let's copy the prompt
and at the end with type two minus, so minus, minus a and R. And now
we added aspiration, for example, 16 to nine, which is a tubercle
aspect ratio for screens. So let's enter and let's
wait for the results. And here's our grid mode with
16 to nine aspect ratio. You see how easy
it is to do that. Let's try it another time. If a different aspect ratio, this time we are
using the opposite. And nine to 16 aspect
ratio is actually not the best aspect ratio for landscape or architectural
photography. Unless you're having a
picture of a skyscraper, normally six to nine, I would use for portraits. But of course in the end is
also your artistic choice, which as ratio to use. Unless of course
you're forced to use a specific aspect ratio because you need that image
for a certain medium, for, for, for video, for a certain social
network, etc. will also provide
you with a list of the most popular
aspect ratio so that you have an idea of
what you can use, what are the popular formats? This way you can then later simply copy them
into your prompts. Now, our example
image is already finished and the thing should
be clear how you can do it. Maybe keep in mind one thing
you cannot use commas, you have to use whole numbers
with the aspect ratio. That's the only limitation. But in fact, for the most
popular aspect ratios like four to three to six into 93 to 221 to nine or nine to 16. You don't need to commerce
you just using whole numbers. So Midjourney is
capable of doing that
12. Lighting and Colors in Midjourney: In a previous video, we have seen different
examples of our loan prompt underwater on Mars and in
a byte West Town, etc. Now, I want to show you
how you can use and utilize lighting to add some traumatic effects
to your image. You might know the meaning of photography is actually
painting with light. So the basis of every good photography
is the right lighting. Lighting can
definitely influence the effect of your
image heavily. That is true for photography's, but of course it
also plays a role in paintings like oil paintings
of watercolor paintings. But of course here you might look more at the
colors you are using. Are we later show you how to
use colors in my journey. But now for the beginning, let's look at some
lighting effects you could use to make your
images more dramatic. Here, e.g. we have our
National Geographic lie on inner white West Town that with disco lighting and lighting. Yeah, I like that effect because it gives
it a funky wipe. Normally I would
use it more indoor, but even here, the outdoor
lighting is interesting. It gives that 50
style to the image. If all the neon lighting, my favorite lighting effect
is cinematic lighting. And you see here you have
the cinema style lighting, a long dramatic shadows. And it gives your image a
very dramatic look here, some mistake with two hats, but yeah, that can happen. But overall, I like
those results. If you believed that your
image is too boring, just add cinematic
lighting and it will make your image
way more dramatic. The best time for making
photos is the sunset time. And yeah, you can add sunset lighting into your
prompt and then you have that golden hour look with the setting sun in
the background, the long shadows,
a lot of goals, and it looks beautiful
in most cases. If you're a photographer, you should already know that. Even if you're new
to meet journey, using standard lighting is something that will always make your image more beautiful and give it a really
romantic touch. So feel free to use
that lighting effect. Besides lighting, you can also use or add colors to your image. We have done it here. We have the standard lighting and with the colors blue and orange. So the typical
Hollywood color scheme worked really well in the second picture
with blue and orange. Also liked the first one
with blue and orange. D others, not so blue heavy thing here you could rewrote
to get maybe better results. So let's continue. And here you can have black and white sip here or add even more
colors like red and green. But in my opinion, that doesn't look
too well a thing, it's a little bit too colorful. So you really have to make sure that your colors
are working together. And that is not just
looking like some rainbow. In my opinion. Adding Carlos works
best if you stick to one or two main colors you
want to have in a prompt, if you're using too
many colors, It's, normally, it's not
looking really good. So let's roll it back a little
bit too. Black and white. Yeah, we've seen it
before for some seconds. Cia, black and white, of course, also a possibility for
colors of non colors. And it gives you that tuple traumatic
black and white look. Or you can also say, let's have a Scipio
looks more like the brownish look
on your picture. That also adds touch off. Having a very old
picture in front of you. I think that was popular on
Instagram ten years ago. That was a short discussion of colors and lighting in
Mitchell's journey. But before we end this lecture, let's make another prompt. Copy that into it here. But instead of the National
Geographic picture, we are now typing in an oil painting and
we add some colors. Let's see, what do we want
to have is red and green. And enter. Let's see what's going
to be generated. As I told you in the
beginning of this video, when you are not generating
realistic photography's, but paintings like oil painting, some watercolor paintings. Of course, colors play
a very important role. And you should definitely try to add colors
to your prompt. We have here red, and you see that
the two colors are dominating the pictures we
have here in a grid mode. What would happen if we
now add some lighting? Yeah, let's do that
and try it out. Yeah, we caught
the prompt and now we add cinematic lighting. And let's see how this will
affect our oil painting. And you'll see that
lighting is indeed heavily influencing the
effect of the image. Now it looks way more
traumatic and you have a much more dramatic
atmosphere in that image. So, yeah, even with
paintings, not trust, Don't just to add colors, but also lighting to
make it more traumatic. And let's make another
test with sunset lighting. Let's see how this will
influence our oil painting. You see as a result again, that golden hour atmosphere. But it's still includes the colors we have
prompted, red and green. So yeah, that was the
result we looked for. And that is the conclusion
of our lecture. With adding lighting and Carlos, you can drastically alter the atmosphere and the
effect of your images. It doesn't matter
if you're using, if you are creating realistic photography's or paintings
or illustrations. In all cases, try to play around with lighting
effects, with colors. And look how you can
influence the atmosphere of your prompts of your images
with those ingredients.
13. Realistic Photography in Midjourney: In the previous video, we have seen how to use
lighting and cameras to influence the
character of the images. One thing we will dive into here is the choice
of the camera, but also of the medium you're
using for your images, e.g. as we've seen previously, you could use an oil painting or a watercolor painting of the loan debt is here walking through the white West Town. You have already seen
it last video that you can also use the
lighting for paintings, for oil paintings, like we can see here with
cinematic lighting, that gives a totally different
character to the fore. Images in that great compared to the other oil
painting we have here, where we didn't
define any lighting, but only some colors, red and green to be precise. As we have discussed,
colors can give, can define the atmosphere
of your images. That's very important for,
of course, paintings, but it plays a role
in photography. Before we have seen a
dramatic cinematic lighting. But of course, you can also use the more soft sunset
lighting that creates a beautiful golden atmosphere that is really popular
among photographers, but also among interests. Of course, you can let your inner artists free and
use tons of different styles. Artists size like
impressionism, expressionism. You can do a
Renaissance painting, you can do cubism. In a more advanced video, I will show you some
dozens of studies. I used image journey
and you can use them as a template for your
own prompts, of course. But let's go back to
realistic photography. But because creating
realistic images is a topic of high-interest
for a lot of people, as it seems to me. What are some things you
can add to your prompt to create even more
realistic photography's. The easiest way is to just add a portrait photo of photography
off allowing this case, this creates a classical
portrait shot. And of course, you could also
now add some emotion like an aggressive line and that generates a very
aggressive loan. Compare to the first
image grid where we have to learn
relaxing the sun. And now we have here
the line in a kind of warrior mode, very
aggressive face. You see adding emotion is another way to alter the
effect of the image. Here tried using
depressing Leon. Yeah, the images, they look a little bit more
depressed, more sad. Maybe it's not as striking as the first version with
the aggressive emotion. But still, I think people
would agree that this is a much more aggressive
but more depressed result. Some emotions work
better than others. That is also heavily
influenced by the object you're
using with the line, of course is the results of bad with aggressive,
with other animals. You would get probably
different recites, of course. For the start we used a very classical photography,
portrait photography. Of course, you
could also go into a totally different
direction. Here, e.g. I. Tried to the prompt
refer 360 degree lens. And the result is not
really good to be honest. But that's because lion and 3.6 degree isn't really working. That's better. It would work much better for
it for architecture, e.g. so as the name suggests, the 3.60 degree lens, it crazy imagery for
free or six degree view of the scenario. And yeah, it's much better working for architectural
or like e.g. if you want to display a
wide West Town, that Yeah. So but of course
that's something you should use for very
special occasions. It's not really working
for most, for most images, for most scenarios
you want to create architecture or
interior architecture, where you want to see
everything in a room, e.g. that could be a good
scenario for such a lens. So let's maybe go
back to the portray. And of course you can have different kinds of
portraits, e.g. close up, portray or
close up photography. Adding close-up photography
to your prompt will create images with the object, a lion in the foreground and the background will
be in the bouquet. It will not be sharp, but a little bit blurred. Pretty similar to close up this awesome macro photography. Sometimes it gets you photography's that are
even closer to the object. So a macro lens is a lance
that's allowing you to capture highly detailed images of small subtracts very,
very close distance, often when I'm using macro
lens and the prompt, Actually I'm not getting
any backgrounds, but just the object, they're highly detailed picture of the object in the prompt. That's not the case here. And it's probably because I was saying a running
loan and of course. Every word in the
prompt can actually alter the outcome
of your images. So of course you have
to take care of that. You're not adding any words set contradict each other
in your prompt. You also have macro would be something like drone
photography like here. That's not really working with the lung in
our work West Town. A thing, It's not
even alone here. Let's show some animal. And in the other image, the proportions of the
learner totally wrong. It's more like Godzilla alone. So you'll see me, Tony has still some difficulties
with proportions. If it's a more
unusual perspective, maybe that will change in one of the newer versions
in the future. Nevertheless, for
some scenarios, drone photography can indeed
create stunning results. You're also seeing
here that are wide. West Town now is a
really modern town in the Southwest of the US. So it's not really falling. The idea, I had read
that prompt and you see the fifth version of material is reviewed in creating logical, realistic pictures,
but, but not so good in creating more unusual
creative settings. To get back to macro, here we have a
macro photography, but without the running
loan, It's just a line. And you see now there's
less background. It's much more
focused on the loan, on the hat of the loan itself. So you see how adding or removing
something from your prompt can really make a huge
difference in the final outcome. That is something you
should keep in mind. So less is often more. You should not
include unnecessary, unnecessary words
in your prompt. So make sure that every
word in your prompt is really having an effect
on the final outcome. And that's something you should test when you're
creating prompts, when you're creating images
to keep your prompt clean. And it's something
you can test really easily with the seats in nutrient because
they allow you to, because they allow you to
create more consistent results. So the seats, how to use seats, That's something I will discuss in the advanced section
of Miami tonic chord. Knowing how to use seats to create more consistent prompts. And for keeping your
prompts concise. That's half a usual e.g. when you are creating
problems for customers or if you want to sell
products online. Let's continue with my examples because there's still tons of things I can show you to make
your prompts even better. E.g. here we have
a wide angle or wide-angle photography
offer lie on. The wide angle lens
is used when you not just want to
capture the object, in that case lawn, but
a wider field of view. And normally the wide angle lens can capture more than human. I could capture and is a
really popular lands, e.g. landscape photography. A striking difference,
our portrait shot from beginning is that now
with the wide-angle, the background is
not in the bouquet. It's not blurry, sharp. And to make sure
that the object is at the eye level, Yeah, You can add eye level shot through your
prompt so that makes sure that the object is
really at the eye level, but in most cases mature and
it creates pictures like that is more challenged to
have it not at the eye level. We received an ex examples e.g. you could want to have
shot from the grounds. In that case, you add
ground level shot. That midtone often cannot or doesn't interpret
that correctly. So it can have to say extreme ground-level
shot To make it even clearer because that
phrase helps mature any interpret that
more correctly. Variation of ground-level is the low-level shot,
like we have seen here. You can also combine both
phrases into one shot and you see a low
level is actually generating good results
like them a lot. Even near that picture
is really good and you can have a
very aggressive lion. The opposite of low annual
is high angle shot. It worked very well with humans, but not so much with the lounge. We have the same effect like
with the drone photography. So the proportions are off. The loan is just too big. It's not really creating
realistic results. So high angle didn't
work in that case, but I managed to do it a bit by saying a shot from 2 m high
that created better results. But to be honest, it
didn't really convince me. With some other scenarios
it work better. You see, when you
try to add angles, It's a lot of trial and error until you get
the correct results. Most photographers in reality wouldn't shoot a picture
of allowing from close-up, but I would rather use a Taylor. Taylor lands. Using the term Taylor lands
in your prompt will create the Safari look you may be used to when seen,
learn pictures. If you are a photographer
you will love the following. You can actually add
lens details, e.g. you can say a 20
millimeter lens. And mutually we
create images that represent images shot from
a 20 millimeter lens, which is a classical
wide angle lens. You can also say 100 millimeter. That's more a portray Lance. And we can create
more, normally, should create a more portray like image compared to
the wide-angle lens. But you can even get
more nerdy here and say, Yeah, I want to have
an image shot for, shot from Sony Alpha camera with 85 millimeter lens
and the focal length of 1.8 m. Really good at interpreting the camera
details and land status. And you see the focal length of 1.8 normally creates images
with a nice bouquet. And if you want to have that, you should add that
to your prompt. So you can go both ways. You can either say, I want to portray or
what happened, okay? Or you can just add the
camera details and mutually will interpret it mostly
most cases correctly. One thing I love about
mutual nice that you can, that you can let your inner
camera nerd out and e.g. create Rachel pictures, e.g. polaroid photography, which creates this
'80s or '90s look. And the funny thing is that toes are the most convincing I patrons because the quality
of a real Polaroid is low. So it's hard to tell if you have now an AI Polaroid
or the real deal, because the real
polarize also have artifacts and look a bit
strange and artificial. So the codec 800 films. So if you don't want to
create a digital image, but a film photography, you want to create that with AI. You can do that as well. And fake creates classical
film photography like it was popular in the 90s or early 2000s before the rise
of the digital cameras. Here's an example
of extreme bouquet. So as you know, you can either use the
camera details of say, CBC bouquet or some other things you want to
have in your picture, or you can use both. And that makes the commodity
for stronger and make it clearer to meet journey
what you want to have. Sometimes that helps to get the desired effect of the
image you want to create. So player out of that and use all the options
you have available. But of course, don't lengthen your prompt
in an unnecessary way. If you repeat commands
in different words, only do that to make the
desired effect is stronger. So we had a Sony Alpha and
here's a Nikon D, a 250. And you'll see we recapped
the settings but change the camera model and the
other images look different. I'm not expert enough to say that was really the Nikon look or
the Sony Alpha lock. So if you're an expert in
that realm, of course, I appreciate your feedback
and please tell me how good midtone can distinguish between
different camera models. So one of the last 6 h
is the satellite image. Well, in fact, it's
more like it looks more like an image from an airplane. It's not from a real satellite. And yeah, we have
the same issues like if the drone photography, it's a really wide West Town. It looks more like a modern
town into southwest USA. Yeah, settled images can be great for landscape
photography. So if you want to have
the satellite look of a sudden landscape for a certain region than
that looks nice. And like with the drone
photography, the proportion, of course, if you insert alone, they're not really working,
but yeah, satellite images. The idea not meant to be used
with something like portray shots or with objects that are
visible to the human eyes, but more to have a look of a certain landscape or maybe
structures from high above. So that was my introduction into cameras settings
in my journey. And what you should have
learned in the lesson. Instead, you can actually use
any camera or any cameras setting to create
the desired results. Sometimes the camera angles
don't work so well than you. Maybe you have to repeat some commands in different ways to make it more
clear for maturity. The awesome thing is
that you can actually use any camera and
the camera settings. So if you in the past could not afford expensive
camera equipment, now you can use that equipment immaturity
and shoot pictures. You always wanted to shoot with AI chess for 20 bucks per month.
14. Exploring the Midjourney gallery: So find is course we have seen and learned how
to create prompts for midtone and how to improve the quality
of your prompts so that you can eventually
make money with midtone. Ideally, you have already generated several hundred
pictures with immaturity. And if you're at that stage, probably the following
question. We look here. How can I access
my old creations? Well, that's easy. First, you have discord. And of course in
your Discord server you could now scroll back. Tried to finding the old
pictures you have created. But that's not the best
way to access them. And in theory, you
could also use a discord search function
to find all the images. Bad. To be honest, I don't think that this
code is the best place for finding images you
have created there. What's the best way to
access your old pictures? Well, simply go to the
website of midtone, that midtone.com slash account. When you enter that
link into your browser, you will access
your home screen. And here you have a gallery if all the images you
have created so far, and you have different
categories, e.g. you have hot, then you have
new top and your top images. And you have favorite
images, you have favorited. Important. You are now
in the upscale modes. So you see only upscale
images and not the grid mode. If you want to see the grid, you have to click on grit. And now all the grids you have
created and click at all. We'll show you all images
up-scaled and a grid mode. And you see a whole collection. But in most cases you will probably search for
upscaled images, not for images in grid
mode or not in old. Searching them is
maybe too confusing, but you want to have images you already upscaled in the past. I will now show you how
to favorite and image. In that way you can find
your best images easier. So let's choose one of my
photography prompts here. And now all we need to do is click on the
heart symbol here. And what the image is favorited. It appears here
among the favorites. As before you can, among
favorite pictures, you can select just
the upscale pictures. Or you can say, I want to see all the pictures script
mode, and up-scaled. This way you can easily
find your old pictures and the best pictures you've created because they should
be among your favorites. But you can also explore images from others by
clicking on Explore. That's the community
feed and you see the creations from other users. And that can serve as a
huge inspiration for you. You see what is possible
with me journey because you see tons of ideas. And if you click on an image, we see the prompt that
was used for an image. And we can copy the prompt or the full command and re-purpose that for our
own France, of course. If you want to become
better and prompting, I highly recommend checking
out that feed regularly. So you can look at
hearts are rising. You can also see e.g. I'll click on new images, but normally the
quality of toes is not so high because that adjusts
the new creations midtone, so hot and rising normally has
the high-quality pictures. And of course top is also
something worth checking out.
15. How to use --seed in Midjourney?: In this video, I will show
you how to use seat numbers midtone to create more
consistent prompts. Your first question might be, what our seat numbers and how
can I use them in Mitchell? Joni Mitchell uses seed numbers to create a field
of visual noise, like you might know
from television static. And that filter is serving as a starting point to generate
the initial image crits. For each image is generating
a random seed number. And you can enter that
into your prompt for saying minus minus seat, and then followed by
the number or minus, minus same seat and then
followed again by the number. So if the prompt tax stays the same and you enter
a seat number, the resulting image
will be the same again. But why is that even important? You might ask now, well, let's imagine that we
are inputting a prompt, a very simple prompt. Let's see, Mickey Mouse. And we entered once, and then we enter
the second time. The second time, the generated image will
be different because MIT journey will interpret the prompts a little
bit different every time you enter it again. To generate the
same image again, you would need the seat number. But now you ask maybe, why do I even want to generate
the same image again? I want to generate
different images each time. Well, in fact, if you are
writing more complex prompts, you actually want to experiment
with your prompt text. You want to see what's
happening with my image. If I add a word or what, what's happening if I'm
removing a word and for seeing the effects
of your changes, you would like or you want to have more consistent
prompts, of course. Because entering the
same prompt again, the if the seed number is
the only way to secure or to make sure that you will
generate the same image again. And it makes sure that when you are altering the
image or the prompt, the resulting image
will not be too different from the first
image you have created. Because your results
and more consistent, you can test the impact of
small changes to your prompt. This becomes very
important when you're starting to write more
complex problems, or e.g. if you are entering the
field of prompt engineering. Okay, That's the theory. Let's move to the more
practical part of the video. And your first
question is probably, where do I get the seed
number of my images? For this, we're going
to discard again. And I have prepared a picture or an image
that's an image of water white in a retro fantasy
setting from the '80s. Now we're going to make some small changes to
the prompt and see how much that will affect
the generated image. But first we need to get a
seat number. How do we get it? Where we need to
add the reaction? And we click here right top. And we need to click on the
envelope or enter envelope. If the envelope emoji is
not appearing here in the, in the selection list. And now we click Enter. And that worked. Now, the medtronic
bought messaged us and we see the message in the left sidebar here
on top, maturity bonds. Let's click here. Yeah, Now that's the
message of the bond and we have to chop ID and
the seat number. Let's copy that. And now we can go back
to our private server. And that's it. That's how you find out the
seed number of your images. Important. You cannot do that
with other users images. If you want to find
out the seed number of an image that was
generated by another user, you have to politely
ask that user. That's at least for a moment
the only way to do that. Okay. Now we know how to
get the seat number. But now I'm going
to show you what you can do this
that seed number. As I told you before, I want to make some changes to
the original problem. So let's type in imagine and copy the prompt with the
seed number we got before. Okay, let's change
some small detail. Maybe we can change
the color of the cloak from yellow to what
do we type in here, maybe what is with green? Let's have a green
cloak and enter. Let's see what's happening. And the image is generated. And we see that there are indeed some changes to the original compared to the original composition
of the image. So there are indeed more changes than chance to cut off
the clock. So e.g. we have the background that is different and even
the character, even worldwide in
the foreground, it looks a little bit different. Okay, let's continue
with our experiment. And let's see if we can remove some words of our
prompt from our prompt. So maybe we can remove
words like retro and cheap. And let's enter and
see what's happening. Here's our image.
And we see that the overall atmosphere
is actually not so different from our original. Now hear this from
our original image. Of course, it's a
little bit different. But it seems that
both words cheap and Rachel didn't influence the
atmosphere in a strong way. So as a conclusion, we can easily delete them
or replace them with words that may be influenced the image or
the prompt stronger. But what is important
to point out here is that as soon as we make a
small change to the prompt, the image, we change as a whole. So we're not going to get the same image just with
a minor detail changed like the color of the clock
or the atmosphere delighting. Know every minor change
we trigger midtone to interpret the image
in a different way. Let's do another experiment. We now call up the prompt
and I will enter the prompt once without seat number
and once we've seed number. No, that's correct. And a second time. But first the seat number. And now we're ready. Let's see what
metronome will do now. And I think that's a
thing I should stress maybe even more because as soon as you change
something to prompt, you will not get a totally
consistent results. It's just if you enter
the seat number, your result will be
more consistent. Okay, now we've
got our two images That's with seed number and we see it's pretty similar
to the original image. I think it's acts, act the same. And now we have the new or the other one without
seat them and receipts, a totally new interpretation
of the prompt. So the takeaway
of this lesson is that using the seed number will not lead to identical results. If you make minor
changes to the image, but it will lead to a bit more consistency
because without seat number, the changes would be
much more severe. So using a CDMA
is helping you to get more consistent results. In mid journey, of course, normally you're not
going to create identical pictures or images, again with the same prompt because that's more
or less unnecessary. If you want to go back to your
image or generated image, you would simply go to the
gallery and then download it. But of course, in
many cases you want to do some minor
changes to your prompt. And in that case, it makes
a lot of sense to use seed numbers to have at least some consistency
in my journey.
Philipp Stelzel, Notion Enthusiast, Writer, Digital Creat
