AI topics have become increasingly popular recently.
Recently, OpenAI's video generation app "Sora. "
※For more information on Sora, please see the following review:
And on December 12 of the same year, Apple released an image generation app for iOS 12 called "Image Playground(I have an iPhone 14 so it's not supported)
And recently, Google's video generation AI "veo2" and image generation AI and other generation functions included in Adobe Photoshop 2025.
I've been paying particular attention to Photoshop 2025's generation functions recently because they are of high quality and have fast rendering speeds.
In the meantime, the image generation AI that I am using to create my blog right now is,Microsoft Designers" I think this is one of the most user-friendly image generation AIs. The reasons for this are:
- It can generate photorealistic images that are not often seen in other apps.
- It is also great for generating Japanese anime-style illustrations.
- Four patterns of images are generated with one command
- It responds to commands quite faithfully.
- If you have an office account it's free to use.
- Rendering is fast.
I think the above points make it stand out from other apps. Of course, it has its weaknesses, but it more than makes up for them by generating images that follow your commands.
So, I would like to introduce it with some examples and how to use it.
What is Microsoft Designer?
Microsoft Designer is an image generation tool provided by Microsoft, and is a service that aims to enable users to easily create high-quality designs using AI technology. In particular, it is characterized by the fact that even beginners can easily create professional images.
It is also possible to create graphic designs from scratch, like its competitor Canva. The core of Microsoft Designer's image generation program is OpenAI's "DALL-E3" engine, which means that images of the same quality can be output by ChatGPT4-o.
However, I personally find Microsoft Designer easier to use.
It generates four patterns of images at once and allows you to choose the image ratio. Image ratio is particularly difficult to control with ChatGPT. It also has various editing functions using AI functions. Microsoft Designer has made DALL-E4 easier to use and more professional.
How much does Microsoft Designer cost?
It's basically free to use. *A Microsoft account is required.
In addition, images are generated by consuming points called boosts, and you can generate up to 1 boosts per day. (Basically, 15 boost is consumed per generation) This is 1 boosts that are compensated every day, so so far I have not had any trouble with running out. Also, since you can create 1 to 15 variations of images at a time, you will not have to worry much about consuming boosts.
I've also included the smartphone version of the app below. The usage is basically the same. (The smartphone version has a few more AI editing functions, though.)
Basic usage
I'll put off the actual examples for now and first explain how to use it. The explanation is based on the PC version.
The PC version has "Enhance prompts" button, which when clicked will use AI to make the prompt more command-like. However, since it also contains unnecessary commands, I don't use it much.
Below is the generated image.
It was pretty easy to create the image.
The good thing about Microsoft Designer is that it has a fast rendering speed, so repeated retries aren't too painful.
Recently, I've been using this service more and more to create eye-catching images for my blog.
What are the disadvantages of Microsoft Designer?
Like any other service, Microsoft Designer does have some weaknesses.
What is that?Weakness in fine detailsFor example, in the photo below of a desk with gadgets lined up, the keyboard and buttons look like they're made of clay.
This problem could not be fixed no matter how many times we tried to generate the image, and it is a phenomenon that has been seen in other image generation services as well, so I think this may be a breakthrough point for image generation services at the moment.
The rest at this point is "Multimodal" is not supported.
Having this feature makes a huge difference in the usability of image generation, so I hope they'll address it soon.
I tried to generate various images
Here, I used Microsoft Designer to generate various images.
First, generate a fashionable photo.
The command is as follows:
A boy wearing a black Comme des Garçons jacket, seen from behind. He has long hair and is looking towards the camera. The image is in black and white.
And the generated result is as follows:
The intended result was achieved.
It's also great that they even design and include the Comme des Garcons logo.
(The spelling is weird though.)
You can choose from three photo ratios, allowing you to create images with a high degree of freedom.
Next, let's try something a little more dynamic.
The following is the command statement.
It looks like a temple somewhere in China. 100 shirtless monks are running towards the camera. The camera is shot from a bird's-eye view, giving the finished product a cinematic feel.
And the generated result is as follows:
This is also pretty faithful to the commands. It's so faithful that it even generates unnecessary cameras. If you look closely, you can see that the image becomes more distorted the further away it gets, but is this within the acceptable range?
Let's move on to the next step. This time I want to create a dynamic image of a drifting car.
The following is the command statement.
A dynamic photo of two cars drifting side by side at the D-1 Grand Prix. The cars are an AE2 and a Skyline GTR. Use zoom blur to bring out the impact.
And the generated result is as follows:
Wow, that looks pretty good.
It's also great to see the effort put into recreating the car model. This is an illustration-like finished product.
I have created three photos so far, and you can see that the results are faithful to the commands.
Next, let's change things up a bit and try to create a moody photo.
The following is the command statement.
A certain shot bar. A man and a woman sit at the bar counter. The man is a middle-aged, sophisticated man. The woman is in her late 20s. Both are Japanese. They toast with a glass of whiskey. There is a smartphone on the counter. The result is a moody photo.
And the generated result is as follows:
it's amazing.
Even though it was a complicated scene, the command sentences were almost reproduced. There was even a smartphone on the counter, so I was a little surprised at the result. There was no image distortion, and the result was very natural.
This is not possible with other services.
So far I have no complaints about the image quality.
Next time, I would like to check the quality of the graphic design rather than the image.
Graphic design creation skills
There are many different types of graphic design, but I would like to create three types of design: a logo design, a greeting card, and a business card.
Logo creation ability
The following is the command statement.
Please design a logo for my brand. I run a gadget blog. Please create a simple, chic design with good typography. The background should be white. The text for the logo should be "Gadget Mania Z." The letters should be solid navy blue.
The generated result is as follows.
The results of this generation are not so different from other apps. I tried it several times, but it was the same American design. It's elaborate, but it's not simple and hard to use. I tried changing the command many times, but I couldn't get a good design. It's quite difficult...
Since the text can be garbled like this, it may take some skill to use as a logo creation tool.
As of now (December 2024, 12), like other tools, the logo generation does not seem to have reached the desired quality.
Next time I'd like to make something with a seasonal feel. To continue testing my graphic design skills, I'll be making a greeting card.
Greeting card making skills
The following is the command statement.
The art style is in the style of Japanese anime. A long-haired girl, similar to Hatsune Miku, dressed as Santa, is floating in the sky above a night city. Snow is falling, creating a dramatic image. The letters 𝑀𝑒𝑟𝑟𝑦 𝐶ℎ𝑟𝑖𝑠𝑡𝑚𝑎𝑠 are placed in a good position. DM card.
The generated result is as follows.
Wow! This is pretty amazing!
The subject matter chosen was perfect, as it is a Hatsune Miku. And the illustration is very dramatic.
He seems to have excellent graphic design skills, being able to create things like simple greeting cards with a single image.
Now for one last graphic design test. I'm going to have you make some business cards.
Business card design creation ability
The following is the command statement.
Design a vertical business card.
A simple business card with a base color of red and an accent color of black. The main text is "Gadget Mania Z", the subtext is name: TOMO, title: graphic designer, and the address and phone number are written appropriately.
The generated result is as follows.
This was the best of the four proposals generated.
Well, this design is a good idea for a brainstorming session.
But the other three cards were no good at all, so the success rate seems low.
I think it would be a better design if there was at least a multimodal function that allows you to share and refer to photos, etc., but it's difficult to do it with a logo. Even small letters are distorted...
General comment
- Photorealistic objects can be reproduced to a great extent. Even those who were dissatisfied with the results of image generation AI can now be satisfied.
- A two-dimensional illustration like a logo mark instantly gives it a strong American feel.
- The greeting cards are simply designed and show great design ability.
- Business cards also look American when they have a logo.
- Since it is not multimodal, prompting skills are required.
Conclusion
Although their image generation skills are strong, their graphic design skills are still in development.
*It is possible that it cannot be used for brainstorming ideas.
Other graphic design features
Microsoft Designer is a graphic design tool, and has a wide range of functions other than image generation AI. Below, we will explain using the smartphone version screen. (The smartphone version has a few more functions.) Select the "Edit with AI" tab at the top of the screen.
Generative erase
I tried to try it, but the loading never finished so I gave up. I think it's a function that uses AI analysis to remove unnecessary parts of the image.
Restyle Image
This function uses AI analysis to change the image to a specified image style. This is quite excellent.
Below are some examples↓
Background blur
This function uses AI analysis to add depth of field to the background of the specified image. This is quite impressive. I feel like it's a bit too blurry though.
Below are some examples↓
Remove background
This function literally uses AI analysis to remove the background of a photo and save it as a transparent image.
There will be some jagged edges when cutting out delicate things like hair, but this may be a useful feature depending on the photo.
Below are some examples↓
Frame Image
This function allows you to choose your favorite frame from a wide variety of options and insert a specified photo into it.
Below are some examples↓
Editing an Image
This is a function that allows you to easily retouch your photos.
There aren't many retouching options, so I felt it would be better to use another photo app.
Edit screen below:
I'll introduce some photo retouching apps that I recommend:
[Fotor]
Expanding video production with the combination of Sora and Premiere
I think it would be a waste to just generate images.
Here are some creative possibilities I can think of:
Example) Creating a greeting video
4. Create 5 to XNUMX images with a Christmas theme using Microsoft Designer.
② Bring the image into the open AI Sora and add animations. * Make it snow, make the reindeer run, etc.
3) Use Adobe Premiere, Vrew, etc. to arrange the animated videos on a timeline, cut and edit them, and add music.
Use Microsoft Designer → Sora → Premiere in three steps.
In this way, you can create anything from still images to videos. It's exciting to think about it, isn't it? All of the operations are simple, so please give it a try.
Summary: Suitable for generating photo-like images
To me, many of the features seem like a toy, and it doesn't really have the appeal of being a graphic design tool.
It will likely continue to mature with future updates, so we look forward to seeing what the future holds.
On the other hand, when viewed as an image generation AI, it appears to boast a fairly high level of completion at this point.
I felt that the quality of the images, especially of people, was very high.
*When using detailed machinery, the images may become distorted in a similar manner to that of clay.
Except for some rare cases, I don't think there are many services out there that have such a high accuracy rate for command statements at this time. (Please forgive me if I'm wrong.)
If you have a Microsoft account, you can use it for free, so why not give it a try? Let's spend this New Year's holiday being creative!
![The magic of "fragrant"... How to use "solid perfume" in a chic and fun way for both men and women [Review]](https://mainoriti.com/wp-content/uploads/2025/02/IMG_0065-3-485x273.webp)
