Nice to meet you!
Thank you for opening this article!
My name is YOS and I post information about cars on Instagram.
Although the car genre is a fairly niche genre on Instagram in Japan, the AI videos introduced here helped the account reach 10,000 followers within three months of launch🎉 (Now, six months later, it has reached 19,000.)
I had no Instagram know-how whatsoever, but the reel video I’m introducing today, which dramatically depicts the evolution of a car using AI, achieved 1 million views from the very first video.
Similar reel videos I posted afterwards also consistently received tens of thousands of views, so I was able to achieve this in a relatively short period of time, even though I was an amateur.
I’ve been incredibly grateful to receive so many comments and DMs from people who’ve watched my reel videos, as well as from people who are already broadcasting and have tens of thousands of followers, 💬asking questions like,
“How do you make them?”
“What kind of AI tools do you use?”
And not just from within Japan, but from overseas too!
So I decided to create a manual that covers the tools, techniques, and tips I use to create car transformation videos📖
If you read this article and put it into practice one by one, even if you are a beginner at AI, you will surely be able to recreate the transformation videos I create!
I believe that learning this AI video generation know-how will be a great tool⚔️ for creators, especially those in the automotive genre, to increase their followers.
In fact, the video I made above of the Stepwagon evolving has been viewed over 3.8 million times and I have gained over 2,000 followers.
That’s an increase of over 2,000 followers just from one video!
On social media, going viral has the potential to bring about a sudden turnaround.
As my track record proves, I believe this transformation video has the potential to bring about a sudden turnaround.
Of course, it can be used for machines like airplanes🛩️, trains🚃, ships🛳️, and robots from anime🤖. Depending on your imagination, I think it could even be applied to the history of animal evolution🐕
I believe there is room for further development with the creativity of all creators. I hope that this article will help you develop the tools you need to survive on social media.
Now, let’s put it into practice together!
What you will learn in this article
1.Tools Used (including paid tools)
2.How to Use Each Tool
3.Workflow Using AI Tools
4.Important Notes and Tips
First, I will explain the entire process.
<Transformation Video Workflow>
- 1. Theme Selection: Select the vehicle model or brand to feature in the video.
- 2. Image Creation: Use AI to generate multiple vehicle images that fit the selected theme.
- 3. Transformation Video Generation: Use AI to create a video based on the generated images.
- 4. Video Editing: Stitch together the generated transformation videos. Add sound effects, text, etc.
- 5. Posting: Add background music and post on Instagram.
目次
1. Tools Used🔗
First, let’s introduce the tools we’ll be using. Detailed usage instructions for each tool will be covered in the next chapter.
Here, we’ll explain the role each tool plays and recommend plans.
① Google AI Studio
This is an AI service provided by Google that functions similarly to ChatGPT.
In fact, it’s completely free with unlimited usage and boasts significantly faster generation speeds.
For the specific task of creating car evolution videos, it’s a complete upgrade.
<🚩Role>
・Generating base car images
・Suggesting vehicles suitable for the theme
<About the Plan>
It’s free with unlimited usage.
There is no need to subscribe to any specific plan.
②Canva
This is an essential tool for posting on Instagram, and I also use it when creating transfomation videos.
<🚩Role>
・Adjusting images generated by Google AI Studio (enlarging, expanding, improving image quality, etc.)
・Organizing generated images
<About Plans>
The 【Pro Plan】 is recommended!
Canva offers both a free plan and a paid Pro plan for personal use.
For this project, please subscribe to the Pro Plan at ¥1,180 per month.
The free plan has limited functionality and lacks the Magic Extend feature we need.
Canva is incredibly useful for Instagram creators, significantly aiding feed and story posts.
Subscribing to a paid plan unlocks more materials and features, making it convenient overall.🔧
③Video generation AI
This is the most important tool for creating transformation videos🏷️
It is used to generate videos that transform the generated images into VFX-like effects.
<🚩Role>
・Used to generate scenes where a car transforms into another model
We recommend [Pixverse]. There are many video generation AIs, but to create a transformation video, you need a video generation AI with [Transition Mode]. Transition Mode is a video generation mode that seamlessly changes and connects the [First Image] to the [Second Image].
<Recommended generative AI services>
・Pixverse
There are many video generation AIs available, but to create a transforming video, you need a video generation AI with a 【transition mode】.
“Transition mode is a video generation mode that seamlessly transitions from the “first image” to the “second image”.
After trying out various types of AI generation, I found the following two types to be useful, but the one I would particularly recommend is Pixverse.・Pixverse
It generates videos of a stable, usable quality. For better or worse, it outputs the results exactly as you input them, so it doesn’t generate useless videos, but it also doesn’t generate videos that go beyond your imagination.・Kling AI
It can generate images comparable to professional CG expressions, and its visual quality is higher than Pixverse.
Also, by adjusting the parameters, it can generate cooler videos that exceed the input prompts.
However, it sometimes generates useless videos that are lower than the desired quality, such as a car driving in reverse, and is significantly inferior to Pixverse in terms of stability and cost.
If you want to create lots of videos without any hassle, try Pixverse.
If you want to create high-quality, cool videos, even if it takes time and money, try Kling AI.
That’s my impression after trying them out 🤔
Both AIs can be generated for free 💸, so you can try them out yourself and use the one you like best, or try using both.
<About the plan>
Minimum [Standard]. If possible, [Pro Plan] is recommended.
If you are serious about posting videos on both Pixverse and Kling AI, you must subscribe to a paid plan.
The free plan only allows you to generate a few videos per month, making it completely useless for continuously posting videos.
Also, a watermark (the name of the video-generating AI service) appears in the bottom right corner of the video. It's said that Instagram reels with watermarks tend to limit their reach, so we strongly recommend subscribing to a paid plan.
Both Pixverse and Kling AI offer the following plans, but it's best to subscribe to the Standard plan at the very least.
If possible, the Pro plan is preferable.
※About Pixverse: Enter the referral code below when registering to receive bonus credits.
referral code：5XD1B4TH
◆Pixverse plans
・Basic: $0/month
・Standard: $10/month (Can generate up to 720P resolution)
・Pro: $30/month (Can generate up to 1080P resolution )
・Premium: $60/month
◆Kling AI plans
・Basic: $0/month
・Standard: $6.99/month
・Pro: $25.99/month
・Premium: $64.99/month
④Video editing software
The files output by the video generation AI are not in a state that can be shown to others, so we perform some simple editing using video editing software.
<🚩 Role>
・Connect videos generated by the video generation AI to create one video
・Insert text such as the car model name and specifications as needed
・Insert sound effects, etc.
<Software to use>
Basically, you can use any video editing software.
I use 【Davinci Resolve】.
The software that was also used to edit the movie 『Avatar』.
The reason I use it is because it's free and has many features.
I also used it when I originally posted videos to YouTube, so I'm familiar with it.
If you already have software you use, or if there's software that suits you, feel free to use that instead.
Needless to say, I’ll use it to post.
I’ll also set the background music.
<🚩Role>
・Setting background music for reel videos
・Posting
Summary of tools used
Now, let’s review the tools we’ll be using.
- ① Google AI Studio: Generate the base image
- ② Canva: Adjust the image generated in ① (enlarge, expand, improve image quality)
- ③ Pixverse: Create an animation of the car transforming from the base image
- ④ Editing software: Stitch together the generated transformation video
- ⑤ Instagram: Add background music and post
２.How to use each tool and work procedures
From here, we will explain the steps to create a reel video using each tool, step by step.
STEP 1: Decide on a theme
First, start by deciding on a theme for which car to evolve.📄
For example, when I choose a car model based on a theme, I consider patterns like these🚗
・Cars that have evolved repeatedly under the same model name (like the Porsche911, FORD MUSTANG,etc)
・Cars from the same manufacturer that are similar to each other (like convertibles or those with the same engine)
Once the theme is set, select models that fit it.
You can base your choices on your own knowledge, or try searching online or asking AI for suggestions🔍
I often ask AI for help when deciding on themes.
<How to get AI to suggest models based on a theme>
Here's how I actually get Google AI Studio to suggest a car model.
(1)Open Google AI Studio
First, open Google AI Studio.
When the home page appears, select 【Gemini 2.5 Pro】 from the AI models in the top right corner of the page.
② Request to AI
Once you enter the chat screen, you will first be shown the following prompt.
I'm going to create a video showing the evolution of cars based on a certain theme. I'd like you to suggest car models that fit the theme I've presented.
The AI responds with the following:
From now on, if you bring up a theme in the chat, the app will suggest car models based on that theme.
For example, to theme a Toyota RAV4, enter the following prompt:
The theme will be the Toyota RAV4.
Please cover all previous models, predecessors, and successors. Please provide suggestions if the base model has undergone a minor change and its shape is significantly different.
In addition to the base model, please also pick out iconic models, concept cars, prototypes, competition vehicles, etc. to introduce.
The AI will then suggest car models based on the theme, as shown below.
Based on the responses, you can eliminate unnecessary ones and add ones that are lacking to decide which car models to feature in your video👍
The Evolution of the Toyota RAV4: A Comprehensive Guide for Your Video
Here is a detailed list of Toyota RAV4 models and concepts to feature in your video, charting its course from a quirky concept to a global bestseller. This guide covers its predecessors, all generations with notable changes, iconic variants, and glimpses into its future.
Predecessor: The Visionary Concept
1989 Toyota RAV-Four Concept
The story of the RAV4 begins with this concept, first displayed at the 1989 Tokyo Motor Show. Its name stood for “Recreational Active Vehicle with 4-Wheel-Drive.” The design was a friendly, rounded departure from the boxy 4x4s of the era, featuring a hidden winch and removable roof panels. It was Toyota’s initial vision for a “neo-urban 4WD” aimed at young, active city dwellers.
First Generation (XA10): The Segment Creator (1994–2000)
The first-generation RAV4 was a trailblazer, effectively creating the compact crossover SUV segment. It combined the high driving position and versatility of an SUV with the nimble handling and fuel efficiency of a smaller car.
- 1994-1997 Models:
- Iconic 3-Door Model: This was the original, quintessential RAV4. Its short wheelbase and “cutesy” off-roader styling gave it immense character. It was marketed as a fun, recreational vehicle.[1][2]
- 5-Door Model (from 1995): Introduced a year after the 3-door, this more practical version offered more space and seating for five, broadening its appeal to families.[1][3]
- Significant Changes and Iconic Models:
- 1998 Facelift: The RAV4 received a slight restyling to its front and rear fascias, including new crystal-style headlights, to freshen its look.[3]
- 3-Door Soft Top/Cabriolet (from 1998): A rare and iconic version that featured a convertible soft top over the rear passenger and cargo area, emphasizing its “fun in the sun” image.[2][4]
- First-Generation RAV4 EV (1997-2003): A limited-production, all-electric version developed to meet California’s zero-emissions mandate.[5][6] While visually similar to the standard RAV4, it was a pioneering effort in the EV space.
Second Generation (XA20): Maturing and Refining (2000–2005)
The second generation saw the RAV4 grow in size and sophistication to compete with a host of new rivals it had inspired. It shed some of its initial quirkiness for a more mature and refined design.
- 2000-2003 Models:
- Significant Changes:
- 2004 Facelift: This mid-cycle refresh brought updated bumpers, headlights, and grille, giving the XA20 a sharper look for the latter half of its production run.[8]
Third Generation (XA30): Growth and Power (2005–2012)
This generation marked a significant evolution, with the RAV4 moving to a brand-new platform and becoming larger and more powerful. It also introduced a notable regional variant.
- 2005-2008 Models:
- Standard 5-Door Model: The RAV4 dropped the 3-door body style entirely. It now featured a more rounded and muscular design, and for the first time, the spare tire was removed from the back door on some versions, which now opened upwards instead of to the side.
- Significant Changes and Iconic Models:
- 2009 Facelift: The front and rear fascias were redesigned for a more contemporary appearance.
- Toyota Vanguard (2005-2013): A key model to feature, the Vanguard was a Japanese-market version of the long-wheelbase RAV4. It was more upscale, featuring a different grille, headlights, and bumpers, and uniquely offered a third row of seats, making it a 7-seater.[9][10][11][12][13] This highlights how the RAV4 platform was adapted for different market needs.
- Second-Generation RAV4 EV (2012-2014): Built on the third-generation RAV4 body, this was a partnership with Tesla, which supplied the electric powertrain and battery.[5][14] It was visually distinguished by a unique front grille, bumper, and badges.
Fourth Generation (XA40): A Sharper, More Modern Look (2013–2018)
The fourth-generation RAV4 adopted a much more aggressive and angular design, aligning it with Toyota’s design language of the time. It also marked the introduction of hybrid technology to the lineup.
- 2013-2015 Models: This redesign featured a sharp front end and a conventional top-hinged tailgate across all models.[15] The V6 engine option was discontinued.[4]
- Significant Changes and Competition Vehicle:
- 2016 Facelift and Hybrid Introduction: This was a major update that streamlined the exterior styling.[4] The most significant development was the introduction of the RAV4 Hybrid, which offered improved fuel efficiency and became a cornerstone of the RAV4’s appeal.[6]
- 2015 Toyota RAV4 Rally Car: For a dynamic segment in your video, feature the rally-prepared RAV4 driven by Ryan Millen. This vehicle competed in the 2WD-Open Class of the Rally America series, showcasing a different, more rugged side of the RAV4.[16] Despite being mostly stock in terms of its engine, its durability proved to be a major advantage.[6]
Fifth Generation (XA50): Rugged Redesign and Peak Performance (2019–2025)
The current generation represents a dramatic shift in design, moving away from a car-like appearance to a more robust, truck-inspired aesthetic. It’s built on the Toyota New Global Architecture (TNGA) platform.[17]
- Base and Hybrid Models: The design is characterized by a bold, upright grille and chiseled body lines.
- Iconic Models and Trims:
- Adventure/TRD Off-Road Trims: These models feature a more rugged front grille, chunkier fenders, and unique wheels, emphasizing the RAV4’s off-road capabilities.
- RAV4 Prime (PHEV) (from 2021): The most powerful RAV4 ever produced, this plug-in hybrid boasts 302 horsepower and an impressive all-electric range.[4][7] It is a must-feature model to showcase the pinnacle of the RAV4’s performance evolution.
- Toyota Wildlander: For a global perspective, include the Toyota Wildlander, the RAV4’s counterpart in the Chinese market. It features distinct styling, particularly in the front and rear fascias, giving it a sleeker, more Lexus-like appearance.[17][18][19][20][21]
Successor: The Next Chapter (2026 and Beyond)
The evolution continues with the upcoming sixth generation, which is expected to further modernize the RAV4’s design and technology.
- Sixth-Generation Toyota RAV4 (Expected 2026): Early looks and concepts suggest an all-new design with cleaner lines, potentially moving to an exclusively hybrid and plug-in hybrid lineup.[7]
- Concept Visions for the Future:
- Toyota RAV-X Concept (2024): Unveiled at the SEMA show, this radical concept is inspired by off-road race vehicles, showcasing the potential for a high-performance, rugged RAV4.
- RAV4 “Dark Side” Performance and “Outdoor Lifestyle” Concepts (2025): Previewed for the Tokyo Auto Salon, these concepts explore two different futures for the RAV4: one a street-focused, sci-fi inspired performance model, and the other a more adventure-focused off-roader. These concepts are perfect for the closing segment of your video, hinting at what’s to come.
※As of October 2025, the cutoff date (learning completion date) for Gemini 2.5 Pro is January 2025, so it is highly likely that models after that date and in the second half of 2024 will not be proposed✒
STEP②：Generate a base image
<Tools used>
Google AI Studio
Once you’ve decided which car model to feature in the video, you can use Google AI Studio to generate a base image. 🖼️The image you want to generate should capture the entire car, as seen in the image below, with the car traveling from the front left diagonal as seen from the car. 📸
As a concrete example, let’s look at the Toyota RAV4, which was the subject of our previous discussion.
The actual prompt used to generate the image above is:
Please generate an image of a car in motion.
<Vehicle Model>
Toyota RAV4 (First Generation) ← Please meticulously replicate the first image.
<Conditions>
・Strictly and accurately maintain the 9:16 aspect ratio.
・Angle of view: Low angle from the front left diagonal side (refer to the second image).
・Location: Unpaved mountain pass road (background to be depicted with slight blur).
・Time: Daytime
・Weather: Sunny
・License plate: White background (no details).
・Driver not depicted.
・Right-hand drive. Location: Unpaved mountain road (blur the background slightly)
・Time: Daytime
・Weather: Sunny
・License plate: White outline (no characters)
・Do not depict the driver
・Right-hand drive
<Expression Conditions>
・ The car is traveling at high speed.
・Ensure the entire body (front to rear) fits within the image without being cut off.
・Generate with neutral daylight white balance (approx. 5000 K) for clear, non-yellowish color tones.
・Brighten the image.
・Realistic photo.
・ Realistically render the car body’s gloss and reflections.
・Meticulously render vehicle details based on the attached photos.
I’ll explain the prompt in detail.
About attached images
The prompt above states, “Copy the first image in detail” and “Refer to the second image.
“The first image you attach will be the source image of the “car you want to generate.”
The image you attach here will determine the color and shape of the generated car.
The second image you attach will be an image to “determine the composition.”
In short, there will be no problem if you attach the image of the RAV4 I showed you above.
Vehicle type and scenario settings
Please change the parts in bold red in the prompt above according to your situation.
(1) Vehicle type For the car model, please enter the manufacturer name followed by the car model name (model number), such as Toyota RAV4 (first generation). (2) Location of the race You can change the background by changing the content here. If you plan to create multiple videos using this method, it's better to change the background every time, as viewers will get bored if you use the same background every time. Try experimenting and changing it to a location that matches the theme 🖼️ The RAV4 we created this time has a strong off-road element, so we set an unpaved mountain road as the background.
(3)Time
Set it to morning, noon, night, etc. 🌃
If you set it to night, it will automatically generate an image with the headlights on.
(If they are not on, you can point it out and it will generate another one)
(4) Weather
You can literally set the weather to sunny or rainy 🌤️
(5) Handle position
It’s just a matter of personal preference, but you can specify whether the car should be right-hand or left-hand drive.
Sometimes the AI won’t listen to you and will set it to the opposite, but this can be fixed by regenerating it a few times.
Regarding expression conditions
There is no need to tweak this part.
I arrived at the conditions for printing images clearly and beautifully through trial and error.
⚠︎Notes⚠︎
Regarding image aspect ratio:
At the time of writing, images generated by Google AI Studio will use the aspect ratio of the last image attached to the prompt. (The reason is unknown.)
For this reason, we recommend that the last attached image be as vertical as possible. (This can be adjusted later, though.)
Also, the maximum pixel count is quite low at 768×1344, which falls short of the full HD resolution of 1080×1920, but this will be adjusted later.
Using the prompts above, you can repeatedly generate images in Google AI Studio to generate a series of images of car models that fit the theme.
※If it doesn’t work, try changing the attached image and try again.
<Regarding image generation after the AI cutoff date>
If you generate images according to the prompts above, you may encounter a situation where you can’t generate a new car model.
This is because information after the cutoff date doesn’t exist in the AI’s database. 🧠
The sixth-generation RAV4 (sixth generation), shown on the far right of the image above, was released in May 2025, so as of the time of this article’s revision, the AI has not yet learned it and cannot generate it using the prompts above.
However, you can generate models beyond the cutoff date by using the prompt below.
Generate an image of a car driving.
<Conditions>
- Location: Unpaved mountain road (the background should be slightly blurred)
- Time: Daytime
📌Please be sure to attach an image of the car you want to generate to the prompt above.
Also, feel free to change the driving location and time.
The image above was generated in landscape orientation because only landscape images were available (as mentioned above).
The reason why prompts compatible with the cutoff date can be generated after the cutoff date is probably because the prompts have become simpler, and there is less information to process, which means there is more capacity to faithfully reproduce the original image.
STEP ③: Adjust, expand, and organize images
<Tools Used>Canva
Images generated by Google AI Studio may not reach full HD quality or may only be generated in landscape orientation. Therefore, we use Canva to expand and adjust them to a 9:16 aspect ratio.🖌️
③-1: Open the Story creation screen
First, open Canva’s home screen and select [Story] to open a new editing screen.
If it’s not located where shown in the image below, please select [Story] from the new project creation screen.
③-2: Add an image and adjust its size and position
Drag and drop the image you created in Google AI Studio into Canva 🖱️
The image will be inserted very small relative to the frame (with a lot of white space).
The image is too small as is, so we’ll enlarge it until the car is fully visible and adjust its position accordingly.
Position the image so the car is roughly centered on the screen. (Adjust this position while visualizing the final product to align with where you’ll place captions or other elements.)
※ Adjust the position of horizontally generated images in the same manner.
Some vertical margin will appear, but this will be addressed in later steps, so it’s not a problem.
③-3: Perform high-quality image processing
In step ③-2, the image quality becomes rough because the display size is enlarged from the original image, so we will process it to improve the image quality.
Click [Edit] at the top center of the screen and select [Image Upscaler] from the Apps tab.
Since there is no problem with doubling the image quality, select [2x].
After a while, the process will be completed, so click [Replace] to finish.
③-4:Removing watermarks
Remove the watermark displayed in the bottom right corner of the image.
Click [Edit] at the top center of the screen, and then click [Magic Eraser] in the Magic Studio tab.
After filling in the watermark area, click [Delete] to complete the process.
③-5: Depicting the blank spaces
The process up to this point has created some blank space on the page, so we will fill that space.
Click [Edit] at the top center of the screen, and select [Magic Extension] on the far right of the Magic Studio tab.
You will be asked to select a size, so choose the middle option, “whole page.”
(Other options are fine, but “Full Page” will produce the highest quality image.)
After about 10 seconds, four images will be generated.
Select the one that looks the most natural and press Done.
Now, one image is complete.
If you compare the image generated by Google AI Studio with the image adjusted in Canva, you can see that it has been enlarged considerably, and the magic-enlarged area looks natural, as shown below.
Left: Image immediately after generation / Right: Image after adjustments in Canva
Repeat this process for each generated image.
Once you have finished organizing all of your models using the same steps, click [Share] in the upper right corner and select [Download] to download all of the expanded images at once.
You can choose the extension when downloading, but please save it as [jpeg] or [png] so that it can be loaded into the video generation AI later.
STEP④：Generate a video
<Tools used>Pixverse (or Kling AI)
In this step, we generate a video that evolves from the old model to the new model based on the images created in the previous step.
As mentioned in the tool explanation, there are two AI services that can be used to generate this video: Pixverse and Kling AI. However, for stability reasons, we will use Pixverse in this explanation.
Once you get the hang of it with Pixverse, you should be able to generate the same results with Kling AI.
Pixverse function explanation
First, open Pixverse.
There are menus on the left side of the screen, so select [Transition].
Transition mode is a video generation mode that does not skip ahead between the first and second images based on the prompt you enter.
Now, when you enter the transition editing screen, the following screen will appear.
Some parts may be difficult to understand for those using video generation AI for the first time, so we will explain each item.
(1) Model
Think of the model as the engine for the generative AI.
Changing this can drastically change the generated results, even with the same prompt.
Please select [v4.5] for the model.
At the time of writing this manual, v5 is the latest version, but the car evolution video generator may not output the desired video, so we recommend v4.5!
(2) Time
You can choose the video length from 1 to 8 seconds.
Five seconds is just right if you want to express a sense of tempo and cool transformation motions.
(3)Resolution
You can choose the resolution of the video you generate.
The highest resolution, 1080P, is only available with a Pro plan subscription. (720P requires a Standard plan subscription.)
Low quality can limit your reach on video platforms like Instagram, so we recommend choosing the highest quality possible.
However, the higher the resolution, the higher the required cost. 💰
(4)Preview Mode
It is possible to check the video generation results in low quality at low cost.
It is recommended for use in experiments with new prompts, etc.
(5)Auto Sound
You can turn this feature on or off as you like. (Turning it on incurs an additional cost.)
⚠️However, there’s one thing to note: the sound effects added by this feature will only be heard in your left ear when listening with headphones.
If you want to avoid this, it’s best to turn it off and add the sound effects yourself when editing the video.
You can also choose not to add sound effects to the video itself.
(6)Auto Speech
It adds human voices 🗣️
It’s not necessary for this video, so please turn it off.
(7)Off-Peak Mode
Available to those with the Pro plan or above 🛌🏻
Turning this on significantly increases the time it takes to generate a video, potentially several hours, but it also reduces the generation cost by 30%.
This is a huge cost savings for those who generate a large number of videos, so I usually keep it on.
(8)Create Count
Specify how many videos to generate per generation.
Increasing the number of videos generated does not significantly change the results, and increasing the number to two or three simply doubles or triples the cost, so we strongly recommend selecting [1].
(9)Seed
It’s like a random number generator, and if the numbers change, different results will be output even if you use the same image and prompt.
Basically, it doesn’t matter whether you tweak it or not, so you can just leave it as it is.
Actually generate the video
Now that you understand each of the settings above, let’s actually generate a video 🎞️
Video generation is done at the bottom of the screen in transition mode.
As shown in the image below, upload the image of the car before the transformation as the first image and the image of the car after the transformation as the second image.
Enter how you want the car to transform in the prompt.⌨
It might be a good idea to come up with the prompts yourself from scratch and have Google Translate translate them, or to have an AI such as Google AI Studio or chatGPT come up with them.
Of course, you can also copy and paste from my bonus prompt collection!
Once you have entered all of the information, press the [Create] button at the bottom right to create the video.
Once the video is complete, a thumbnail will appear in Pixverse’s transition mode as shown in the image above, and you can simply download it from there.
Once you have done this for multiple models, this step is complete!
STEP ➄: Video editing
<Tools used>Video editing software
Once you get this far, you’re almost there!
Use video editing software to finish the video generated in STEP ④.
This time, I will be explaining using Davinci Resolve, which is the software I use.
It may be difficult to understand if you are unfamiliar with editing software, but don’t worry, we will only be using basic functions that are found in any software.
Here’s what I do in this step:
・Video splicing
・Vehicle specifications
・Sound effects
Splicing videos is a must, but feel free to do anything else as needed.
If you want to add a description or make your own personal differentiation, then go ahead and do so.
Tips when editing
Arrange the videos generated using the steps above on the same timeline as shown in the image below.There are two ways to improve the tempo.
(1)Video speed is 1.2 times faster
To increase the sense of speed and tempo, we recommend speeding up the video by about 1.2 times.
(2)Shorten each video by [-0.05 seconds]
If you simply splice together the videos generated by Pixverse, the car will stop at the joints between the videos.
It seems that Pixverse transition mode videos are designed to freeze like a still image at the end.
Therefore, by shortening each cut slightly 【-0.05 seconds】, you can create a video with smooth movement even at the transitions between cuts.
STEP➅：Post
If you’ve gotten this far, you’re 95% done.
All that’s left to do is select background music, write a summary, and post your video on a video platform!
When you go to the reel video posting screen, there is a [♫] button at the bottom of the screen. Click on it to add music.
Any music can be inserted, but from my experience, I recommend the following.
・Commercial songs related to the theme car model
・BGM for media in which cars play a central role, such as 『Fast and Furious』.
・Otherwise, up-tempo music
All you have to do is add a description in the description section.
Only the first line of a reel video description is displayed during playback, so as long as you can concisely describe the content of the video, it should be fine.
This is because almost no one will bother to click on the description section to check it out.
I only wrote about the “Evolution of (car model name)”.
In fact, I didn’t even use a hashtag. (Instagram has officially stated that using hashtags is meaningless in today’s world.)
⚠️Note⚠️
At the bottom of the posting screen, there is an option called “Add AI Labels.” Since this is an AI-generated video, be sure to turn this on.
All that’s left to do is press the share button and you’re done!
Special favor
We’ll give you a list of 23 prompts you can use on Pixverse.
The outline of the transformation and its prompts are written in an Excel sheet, so please copy and paste it.