Adobe's "new course" is here: Teach You a Sentence P Video.
Back then, Premiere Pro, Adobe's video editing tool in the family bucket, was a regular guest on college electives, and for this reason, many students temporarily crammed at station B, and then used it to forget all their knowledge.
The videos that have accumulated dust in your favorites haven't been thrown away, so it's better to start learning how to command the AI to conquer the PR now, and it's still too late to go from beginner to proficient.
Change a sentence to a video, Sora and Pika are all at your service
Adobe released a big event yesterday: a PR enhanced by generative AI, which is scheduled to roll out within this year.
Unfortunately, the exact time has not yet been announced, but the official demo is already eager to get started immediately.
After all, the effect looks too intuitive. Enter text prompts to add, change, and delete objects, add auxiliary shots, and even extend a shot by a few frames.
There aren't enough diamonds traded underground in gangster films, it seems that the crew is poor and the actors are not forced? It doesn't matter, use the pen tool to select the box, enter the prompt "a bunch of big diamonds on a black velvet cloth", and the AI can create something out of nothing.
In the same way, the set can also be "not physical if it is not necessary", add a vase on the table, and hang a painting in the study.
Conversely, redundant props can be easily removed, and the signs on the wall and the bags on the ground can disappear in place with one click. If the microphone or light stand accidentally gets into the lens, it can also be quickly remedied. Don't choose where to choose, you can save snacks in the field and later stage.
An actor's make-up can also be the icing on the cake, changing the brand to the watch and putting a tie on the shirt.
Even the length of the video can be changed at will, and if the actress's close-up isn't long enough, click and drag to the right position at the beginning or end of the clip, and the AI can help extend it by a few frames.
To implement these AI capabilities, Adobe is not only developing its own video models, but also integrating third-party video models in PRs, giving users the freedom to choose the one that performs better.
Runway, Pika, OpenAI's Sora, and the most famous sweets in the AI video track, have all been included in Adobe's cooperation.
Judging by the current demo, several third-party models have their own strengths. Pika is primarily used to extend videos and make scene transitions smoother.
Sora and Runway, which are used to generate secondary shots, i.e. B-rolls, for scenes such as "City Skyline View in the Rain at Night" and "Driving Through the City at Night from the Driver's View".
Letting the AI take care of the difficult empty shots can not only make the rhythm of the whole video better, but also serve the story, portraying the psychological activities of the protagonist with scenery and emotion.
Of course, if you don't want to use the AI function anymore, you can also regret it, because the AI material and the original material are independent, and they are either split in the same timeline or simply in different timelines.
Even though it's just a demo, I still want people to sigh that Adobe is still the same Adobe, when I learned to use PS to pick my hair and PR to pull the speed line, I was painful and happy, and now everything is getting fresher and simpler.
Tools are in their own right, whether it's a trash can or a fire hydrant, whether it's in the city or in the wilderness, it's up to the script and our imagination.
Let the initiative of AI be exchanged for the loyalty of users
Adobe's family of AI tools, called "Adobe Firefly", was integrated into PS in May last year, and this time it's PR's turn.
PS's generative AI function, similar in idea to PR, is to make some details that humans find troublesome easier to deal with through text prompts: add content, remove or replace parts of an image, and expand the image through text......
Before you know it, PS's AI feature, "generative filling", has been integrated into the editorial workflow.
The video team used it to set the scene, such as the ornaments under the TV in this picture, which are all generated by AI, but naturally integrated into the picture, without deliberately speaking, it may not be visible, which really saves props and shooting time.
In the words of the video editor Wang, "It's really easy to use, and it feels magical."
As heavy users of Adobe's family bucket, the design team is naturally also actively exploring new toys. Currently, AI-enhanced PS is more suitable as an adjunct.
For example, to remove an object or a person, generative AI is more efficient than PS's original content recognition fill function, and the edge processing is softer, without distorting the ground, etc.
▲ In order: original image, content recognition filling, generative AI.
The function of expanding the image comes in handy, Party A often gives some materials with insufficient size, and the response at this time does not have to be to pull back and forth in the group @ the other party, and hand it over to AI to get it done, so that the material can be better used.
The AI material similar to PR remains unique, and the generative filling of PS is also a separate layer.
If PR and PS have become more refined because of AI, the efficiency has increased by an order of magnitude, but no, for example, AI-generated pictures cannot be disassembled and fine-tuned, and AI-generated videos still make people play, which has an indescribable AI flavor.
While a lot of the resulting assets can't be used directly, AI is enough to bridge some of the creative and workflow gaps.
A lot of tedious, time-consuming, and boring things, such as cutouts, don't have to be done manually by humans. When customers make a request, AI can also quickly generate a reference effect, reducing the cost of communication with each other.
Previously, in order to add generative AI capabilities to PRs, Adobe conducted a deep survey of user groups to find an answer: How does AI solve problems in their real work?
Adobe has found that there are three core demands of users:
- AI must be integrated into workflows and the tools they use every day;
- AI must edit and process content faster, while still giving users precise control over the content;
- Users have the right to choose about AI, and they can also use AI models with their own strengths.
Judging from the results of this demonstration, Adobe really starts from the needs of users, even if it is not as impressive as Sora at first glance, but the crowd base is strong, and for loyal users, it is attractive enough to solve problems better and faster with only one software.
In fact, in the corners where few people pay attention, Adobe has introduced more AI models.
Firefly Image 2 文生图模型、Firefly 矢量图形生成模型、Firefly 文本到模板设计模型...... Adobe Illustrator 作为全家桶一员，早已沾到了 AI 的光。
Many netizens also left messages in the YouTube comment section of Adobe: When will it be AE's turn?
A veteran creative company that has revolutionized itself and wants to be a clear stream in the AI world
Adobe's situation in the creative field and Google's situation in the search track are inexplicably similar, they are both waiting for a self-revolution because of the arrival of AI.
Many designers, illustrators, and photographers have a negative attitude towards generative AI, feeling that the aesthetics of AI are broken, infringing on human works, and threatening their jobs.
At the same time, Wall Street is not bullish on Adobe. Stocks of Adobe, which fears that its core business will be damaged by AI, have fallen about 20% this year.
Adobe has made its own choice, embracing AI and even actively partnering with AI startups that are making their own lives.
On the other hand, on the side of artists, they only use licensed materials such as Adobe Stock or materials whose copyrights have expired, and provide a compensation system for Adobe Stock contributors to flaunt their professional ethics.
But Adobe isn't always immune.
Some time ago, the media broke the news that part of Adobe's data came from Midjourney, and Adobe responded that this part of the material had been strictly reviewed, and this time Adobe reached a cooperation with Sora, but Sora's training data was also unclear, and even OpenAI CTO said "I don't know" with flickering words, and the on-the-spot reaction was made into a meme.
Adobe doesn't pay a lot of money, even if it's from formal sources, with Bloomberg reporting that images cost between 6 cents and 16 cents per image, and videos cost between $2.62 and $7.25 per minute.
When announcing the generative AI feature of the PR this time, Adobe at least made a clear attitude, emphasizing that each video will be labeled with a "content tag" to tell you that the video has AI involved, and even tell you which model it is.
The controversy has always been there, but it can't be hindered, and the tools have gotten better and faster.
Adobe's speed of entering generative AI is not sharp, and it is not the first in the fields of pictures and videos, but it is relatively steady in its thinking, taking the path of incremental innovation, while directly providing AI functions based on the whole family bucket, and integrating third-party services on the other.
This may be a good instinct for users, and we may not be paying for new, alternative AI products, but we may be paying for productivity software that we are used to and that adds AI capabilities.
From the user's point of view, if PS improves the work efficiency of artists, the Meitu App allows mass users to experience the joy of face slimming with one click, and the AI version of PS and PR empowers both industry insiders and amateurs.
Several Adobe employees told Business Insider that the work of some artists has become working for AI and retouching images for AI. Some advertising agencies are also downsizing their graphic design teams.
Rather than saying that everyone is a creator, everyone is a participant and runner in technology, and there is still a threshold for writing scripts and storyboards, and how to make AI generate ideal pictures and videos has given rise to new knowledge.
At this juncture, the most amazing thing is that companies that serve creativity are also reimagining the creative process to serve users with AI-powered tools without compromising their core strengths. For everyone, and even companies in the pre-generative AI era, there is a mountain waiting to be crossed.