Hi everyone, this is Sister Orange, who uses AI first whenever there's a problem.
It blew up, it blew up.
The entire internet, and even the global creator community, has been blown away by ByteDance's "new toy"—Seedance 2.0.
From its launch on Jimeng on February 7 for small-scale beta testing, the topic shot straight to Weibo's trending searches on February 9, and in the following days, discussions, reviews, and stunning examples about Seedance 2.0 swept across all content platforms.
Netizens said: "Last year DS, this year SD," "China's Sora is here," "Directors are losing their jobs, editors are being laid off"...
Why exactly is Seedance 2.0 so wildly popular?
You can think of Seedance 2.0 as a "super brain" with a top-tier team of directors, cinematographers, editors, voice actors, and composers.
The core is just one thing:Turn the vague ideas in your head, or the fragmentary materials at hand, into a decent-looking short film。
Seedance 2.0 supportstext, images, videos, audio, and you can throw them all in as references.
Seeing this, don't you really want to try it out too? There are three free trial entry points; those interested can go try them in advance.
01. Free Trial Entry Points
1. Xiaoyunque
The first entry point isXiaoyunque Web Version。
You need to download the mobile app first, then log in to the web version to use it.
Search for "Xiaoyunque" in your phone's app store to download the app, register, and log in.
Open the Xiaoyunque official website on your computer:https://xyq.jianying.com/，点击网页版，用APP注册的手机号登录。(移动端也支持体验了）
SelectImmersive Short Film Generation Mode, that's Seedance 2.0.
You can only use it 3 times. If you have multiple accounts, that's x3 times, you know what I mean.
2. Volcano Engine - Experience Center
Open the Volcano Engine official website:https://www.volcengine.com/
Log in and register, select AI Experience Center, select vision models, and set the model toDoubao-seedance-2.0and you're done.
3. Jimeng
Open the Jimeng official website:https://jimeng.jianying.com/
Select video generation, and see if there's a Seedance 2.0 option. If not, you haven't been included in the rollout.
There's a workaround worth trying.
Jimeng hasa limited-time offer of a 7-day membership trial for 1 yuan.
Buying this might let you experience Seedance 2.0. (It's now online, but generation requires queuing)
You can't lose or be fooled for 1 yuan. If you buy it and it's not there, or you don't want to use it later, quickly turn off auto-renewal.
I think Xiaoyunque is the most convenient.
02. Hands-on Video Creation
1. Food Video Production
Open the Xiaoyunque web version https://xyq.jianying.com/
I found a picture of braised prawns online, uploaded it as a reference image, entered the prompt below, and selected 15s.
Let's see the result:
I didn't upload audio or video references, but the generated result included human voiceover, and it was a very fitting deep, rich male chef's voice.
Maybe the prompt I entered was too simple—at the start of the video, the prawns are almost cooked... haha
Still, I was quite surprised.
2. Turning Everyday Photos into Videos
I found a photo of a snow penguin made during Beijing's first heavy snowfall, and a photo of my son climbing a tree, and linked them together.
Enter the prompt below
This time the prompt is written a bit more richly.
Let's see the result:
Likewise, the background sound effects fit the scene very well, magical yet playful. If the child in the back were also moving when the penguin comes to life, it would be even more perfect.
It seems the prompt still needs practice and learning.
But for ordinary people, this kind of result is already very stunning.
3. Product promotion/good finds sharing videos
When everyone is scrolling through good finds sharing, good finds unboxing, and other videos, they will more or less be persuaded to buy one or two items.
Everyone has items they like/love to use in life, and also the desire to share, but it stops at the cumbersome shooting and editing.
So now, try using Seedance 2.0~
I found a few product images and used AI to generate a photo of a beautiful woman to create a good finds sharing video.
Let's see the result:
Isn't that done? From now on, you can totally use your own photos to generate videos.
Although it currently only supports 15s, then just generate a few more 15s clips and splice them together with any editing tool.
The higher the sharing frequency and the richer the content, are you still afraid the account won't take off?
Believe that great effort produces miracles.
03. Final words
The popularity of Seedance 2.0 is backed by several solid technological breakthroughs
It adopts the "dual-branch diffusion transformer" architecture. You can understand it as its brain being divided into two parallel departments: the video department and the audio department. From the first second the instruction is issued, these two departments start working simultaneously and collaborate closely.
It solves:character consistency and physical coherenceissues. In multi-shot scenes, the protagonist's face will not change inexplicably, and the patterns on the clothes will not flicker.
Seedance 2.0's "director mindset". It can understand the logic and timing in the long prompt you input, automatically break it down and plan it into a series of related shots.
Seedance 2.0 has raised the "usability rate" of AI video generation from an industry average of less than 20% to over 90%. AI video truly has the possibility of industrial application.
A person with ideas and stories can use AI to present their visual universe.
For ordinary people like us, it is redefining the boundaries of "creation" and "expression." Video will no longer be the privilege of professionals; like text and images, it will become part of everyone's daily expression.
I heard it will be fully opened soon, and will also be available on national apps like Doubao, popularized among the masses and entering thousands of households.
Original link:ByteDance AI video "new toy" Seedance 2.0, it's exploded