AI image & AI video news and information delivered by Mintbear. Only the news you shouldn't miss. With Mintbear's insights. Easy and useful.
All
AI Films
Sora
Gen-3
Kling
Hello
Hunyuan
Luma
Pixverse
Flux
NVIDIA
Mintbear
AI LLM
OpenAI
Krea
Gen-4
[Jocoding x Mintbear] AI Image Special Lecture
AI Image
AI Video
AI Sound
AI Films
Mintbear
Sora
Gen-4
2025/04/08
mintbear
Video Production Tools Comparison: Kling & Hailuo & Luma & Pixverse & Sora
Kling
Hello
Sora
Luma
Pixverse
2025/03/10
mintbear
1
Workflow Study Zoom Meeting (Public)
Updates
Mintbear
2025/02/20
mintbear
Krea Chat: Images & Videos Created Through Conversation
AI Image
AI Video
Updates
Krea
AI Films
Flux
Hello
Kling
Luma
2025/02/08
mintbear
Deep Research using ChatGPT 'o3' model
Updates
AI LLM
OpenAI
2025/02/06
mintbear
2025 AI Era Human Intelligence Conference (On-site Sketch)
AI Video
AI Films
Mintbear
2025/01/12
mintbear
2
2025 AI Era Human Intelligence Conference (Mint Bear Lecture)
AI Video
AI Films
Mintbear
2025/01/12
mintbear
Sora is Open!
AI Video
Sora
AI Films
The long-awaited Sora-nim has arrived. As expensive as it is, the realistic simulation and presentation are excellent. I am looking forward to various functions. However, in I2V, there are some parts that are slightly different from the presentation of the reference image. I will continue to test it.
2024/12/10
mintbear
Sora-01
Sora In London
AI Video
Sora
2024/12/09
mintbear
Hunyuan Video by Tencent
AI Video
Hunyuan
2024/12/07
mintbear
Hailuo I2V-01-Live
AI Video
Hello
2024/12/04
mintbear
Gen-3 Video Keyframing (Prototype)
AI Video
Updates
Gen-3
2024/12/03
mintbear
Fugatto: NVIDIA-generated AI heralds audio revolution
AI Sound
NVIDIA
2024/12/01
mintbear
Korea AI International Film Festival Grand Prize Winner, Mateo Online Screening (Event ended / Please wait for the next screening)
The long-awaited Sora-nim has arrived. As expensive as it is, the realistic simulation and presentation are excellent. I am looking forward to various functions. However, in I2V, there are some parts that are slightly different from the presentation of the reference image. I will continue to test it.
On 12/10/2024 at 12:00 AM, I subscribed to a $200 Sora Pro account.
(I will skip the story of the time I worried about the cost, as worrying only delays the process.)
Menu
For a detailed menu, please refer to the well-organized posts by Traveler J and James. (It's in the Reference section below.)
Speed
Actually, I tried generating min/max video I2V and the speed is good.
•
Minimum size: 480p * 5s * 1 = about 2 minutes
•
Max size: 1080p* 10s* 1 = about 6 minutes
Length
For Plus Plan ($20): Only 5 seconds are generated in 480p and 720p resolution.
Based on Pro Plan ($200): Videos in 480p and 720p resolution can be created in 5, 10, 15, and 20-second lengths. However, videos in 1080p can only be created in 5 and 10-second lengths. Videos longer than 15 seconds in 1080p cannot be created.
And the one-minute video creation feature that has been promoted since February is still not on the menu.
Creation results
Original Image (Midjourney)
Let's compare Sora and Hailuo videos using images generated by Midjourney.
Sora Video (I2V)
🍀🧸 : Why are you coming back from behind.. Why are you laughing and passing by..
Sora clearly seemed to understand the style, structure, and components of the original image.
But instead of starting the video from a reference image, I created a new movement and story. The bag changed too. It was very realistic and fresh, but I couldn't control it.
The reason I say this is because Hailuo recognizes the inserted image as a keyframe and uses it explicitly as the starting point of the video. Compare below.
Hailuo Video (I2V-01-Live)
Hailuo is a great director who has done a great job of producing the direction I wanted, and all the information included in the image reference. This is common sense production.