Show HN: Sora Videos – Generated by the Sora AI model, currently with 107 (soravideos.media)

1 points by yangxiaobo 2y ago ↗ HN

一、Website Features The website has compiled basically all the videos currently generated by Sora (some videos are not fully displayed because they are just regenerated by changing some parameters, and there are many similar parts). Currently, there are 107+ videos, and newly generated videos will be updated in time.

1、Timely updates: New videos generated using Sora will be updated on the website in a timely manner, currently it should be the most up-to-date and comprehensive;

2、Video categorization: Videos are categorized according to the way they are generated: text-generated videos, image-generated videos, video-to-video stitching parts, and parts where there are some defects in the currently generated videos. This allows for a more convenient experience of its different capabilities;

3、Includes prompt words: Videos with prompt words also display the corresponding prompt words under the video, and you can try using the same prompt words in other video generation software;

4、Source display: The source of each video is displayed. Most of the early videos were released by the official openAI, so the source and time are the first time Sora was released; subsequent videos were released by the main creators on Twitter;

5、Display release time: Each video has a corresponding release time, sorted in reverse chronological order, so you can see the latest videos every time you come in;

Everyone is welcome to experience online: https://soravideos.media

If you have any ideas or suggestions, you are also very welcome to leave a message.

1 comment

[ 8.7 ms ] story [ 9.5 ms ] thread

yangxiaobo 2y ago ↗

On February 16, openAI launched the Sora model, which seems to have revolutionized previous video generation technology. Whether it's the duration of the generated video (it can generate videos up to 1 minute long, while previous video generation software was still struggling with 4s, 16s timelines), or the consistency of the video (it has solved the problem of inconsistent content and flickering in previous generated videos), there has been a qualitative leap. Especially in terms of understanding the physical laws of the real world, it feels like a technology that can directly generate videos from text, rather than stitching videos together in the form of generating keyframes of pictures like previous models. Everyone is very curious and enthusiastic about everything about Sora, but since it is not yet open for use, we can only see some of the effects of videos generated by Sora. However, the videos are scattered everywhere, making it inconvenient for unified learning and viewing.

So I developed this website.