Hello all,
Thank you for being part of the Sounding Video Generation Challenge 2024! Whether you finished at the top or explored the task for the first time, we’re glad you joined us in this first edition and pushed the boundary of research.
We’re excited to share the winners of this challenge (see below) — but before you scroll, we want to invite you to be part of what comes next:
If you built something you’re proud of, learned something unexpected, or just have thoughts on how the challenge went, we’d love to feature your story.
-
Post a short reflection or summary of your solution in the challenge thread, or share it on LinkedIn using #SoundingVideoGen and tag @AIcrowd. We’ll highlight selected submissions across our channels.
-
And if you have 2 minutes, we’d be grateful if you filled out this quick feedback form. This was a first challenge of its kind, and your feedback will directly shape how we run the next one.
A message from the organisers
We sincerely appreciate your participation in the Sounding Video Generation Challenge 2024! We are excited to announce the top three submissions for each track and would like to congratulate the top-ranked participants on their excellent achievements.
Please note that the rankings are based on ratings collected from the Amazon Mechanical Turk platform, rather than our personal preferences. The evaluation process took longer than anticipated, and we appreciate your understanding and patience during this time.
We would like to extend our heartfelt thanks to all participants. This is our first attempt to organize a challenge in this research domain, and we were unsure if researchers would be interested. However, many participants submitted outstanding systems/models. We appreciate all of your great work. This challenge has provided us with valuable lessons, and we will apply what we have learned to future challenges/projects. Thank you for your participation and support.
Final Results (based on human evaluations)
Temporal Alignment Track (contact: @Akio Hayakawa)
| Rank | Participant | Submission ID | Score | Prize |
|---|---|---|---|---|
| 1
|
@ni_kai_hua | 281469 | 0.87981 | $10,000 |
2 
|
@OwO | 281046 | 0.87585 | $5,000 |
3 
|
@kartana | 281375 | 0.86340 | $2,500 |
Spatial Alignment Track (contact: @Kazuki Shimada)
| Rank | Participant | Submission ID | Score | Prize |
|---|---|---|---|---|
| 1
|
@nht1990 | 281487 | 13.89 | $10,000 |
| 2
|
@wangzhiyu918 | 281316 | 13.33 | $5,000 |
| 3
|
@kartana | 281166 | 13.18 | $2,500 |
Each track used a different metric to rank the submissions.
Let’s keep this going
Whether or not you landed in the top ranks, your journey can inspire others. We’d love to hear about what you built, how you approached it, or what you learned along the way.
Drop a comment in the challenge thread
Or share your reflections on LinkedIn using #SoundingVideoGen and tag @AIcrowd, we’ll reshare highlights with the community.
And if you have just 2 minutes, please fill out this quick feedback form. It’s short, but your input will go a long way in helping us make future challenges even better.
Thanks again for being part of this journey. We can’t wait to see what you do next.