The New Lossless Scaling 2.0 Plugin Doesn't Make Frame Generation Better

まだの方は、X(Twitter)、Mastodon、Blueskyでフォローし、新しい記事が投稿されたらお知らせください。また、Redditに参加したり、YouTubeに登録して、私たちのコンテンツをチェックしたり、素晴らしいコミュニティと交流してください。応援ありがとうございました！

With all the tools at our disposal to improve our experience on the Steam Deck, one of the most talked-about is frame generation. More specifically, people talk about lossless scaling and using it to artificially double frame rates on the Deck. There are some major drawbacks to using it, which is why we don’t recommend it usually, but it still persists in conversations about performance for games on the Deck.

The most common way to get it running is through a Decky plugin, and there are a few created to facilitate it (like Decky-lsfg). However, over the past week, I have been seeing posts on Reddit about a magical lsfg-vk 2.0 going around, which is essentially a new plugin that utilizes the developmental build of lsfg-vk, a fork of lossless scaling frame generation that works with Vulkan/Linux. However, it has been billed as a major upgrade, reducing ghosting, adding an FP16 mode for increased FPS, and more. But a lot of these require major disclaimers, and this plugin isn’t what it’s made out to be.

The creator of lsfg-vk, PancakeTAS, responded to the post, debunking a lot of the information that was published in it. First, the reduced ghosting was a bug that was actually fixed in the development build of lsfg-vk 1.0, and it only fixes certain ghosting issues. It does not improve image quality. On top of that, the author didn’t introduce FP16 acceleration for better performance. This already exists in the original plug-in, but it was hidden because turning it off makes no sense and will only harm the experience.

Pancake also mentioned that some of the claims on the GitHub page were completely fabricated. This includes the fact that the reworked Vulkan backend and VRAM handling are not available to the public at all, since they require a file that doesn’t exist to the public, and there is no frame-pacing work that is magically making the game run smoother.

Just to see ourselves, we did take a look and tested the current Decky-lsfg and this experimental version, and while there are some ghosting improvements, it is exactly the same otherwise. We still have the same input lag issues; performance is about the same, and it doesn’t feel much different.

I also took a look at the plug-in itself. I don’t see any mention of AI coding, but I wouldn’t be surprised if there is some AI involvement because a lot of the recent projects from the creator of the experimental plugin either work with generative AI or use CLAUDE. So, while I can’t confirm that this uses AI from my own knowledge, it is an accusation that should be taken seriously, especially since it’s from the creator of lsfg-vk.

Ultimately, this doesn't change my views on lossless scaling, nor does it make some major improvements over the more popular version right now. Apart from the ghosting improvements through the bug fix on the development branch, nothing else has been changed. I would be wary about using lossless scaling on the Steam Deck in general, but this new "2.0" plugin won't be making any huge differences to the experience.

この記事を楽しんでいただけたなら、SteamDeckHQ の他のコンテンツもチェックしてみてください！あなたのゲーム体験に役立つゲームレビューやニュースを幅広く取り揃えています。ニュース、ヒントやチュートリアル、ゲーム設定やレビュー、最新のトレンド情報など、あらゆる情報をお届けします。 カバーしています！