What is DLSS 3.5? how it works and the difference with dlss 3
Graphic technology continues to grow rapidly, especially in the realm of gaming and 3D rendering. One of the latest innovations from NVIDIA is DLSS 3.5, the latest version of technology Deep Learning Super Sampling (DLS) was first introduced in 2018.
Then, what exactly? DLSS 3.5, how does it work, and what is the difference with DLS 3 previously present on the RTX 40 series graphics card?
This article will discuss in full about the meaning, mechanism of work, and the advantages of DLSS 3.5 compared to its predecessor with an easy-to-understand language style.
What is DLSS 3.5?
DLSS 3.5 is the latest version of technology Deep Learning Super Sampling artificial NVIDIA, which uses artificial intelligenceAi) and Machine Learning To improve the quality of graphics and performance in games and professional graphics applications.
Unlike the previous version, DLSS 3.5 introduced a new feature called Ray Reconstruction, which is designed to improve the quality of lighting (Ray Tracing) to be more realistic, smooth, and noise free.
According to NVIDIA, DLSS 3.5 focuses not only on improving FPS performance (frame per second), but also on improving Fidelity Visual, namely the sharpness, detail, and accuracy of the reflection and shadow in the game.
a glimpse of the development history
Before understanding DLSS 3.5, it is important to know how this technology evolved:
| DLSS version | Release year | Main features | Note |
|---|---|---|---|
| DLSS 1.0 | 2018 | Simple AI Upscaling | Visual quality is still not consistent |
| DLS 2.0 | 2020 | Temporal AI + Improved Details | Sharper and more stable results |
| DLSS 3.0 | 2022 | Frame Generation | significantly increase FPS |
| DLSS 3.5 | 2023 | Ray Reconstruction | Focus on the quality of visual ray tracing |
From the table above, it can be seen that each generation brings significant improvement. When DLSS 3 focuses on Frame Generation, then DLSS 3.5 focuses more on Improved Visual Ray Tracing Quality.
How DLSS 3.5 Works
Technically, DLSS uses a neural network (Neural network) trained by NVIDIA supercomputers. The goal is to “predict” high resolution images based on low resolution images of the game being run.
In DLSS 3.5, there are three main components:
1. Super Resolution
This component has been around since DLSS 2.0 and serves to increase the resolution of the image without burdening the GPU. With AI techniques, low resolution images can be upgraded to high resolution like 4K without losing important details.
2. Frame Generation
This feature was introduced in DLSS 3.0. He is able to create Additional frames Among the original frames generated by the GPU. That way, the game can run more smoothly, even above 100 fps, even though the hardware is not very strong.
3. Ray Reconstruction (new feature in DLSS 3.5)
This is the most prominent feature in DLSS 3.5. Ray reconstruction replaces some algorithms denoising traditional with a smarter AI model.
AI is able to analyze information from various sources—such as object position, direction of light, and reflection—to create results Ray Tracing more realistic and visual distractionArtifact).
According to a report from TechRadar, the visual results of DLSS 3.5 make shadows and reflections look much more natural, especially in games that use many lighting elements such as Cyberpunk 2077: Phantom Liberty.
difference between dlss 3.5 and dlss 3
Many think that DLSS 3.5 is just a minor update, but there are a number of important differences that have a big impact on the visual results and device compatibility.
| Aspect | DLS 3 | DLSS 3.5 |
|---|---|---|
| Main features | Frame Generation | Ray Reconstruction |
| main focus | increase FPS | improve the quality of visual ray tracing |
| GPU compatibility | RTX 40 Series (No Lovelace) | RTX 20, 30, and 40 series |
| Performance in non-ray tracing games | Optimal | same as DLSS 3 |
| Visual Ray Tracing | still dependent on Denoiser traditional | More Accurate and Natural with AI Ray Reconstruction |
One of the big advantages of DLSS 3.5 is wider compatibility.
If DLS 3 can only be used by RTX 40 series graphics card users, then DLSS 3.5 can also be run at RTX 20 and RTX 30 Series.
This is certainly good news for gamers who have no plans to upgrade the GPU.
Advantages of DLSS 3.5
Here are some of the main advantages that make DLSS 3.5 interesting:
1. Visual quality is more realistic
With ray reconstruction, light, shadow, and reflection effects look smoother and more natural, even in complex lighting environments.
2. Compatible with old GPUs
As mentioned above, DLSS 3.5 is not exclusive to the RTX 40 series. This means that users of RTX 2060 or RTX 3060 can enjoy improving visual quality without having to change the GPU.
3. more consistent performance
DLSS 3.5 uses AI to balance visual quality and performance, so that the frame rate remains stable without sacrificing details.
4. Improve rendering efficiency
In applications such as Blender or NVIDIA OMNIVERSE, DLSS 3.5 helps the rendering process to be faster, making it also suitable for 3D designers and professional animators.
examples of games that support dlss 3.5
According to NVIDIA, some games and applications that have or will support DLSS 3.5 include:
Cyberpunk 2077: Phantom Liberty
Portal with RTX
Alan Wake 2
D5 rendering
NVIDIA OMNIVERSE
games like Alan Wake 2 It even makes DLSS 3.5 one of the excellent features in showing off realistic cinematic lighting.
Can DLSS 3.5 increase FPS?
Technically, DLSS 3.5 does not directly increase FPS Like DLSS 3 because the main feature is Ray Reconstruction, not Frame Generation.
However, because the visual results are more efficient and NoiseIt is reduced, the GPU can work lighter in some situations, so a small performance increase is still possible.
In other words, DLSS 3.5 focuses more on image quality than raw performance. If you want to increase FPS, DLSS 3 remains ahead. But if you want the most realistic graphics with the best light effects, DLSS 3.5 is the answer.
DLSS 3.5 challenges and limitations
No technology is perfect, including DLSS 3.5. Some of the limitations include:
Not all games support DLSS 3.5 — Developers need to manually integrate the Ray Reconstruction feature.
The effect of “ai sharpening” is sometimes excessive, makes some textures look too sharp.
Performance depends on the latest drivers and optimization of NVIDIA.
However, when viewed from the direction of development, DLSS technology will continue to be improved to overcome these things in the future.
DLSS technology future
With the presence of DLSS 3.5, NVIDIA shows its ambition to make AI the backbone of the entire modern graphic process.
According to Digital Foundry, in the future, it is not impossible that DLSS will completely replace Traditional Rendering Pipeline, because the efficiency is much higher.
We can also hope that the next DLSS version—maybe DLSS 4.0 — will combine Frame Generation and Ray Reconstruction which is more adaptive and able to work on various platforms, including mobile devices and cloud gaming.
FAQ about DLSS 3.5
1. Is DLSS 3.5 only for RTX 40 series?
No. DLSS 3.5 supports RTX 20, 30, and 40 series, especially for ray reconstruction features.
2. Does DLSS 3.5 increase FPS?
not directly. The focus is on improving visual quality through ray reconstruction.
3. What is the main difference between DLSS 3 and DLSS 3.5?
DLSS 3 Focus on Frame Generation for high FPS, while DLSS 3.5 focuses on Ray Reconstruction For more realistic visuals.
4. What games already support DLSS 3.5?
Some of them are Cyberpunk 2077: Phantom Liberty, Alan Wake 2, and Portal with RTX.
5. Does DLSS 3.5 require a special driver update?
Yes, users need to install the latest drivers from NVIDIA so that DLSS 3.5 can function optimally.
Conclusion
DLSS 3.5 is NVIDIA’s big step in delivering realistic visuals through artificial intelligence.
If DLSS 3 focuses on improving FPS performance, then DLSS 3.5 brings great improvement in terms of lighting quality and ray tracing.
With support for the previous generation RTX GPU, DLSS 3.5 is an interesting solution for gamers and creators who want to enjoy the best visual quality without having to upgrade hardware on a large scale.
This technology is a clear proof that The future of computer graphics will be highly dependent on AI and machine learning, bringing the visual experience to a level that is getting closer to reality.
