[Brainstorming] How do you think AMD would implement the Frames Generation & FSR 3?

[ryzenmaster]

Hey fellow tech enthusiasts,

I wanted to initiate a discussion about AMD's potential implementation of frame generation in their upcoming FidelityFX Super Resolution (FSR) 3 technology. FSR has already made waves in the gaming community by providing an alternative to NVIDIA's DLSS (Deep Learning Super Sampling), and with the advent of FSR 3 on the horizon, it's intriguing to ponder how AMD might take this technology even further.

FSR has been a game-changer in terms of performance optimization and image upscaling, leveraging various rendering techniques to enhance gaming experiences without significant hardware overhead. With FSR 3, I believe AMD will continue to refine and enhance their frame generation methodology.

Here are a few possibilities I've been considering:

1. AI-Based Approach: Following NVIDIA's footsteps with DLSS, AMD might incorporate machine learning algorithms to generate frames in FSR 3. By training a neural network on high-quality frames and using it to predict missing information in lower-resolution frames, AMD could potentially achieve superior image upscaling and preservation of detail.
2. Temporal Techniques: Building upon their existing FSR technology, AMD could further explore temporal techniques in FSR 3. This approach involves analyzing multiple frames in a sequence to generate a higher-quality output frame. Techniques such as motion estimation and frame interpolation could be employed to improve image clarity and reduce artifacts.
3. Advanced Upscaling Algorithms: AMD may also focus on refining their upscaling algorithms in FSR 3. By utilizing advanced mathematical techniques and image processing algorithms, they could enhance the accuracy of frame generation while minimizing the loss of detail. This could result in sharper, more realistic visuals in games.
4. Frame Generation (AMD Fluid Motion): To cater to the diverse preferences and hardware capabilities of gamers, AMD could for example introduce customizable frame generation (Fluid Motion) options in FSR 3. This could enable users to adjust the trade-off between performance and image quality (error correction per generated frame) according to their specific requirements. AMD may opt to integrate AI cores within the cutting-edge RDNA 3 architecture to bolster the computational capabilities of FSR 3 or other graphically intensive operations such as Fluid Motion. This strategic decision would harness the power of dedicated artificial intelligence processing units to amplify the overall performance and efficiency of these tasks, thereby empowering users with unparalleled visual fidelity and immersive gaming experiences.

Of course, these are merely speculations and it's difficult to predict precisely how AMD will implement frame generation and FSR 3. It's worth noting that AMD has always prioritized open-source technologies, so we might see an emphasis on community-driven development and collaboration to ensure the widespread adoption and success of FSR.

What are your thoughts on this? How do you think AMD will approach frame generation in FSR 3? Do you have any additional ideas or theories? Let's discuss and share our excitement for the future of gaming visuals!

 

[crazyeyesreaper]

Hey fellow tech enthusiasts,

I wanted to initiate a discussion about AMD's potential implementation of frame generation in their upcoming FidelityFX Super Resolution (FSR) 3 technology. FSR has already made waves in the gaming community by providing an alternative to NVIDIA's DLSS (Deep Learning Super Sampling), and with the advent of FSR 3 on the horizon, it's intriguing to ponder how AMD might take this technology even further.

FSR has been a game-changer in terms of performance optimization and image upscaling, leveraging various rendering techniques to enhance gaming experiences without significant hardware overhead. With FSR 3, I believe AMD will continue to refine and enhance their frame generation methodology.

Here are a few possibilities I've been considering:

1. AI-Based Approach: Following NVIDIA's footsteps with DLSS, AMD might incorporate machine learning algorithms to generate frames in FSR 3. By training a neural network on high-quality frames and using it to predict missing information in lower-resolution frames, AMD could potentially achieve superior image upscaling and preservation of detail.
2. Temporal Techniques: Building upon their existing FSR technology, AMD could further explore temporal techniques in FSR 3. This approach involves analyzing multiple frames in a sequence to generate a higher-quality output frame. Techniques such as motion estimation and frame interpolation could be employed to improve image clarity and reduce artifacts.
3. Advanced Upscaling Algorithms: AMD may also focus on refining their upscaling algorithms in FSR 3. By utilizing advanced mathematical techniques and image processing algorithms, they could enhance the accuracy of frame generation while minimizing the loss of detail. This could result in sharper, more realistic visuals in games.
4. Frame Generation (AMD Fluid Motion): To cater to the diverse preferences and hardware capabilities of gamers, AMD could for example introduce customizable frame generation (Fluid Motion) options in FSR 3. This could enable users to adjust the trade-off between performance and image quality (error correction per generated frame) according to their specific requirements. AMD may opt to integrate AI cores within the cutting-edge RDNA 3 architecture to bolster the computational capabilities of FSR 3 or other graphically intensive operations such as Fluid Motion. This strategic decision would harness the power of dedicated artificial intelligence processing units to amplify the overall performance and efficiency of these tasks, thereby empowering users with unparalleled visual fidelity and immersive gaming experiences.

Of course, these are merely speculations and it's difficult to predict precisely how AMD will implement frame generation and FSR 3. It's worth noting that AMD has always prioritized open-source technologies, so we might see an emphasis on community-driven development and collaboration to ensure the widespread adoption and success of FSR.

What are your thoughts on this? How do you think AMD will approach frame generation in FSR 3? Do you have any additional ideas or theories? Let's discuss and share our excitement for the future of gaming visuals!

I have a feeling AMD will utilize MCFI otherwise known as motion compensated frame interpolation. Using 2 frames to generate a new frame. Typically TVs can already do this however, driver level in game I could see AMD using this at lower frame rates then disabling at higher frame rates. It wouldn't require much compute and would improve bottom line performance by making the game playable appear smoother and with less judder. However, at higher frame rates I imagine there would be more artifacts due to the amount of new frames being generated. That is just my guess based on the fact AMD's equivalents to NVIDIA have been industry standard techniques thus far.

 

[Space Lynx]

I have a feeling AMD will utilize MCFI otherwise known as motion compensated frame interpolation. Using 2 frames to generate a new frame. Typically TVs can already do this however, driver level in game I could see AMD using this at lower frame rates then disabling at higher frame rates. It wouldn't require much compute and would improve bottom line performance by making the game playable appear smoother and with less judder. However, at higher frame rates I imagine there would be more artifacts due to the amount of new frames being generated. That is just my guess based on the fact AMD's equivalents to NVIDIA have been industry standard techniques thus far.

is this what youtube does with 60 fps videos? when I watch a yuotube video and change it to 60 fps, it looks insanely smooth... but regular 60 fps gaming doesn't look that smooth.

 

[eidairaman1]

I say go work for them for a bit and get back to us