Microsoft is Testing Video Super Resolution Feature in Edge Canary for AMD and NVIDIA GPUs, Upscaling Web Videos Below 720p and Removing Artifacts
News

Microsoft is Testing Video Super Resolution Feature in Edge Canary for AMD and NVIDIA GPUs, Upscaling Web Videos Below 720p and Removing Artifacts

Table of Content

In this article, we’ll explore Microsoft’s latest experimental feature in Edge Canary that promises to revolutionize video streaming for Windows users.

Key Takeaways:

  • Microsoft’s Video Super Resolution (VSR) feature uses machine learning algorithms to enhance video streaming quality for Windows users.
  • VSR removes blocky compression artifacts and upscales video resolution, resulting in clear and sharp videos on streaming platforms like YouTube without sacrificing bandwidth.
  • VSR works on both AMD- and NVIDIA-powered GPUs, but users need at least a GeForce RTX 20 series or Radeon RX5700 to access this feature.
  • Requirements for VSR to work include Windows OS, video resolution less than 720p, device connected to power, video height and width greater than 192 pixels, and no digital rights management (DRM) system.
  • The technology behind VSR uses GPU-agnostic algorithms, ONNX Runtime, and DirectML to make it accessible to more users.
  • Microsoft plans to expand VSR’s availability and supported GPUs in the coming weeks.
  • MSRA’s machine learning models, compressed by Microsoft DeepSpeed, improve video quality and subjective evaluation metrics were developed by the MSRA team.

Microsoft’s Video Super Resolution: Enhancing Video Streaming Quality

Microsoft’s latest experimental feature in Edge Canary, called Video Super Resolution (VSR), promises to revolutionize video streaming for Windows users. 

This feature utilizes machine learning algorithms to enhance the quality of videos viewed in Microsoft Edge. 

VSR eliminates the blocky compression artifacts and enhances the video resolution, giving the audience clear and sharp videos on different streaming sites like YouTube without using much bandwidth.

VSR for AMD and NVIDIA GPUs

Previously, NVIDIA video card owners had the luxury of video upscaling in Chrome or Edge. But, Microsoft is now testing the Video Super Resolution feature in Edge Canary, which works on both AMD- and NVIDIA-powered GPUs. 

Users will need at least a GeForce RTX 20 series or Radeon RX5700 to access this feature, but it could help them watch old YouTube videos or save bandwidth on a capped data plan.

Conditions for VSR to work

To use VSR, certain requirements must be met. Firstly, VSR is only accessible to Windows users. 

Secondly, the video must be played in a resolution lower than 720p, and the device must be connected to a power source. 

Thirdly, the video’s height and width must be greater than 192 pixels. 

Finally, videos with digital rights management (DRM) systems like PlayReady or Widevine cannot be processed by the browser, and frames from these protected videos are not accessible for enhancement.

Hybrid graphics setups on laptops

At present, users with hybrid graphics setups on their laptops will have to force Edge to use their dedicated GPU to access this feature. 

However, automatic hybrid video support is currently in development, according to Microsoft.

Technology behind VSR

The technology behind Video Super Resolution relies on GPU-agnostic algorithms. 

Microsoft had to add a DirectX 12 pipeline to Google’s Chromium engine so that its machine learning framework could interact with the browser. 

This compression helps keep the computational load relatively light.

Availability and expansion of VSR

Currently, Video Super Resolution is only available to a handful of Edge Canary users. 

However, Microsoft plans to make this feature available to additional users over the coming weeks. 

The company also plans to expand the list of supported GPUs, so if your hardware isn’t initially compatible, don’t be disheartened. 

It will likely take a while before betas and polished versions support the feature.

Working behind the scenes

Video Super Resolution uses ONNX Runtime and DirectML to make it accessible to more users. 

These teams have spent years refining their technology to make the most of your graphics card’s processing power. 

The MSRA’s machine learning models are a significant part of the VSR feature. These models were distilled and compressed by the Microsoft DeepSpeed team from a much larger model. 

Although the base machine learning models improve the quality of the video, they require high compute cost, making it challenging to achieve real-time inference on consumer GPUs. 

To overcome this challenge, the DeepSpeed team used advanced compression techniques to reduce the model size and inference cost. 

Microsoft also focused on video enhancement metrics, which include LPIPS and FVD, but these metrics cannot fully reflect human perceptual preference. 

The MSRA team built an end-to-end pipeline to subjectively evaluate video enhancement tasks’ performance.

Conclusion

Video Super Resolution is an exciting new feature that promises to revolutionize the quality of videos viewed in Microsoft Edge. 

While the feature is currently only available to a select few Edge Canary users, Microsoft plans to expand its availability over the coming weeks and widen the list of supported GPUs. 

By utilizing the ONNX Runtime and DirectML, Video Super Resolution is portable across GPU vendors and makes the most of the processing power of your graphics card. 

The machine learning models from MSRA, which were compressed by the Microsoft DeepSpeed team, improve visual quality over traditional up-sampling methods, and evaluation metrics for subjective evaluations were developed by the MSRA team. 

Video Super Resolution is set to enhance the streaming experience for Windows users, providing crisp and clear videos on various streaming platforms without sacrificing bandwidth.

share

Written by

gabriel

Reviewed By

Judith

Judith

Judith Harvey is a seasoned finance editor with over two decades of experience in the financial journalism industry. Her analytical skills and keen insight into market trends quickly made her a sought-after expert in financial reporting.