Last Updated: April 01, 2022
BitClear uses deep perceptual denoising and upscaling for video delivery systems. The technology provides a scalable solution to denoise and upscale severely distorted low-resolution video assets. BitClear is an AI-based video processing technology that removes compression artifacts (like blurring and blocking artifacts) from user-generated (or heavily-compressed) content. Video content that has been through multiple transcoding iterations can be revived to the maximum possible quality without affecting the artistic intent of the original creators. The process can also allow for video upscaling, all with as little as 5ms processing latency on GPUs or high-performance CPUs.
- Simple integration
BitClear post-processes video content without changing the encoding. These gains are attainable without any of the disadvantages associated with needing access to (or replacing) components of encoders, streaming infrastructures, or encoding devices.
- Codec independent
BitClear is compatible with any existing codec (including MPEG AVC/H.264, HEVC/H.265, EVC, VVC and AOMedia VP9, AV1, AV2), and supports all resolutions up to 8K.
- Single pass
Only a single pass over the content is required (with single frame latency), prior to any number of subsequent encodings or processing steps.
- Breakthrough speedup vs. all competing approaches
Currently, BitClear models can run on 4 CPU cores or higher, achieving real-time processing on most cloud instances, and in general 7x to 32x faster runtime versus the best-performing deep neural network based denoising and upscaling solutions.