Topaz Video AI v3.0.7—and all of the weekly updates since launching v3.0—delivers many performance and stability improvements, including optimizations that improve model quality and make the app run more smoothly. We’ve also improved codec and GPU driver support.
We’ve updated how some of our models work, which improves export video quality. Our AI models perform best when the input video frames or image sequences are similar to the media used for training in terms of content, artifacts, color space (e.g. YUV, RGB, BGR), and other variables. With Topaz Video AI v3.0.7, the AI models will receive the input media in the same color space that they were trained on, thereby increasing video quality.
Mac users will now experience improved quality when exporting videos using h.264, h.265, and h.265 Main 10 encoders, and we added support for the AV1 encoder on Intel ARC GPUs. We also improved support of h.265 encoders for NVIDIA GTX GPUs.
We’ve implemented many bug fixes and optimizations to make Topaz Video AI run more smoothly across a wider range of hardware. Some of the more notable benefits involve:
We’ve also redesigned the app’s Preferences window layout to be easier to navigate and have made error messages more helpful by providing possible solutions to consider.
In addition to these improvements, our Topaz Video AI product roadmap outlines what we're actively working on, including enhancing performance, the preview experience, and model quality. If you haven’t tried Topaz Video AI yet, be sure to download your free trial to experience all of these improvements and see how much better your video quality can get.