NVIDIA RTX GPU Connects DaVinci Resolve to Power

Two of the most important questions that the Creative COW audience asks themselves are: How can I make my workflow more efficient? And, what are the latest tools that can improve the quality of my work? Here in 2024, both of those questions can often be answered by only two letters – AI.

For professional content creators, artificial intelligence has gone from innovative novelty to downright necessity. If AI gives everyone an edge, you’ll certainly need to compete by keeping the best AI tools at hand. And if AI is an editor’s hand tool, what would an AI power tool look like? The answer to that is hardware that can drive AI tools to their limit of function and beauty.

For content creation, Blackmagic Design’s DaVinci Resolve has the tools, and NVIDIA RTX GPU can deliver more power.

The DaVinci Resolve app is already attractive to users by providing a standalone place to process footage –  basic non-linear editing, audio, VFX, color grading, and motion graphics – all in tabbed pages on a single software interface. According to Shawn Carlson, Product Specialist at DaVinci Resolve, this design was based on feedback from customers.

“One of the challenges we heard about is that post production creatives don’t necessarily want to take a project from one application to another; rather, many prefer to stay within a familiar environment. Our team has worked tirelessly to provide options that help true creativity blossom,” said Shawn in a conversation with Creative COW.

And within that convenient interface, DaVinci Resolve has incorporated AI into many of their tools, starting with Neural Engine in 2019, and totaling over 100 AI implements to date. The April 2024 release of DaVinci Resolve 19 public beta has increased the utility of all these AI tools and folded in a couple of exciting new ones within DaVinci Resolve Studio.

Now available in DaVinci Resolve Studio, Ultra NR is an AI alternative to Spatial Noise Reduction, replacing the Faster, Better, and Enhanced options with a single Analyze button. Hitting Analyze places a draggable square into a flat area to detect noise in the shot, then machine learning quickly deduces the appropriate luma and chroma levels needed to effectively reduce or eliminate noise in the image.

With these AI tool developments at our fingertips, the next step left toward improving workflows lies in turning DaVinci’s AI implements up to full capacity with a powerful GPU.

The argument over what magic combination of PC hardware gives the best output from the DaVinci Resolve 19 update is settled by a conversation with esteemed authorities in the computer hardware field, the folks at Puget Systems. 

Puget Systems builds custom workstations for small to enterprise sized post houses, assembling reliable machines for post-production for over 20 years. Puget takes the hassle out of running intensive NLE and DAW softwares for content creators with extensive testing, and DaVinci Resolve is a common consideration for Puget Systems to build workstations around.

We spoke to Puget Systems Lab Supervisor, Matt Bach, to learn what hardware configuration would give the best performance for the latest AI tools in DaVinci Resolve 19.

“For most users, there are a few rules of thumb for DaVinci Resolve in general, and they apply well to the new AI tools Blackmagic Design is adding as well. First, use an NVIDIA GPU,” said Matt. “They tend to have higher performance than other GPUs, and the ‘Studio’ drivers, in particular, are very stable. Second, make sure you have enough VRAM (GPU memory). Most cards these days have at least 8GB of VRAM, which is enough for 4K timelines. If you want to work with 8K timelines, or heavily utilize many of Resolve’s GPU-heavy features, however, you might want a card with as much as 24GB of VRAM.”

A graphics card of that size includes the NVIDIA GeForce RTX 4090, which is powered by the NVIDIA Ada Lovelace architecture and offers G6X memory to easily power DaVinci’s AI tools in post-production. The power advantage also kicks in for export time in DaVinci Resolve Studio.

With regard to the best configuration, Matt continued, “Last, prioritize a single, higher-performance GPU over multiple GPUs. Multiple cards can improve performance for some tasks in Resolve (OpenFX and noise reduction in particular), but other things like processing RAW footage and many of the new AI features are currently limited to a single GPU. The exact card to use depends on how often you utilize the GPU-based features in Resolve, including OpenFX, noise reduction, editing of RAW footage, and the newer AI tools. The more modern and powerful the GPU, the faster those features will be processed and the more time you will save.”

The advantage with NVIDIA RTX GPU in AI applications is clear when it comes to speed. Says Matt, “NVIDIA has positioned themselves very well for these new AI tools, and it shows in the raw performance numbers.”

“In our testing so far, using an NVIDIA RTX GPU can be around 2x faster than a comparative GPU on average, or nearly 3x faster than the latest Mac-based systems. If you anticipate using (AI tools) regularly, a modern NVIDIA RTX GPU can save you a significant amount of time.”

One of DaVinci’s new tools in the 19 public beta, Ultra NR, also sees significant time reduction with NVIDIA RTX GPU, specifically the GeForce RTX 4090, performing three times faster than the Mac M2 Ultra.

With the introduction of their Neural Engine in 2019, BlackMagic Design integrated new AI features and tools into Davinci Resolve such as up-scaling facial recognition, color matching and more. With the addition of NVIDIA’s TensorRT to the DaVinci platform in September of 2023, which provides library referencing technology to optimize DaVinci’s deep learning models, the speed and efficiency doubled for AI functions in DaVinci in update 18.6.

This provides an advantage for pairing an NVIDIA RTX GPU into the latest DaVinci Resolve workflow, as the integrated ability for the NVIDIA GPU to work hand in hand with Tensor RT makes for a strong combination. The TensorRT adds a fantastic boost for AI tools for video editing, color correction, visual effects, motion graphics, and audio post-production.

Matt Bach of Puget Systems agreed that utility AI tools, such as those available within DaVinci Resolve, can give creators both an edge and a new quality of work-life that we might not want to do without.

“The AI tools in Resolve are currently focused on making the user’s life easier,” said Matt. “The masking tools, in particular (Magic Mask and Automatic Depth Map), are terrific and are already becoming a regular part of people’s workflows. Others, like Text Based Timeline Editing, have a lot of people interested, but will take time for users to adjust their workflow to take advantage of them fully.”

To stay on top of upcoming production-easing AI developments facilitated by NVIDIA RTX, check out AI Decoded. This NVIDIA news spot demystifies AI once a week with RTX-powered AI news, insights and inspiration.

---

---

Enjoying the news? Sign up for the Creative COW Newsletter!

Sign up for the Creative COW newsletter and get weekly updates on industry news, forum highlights, jobs, inspirational tutorials, tips, burning questions, and more! Receive bulletins from the largest, longest-running community dedicated to supporting professionals working in film, video, and audio.

Enter your email address, and your first and last name below!