The new Neat Video 6 is optimized for speed, efficiency and smoother editing workflow. It’s packed with innovations, improvements and smart solutions that make it leap ahead in render performance.

In this article, we will review the available ways of measuring and optimizing the performance of Neat Video and offer recommendations for specific scenarios.

We will also look at the actual speed improvements of Neat Video 6 vs Neat Video 5.

Traditional optimization tools

NeatBench

If you used Neat Video before, you may already be familiar with NeatBench, a standalone benchmarking tool designed to measure the speed of the noise reduction filter (without running or even having the video editing application) on any computer.

This is primarily a benchmarking tool: it is intended to check the performance of a specific system and compare it with other systems. It evaluates both CPU and GPU, giving a clear idea of their individual and combined performance. This can be used to compare different models of CPUs and GPUs, for example when choosing components for a new system.

You can download NeatBench from the Neat Video website and run it on any compatible computer (Neat Video itself is not required for that).

Optimize Settings

There is also the built-in Optimize Settings tool (available in the Preferences > Performance) that measures the speed of the noise filter with different combinations of available CPU cores and GPU(s) and finds the best combination, that you can then choose to use for rendering.

It’s best to use Optimize Settings at the start of a new project — especially when working with different frame sizes and filter settings — to ensure Neat Video runs as efficiently as possible for that specific project.

Starting from version 6, the test run by the Optimize Settings tool measures the speed of Neat Video's own processing together with simulated host-plugin data transfer. This makes the test more suitable for optimizing Neat Video for a specific video editing project and less so for benchmarking computers. Therefore the results of the Optimize Settings test should not be compared with those of NeatBench as well as between computers, video editing applications, or versions of Neat Video.

Limitations of Optimize Settings and NeatBench

There are two key drawbacks of using Optimize Settings and NeatBench.

First, you have to run each of these tools manually, check the results and then decide what settings to adjust in Neat Video Preferences. This requires some manual work, even if not very complex. Yet if you do not run Optimize Settings after setting up Neat Video on a new computer, it will not use all available computing power and render speed will likely be lower than it could be.

Second, both of these tools report Neat Video’s performance in an ideal isolated environment. The FPS (frame per second) numbers they measure show the speed achieved when applying the Neat Video filter as the only processing load, without taking into account any additional load placed, during rendering, by the video editing application, by its codecs and other effects that may be present in the project. Ignoring this extra load during optimization may also lead to sub-optimal render speed.

Next-Gen Optimization: Auto Mode

Auto mode introduced in Neat Video 6 — the next step in maximizing performance. The Auto mode takes a more automated and holistic approach to optimization than anything we had before. Its primary goal is to automatically ensure the best possible denoising speed during render/export.

The Auto mode is enabled by default, so it will choose the best performance parameters automatically once you install and start using the plug-in, without requiring any input from you.

The Auto mode considers your computer’s entire workload, not just Neat Video alone. It aims to optimize Neat Video performance directly during render, accounting for the activity and resource usage of a video editing application, codecs and other video effects.

It also dynamically adapts to hardware changes when you re-configure computer mid-project.

While Optimize Settings offers a hands-on way to manually measure and optimize performance of the filter in isolation, the Auto mode does all that and then some more, to automatically and dynamically adjust for the best Neat Video performance during render/export. Because the Auto mode does all that automatically, it will consistently deliver fast renders without you having to manually optimize settings.

Render speed improvements

As we mentioned earlier, the new Neat Video 6 includes several new solutions to increase the render speed, efficiency and provide smoother editing workflow. The Auto mode is one of them and there are other elements that positively affect the overall performance.

Let’s see what kind of speed improvements we can get when using Neat Video 6 as compared with Neat Video 5. We will check that using a test project with one UHD (3840x2160) clip in Premiere Pro 2025 on a machine with AMD Ryzen 9 3900X and NVIDIA GeForce RTX 3090.

First we check Neat Video 5.

Initially, we measure the speed of Neat Video 5 itself using NeatBench (in FPS or frames per second):

Then we instruct Premiere to export that test project into a similarly sized UHD output clip and measure the time (and then calculate the speed in FPS) it takes to complete the export. The export process includes Premiere itself, its input and output codec, and Neat Video 5, all working in a chain to produce the final video file. We export to one of two output codecs (ProRes 422 and H.264) and get the following results:

As you can see, the export speed of Premiere is significantly lower than the own speed of the filter (Neat Video 5) used in the project. Which is to be expected because the export process includes more elements than the filter alone, all working one after another which obviously takes more time.

The export speed also appreciably changes when different output codecs (or different codec modes) are used. A “slow” codec can reduce the export speed quite a lot. Generally, the export speed may also depend on other effects used in the project, on the codecs used in input clips, on the version of Premiere Pro and its settings, etc.

Now let's check Neat Video 6.

First, measuring the speed of the filter alone using NeatBench 6:

There is a 1.38x speedup with Neat Video 6 on that hardware.

And now let’s measure the render speed. We run the same test export in Premiere Pro and compare the export speed when using Neat Video 5 vs Neat Video 6:

As you can see, export with Neat Video 6 is 1.80–2.47x faster.

This export speed improvement (1.80–2.47x) is clearly higher than the speed increase of the filter itself as measured by NeatBench (1.38x). This is because Optimize Settings and NeatBench cannot measure the contribution of all stages of the render process that become faster when using Neat Video 6. So comparing the results of Optimize Settings or NeatBench is less indicative than comparing the render times with Neat Video 6 vs Neat Video 5. Therefore if you want to know how faster your renders can be with Neat Video 6, it makes more sense to compare the actual render times.

Why the render with Neat Video 6 on that computer is 1.80–2.47x faster than with Neat Video 5? There are several main reasons for that:

1. Core filtration functionality of Neat Video 6 is more efficient than that of Neat Video 5;
2. Neat Video 6 balances work between CPU and GPUs better, achieving must faster speed in multi-device configurations than version 5;
3. Neat Video 6 interacts with video editing applications better than Neat Video 5, and that allows the plug-in and the applications to work together faster than before;
4. Neat Video 6 uses GPU-based integration with Premiere utilizing faster exchange of video data, which shortens rendering times even further.

The numbers above were measured using a very simple test project in Premiere with a single clip in the timeline and with only one effect applied. The speedup can vary a lot depending on the complexity of a project. In some cases (especially when using Lumetri), we’ve seen Premiere renders complete up to 10x faster with Neat Video 6 compared to Neat Video 5.

In other video editing applications, export typically becomes faster as well. Specific figures depend on the application, hardware, OS, drivers, other effects and codecs used in the project. For example, a similar test project exported by Final Cut Pro 11 on MacBook Pro M3 Max with Neat Video 5 and Neat Video 6 shows the following results:

It is clear that there is a significant speedup of the export process in Final Cut too.

As you can see, it can be very beneficial to use Neat Video 6 compared with Neat Video 5. There are many performance improvements over the older version related to the closer integration with the host application and general optimization of the CPU and GPU code. In addition, the default Auto mode introduced in Neat Video 6 enables higher performance automatically, allowing the editing and render processes to be most efficient out of the box, while in Neat Video 5 that required certain manual operations from the user.

Recommendations

Both manual and automatic measurement and optimization tools available in Neat Video 6 can be useful in specific scenarios. Here are the key recommendations based on different use-cases:

1. Automatic performance management
   If you want to just let Neat Video automatically manage its performance parameters to consistently achieve fast render speeds, then rely on the Auto mode setting (default) available in Neat Video 6.
    
2. Manual control over performance
   If you need to manually measure the performance and control the performance settings in specific conditions (for example if you need to assign a specific GPU to Neat Video and reserve another GPU to the host application), then use the Manual mode available in Neat Video 6 and in older versions. Test the speed using the Optimize Settings tool and adjust the CPU and GPU settings in Neat Video Preferences.
    
3. Check render speed
   If you want to evaluate the actual render speed, then it is best to directly run a test render with a short clip (like 30–60 seconds) and measure (not just check an estimate by the video editing application) the total render time or render speed. Do not use the measurements of Optimize Settings or NeatBench 6 since they measure the speed of the filter alone, not render/export speed.
    
4. Compare render speed of Neat Video 6 vs Neat Video 5
   If you want to compare the speed of Neat Video 6 vs Neat Video 5 (or another version), then it is best to run two test renders with different versions of Neat Video applied to the same clip and to measure and compare the render/export times or speeds. 
    
5. Check hardware speed
   If you want to quickly check the speed of a new computer or to compare one hardware (CPU or GPU) to another, then use NeatBench 6, which makes quick direct measurements and doesn’t require a complex setup (no need to install the whole video editing application and set up a test project). This tool is primarily intended to evaluate and compare hardware. It measures the speed of the filter alone. It cannot measure the render/export speed of your video editing application; for that use the options 3–4 above.