Enabling the previous model versions allows you to download all legacy versions of the models that have been removed from the most up-to-date build.

We consider the Legacy Models unsupported. We will not improve quality concerns or fix compatibility with these models.

To access these models, open Topaz Video AI and select 

Previous Model Versions

Topaz Video AI uses ffmpeg to produce output files and apply different encoding settings. 

While the graphical menu for export options includes some of the more popular encoders and containers, there is a way to add custom settings that can be used for more advanced workflows. 

The video-encoders.json file can be modified to add additional options for encoding. This file is found in the 'models' folder on both Windows and macOS versions of Video AI:

Windows

macOS

Linux

As an example of a custom encoder option, we will be updating the "H265 Main10 (NVIDIA)" encoder setting to use the 'slow' preset and B-frame referencing for more efficient compression. 

Some of these features are only available on certain GPU models, so it's recommended to research which exact encoder features your specific graphics card supports. 

Copy the preset that most closely matches the custom option you'd like to create

In this case, "H265 Main10 (NVIDIA)" will be duplicated directly underneath the original in the video-encoders.json file

In addition to options for the video encoder, this json entry can be edited with a different audio transcode setting, maximum bitrate, bit depth, and OS compatibility to prevent settings being shown on incompatible devices.

It is highly recommended to test any custom video-encoders.json entries with a short video and inspect the result using 

 to ensure that the output matches the expected results.

Custom Encoder Options

Topaz Video AI supports executing scripts using a command line interface.

This is designed for advanced users comfortable working in such an environment and offers more flexibility in customizing a variety of scripted processes.  

We highly recommend using the app’s user interface for those not comfortable working in a command terminal.

The majority of the commands for this build will be FFmpeg commands.

 There is no need to install FFmpeg, it is automatically included with the TVAI installer. This article will outline the basic functions for TVAI’s CLI, however, you will want to familiarize yourself with FFmpeg commands for more complex use cases.

Before using the CLI for the first time, we recommend launching the GUI and logging into the app. This eliminates the need to use a command to log into the app and will allow you to launch the terminal directly from the GUI. 

After logging in, select Process > Open Command Prompt, this will set the model directory automatically. The next time you want to launch the CLI without the GUI, follow the steps below:

This variable should be set to the folder where you want model files to be downloaded. A location with ~80 GB of free space will work best.

Chosen during initial installation (Windows)

/Applications/Topaz Video AI.app/Contents/Resources/models (macOS)

This variable should be set to the folder containing the model definition files (.json), your authentication file (

In most cases, this value should not be changed from its default setting.

TVAI is used as an FFmpeg filter, and all models will work on graphics devices from Intel, AMD, Nvidia, and Apple using a command like this example: 

However, different graphics cards may support different encoders and options. Similarly, different encoders support different options, so you may need to tweak settings on different machines. The following options can be used to take advantage of hardware acceleration features from different GPU manufacturers: 

Intel

NVIDIA

macOS (Intel & Apple Silicon)

Additional Information on the Stabilization Models: 

To use the stabilization model, there are two commands that need to be run one after another.

Command Line Interface

You can add a Second Pass Enhancement to create a higher-quality output file from Video AI. 

By default, this option is disabled as not all users will have a machine capable of running multiple enhancements at one time.

Proceed with caution as adding a Second Pass Enhancement is very dependent on your computer system’s capabilities. 

Always keep in mind, that every use of Enhancement, Frame Interpolation, Stabilization, Motion Deblur and or Grain will add to the processing render time and resources needed from your system to calculate and produce results. 

If your machine is not able to use this feature, you are able export a first 'layer' of Enhanced footage to then be re-imported to Video AI and processed again.

Rendering out a first version and re-importing to add another enhancement is another good way to add multiple treatments. This is a smart way to still use combinations of models without stressing your system too hard. 

In the VFX world, these render passes in compositing are called pre-comps. They are essentially more resource-friendly passes that are combine (composited) to make for the final render.

If you’re still onboard for multiple enhancements, this is where you can enable it. 

Navigate to the upper screen left corner, select the 

 tab and follow the dropdown menu until you find 

. On the top selection of "Show second enhancement feature", turn the slider on. Be sure to 

This will show the "Add second enhancement" under the initial Enhancement.

There are many possible combinations of filters that can be used in this mode, but here are some common scenarios that may be a good use case for Second Enhancement:

Deinterlacing and enhancing footage with Iris or Dione, followed by Nyx or Proteus for scaling and denoising.

Scaling standard-definition video to HD using Artemis, followed by Gaia HQ to further scale to 4K.

There is an additional toggle available in the Second Enhancement menu that allows users to manually set the 'intermediate resolution' -- the size of the output of the first Enhancement filter pass. Think of these as weights. How much weight/strength will the first enhancement take over the second enhancement when both are applied to your footage.

These options are equivalent to selecting 1x, 2x, or 4x Upscale from the Resolution Control setting in Single Enhancement mode.

As an example, we can look at a use case where a video clip is being scaled from 540p to 2160p. This is a 4x increase in dimensions, and can be accomplished in several ways:

In Auto mode, the Intermediate Resolution will be set to 2x, and the Second Enhancement filter will also run in 2x mode. If you'd prefer to run a specific model in a specific scaling mode, the manual option can be used to specify the scale factor of the first pass.

Advanced Functions

Previous Model Versions

Custom Encoder Options

Command Line Interface

Second Pass Enhancement