Copyright © 2023 Acrovid. All trademarks are property of their respective owners
Support
Products
Resources
About
Sign up for our newsletter to receive
news and special offers
Press/Affiliates
Share
About
About
User Guide
Configuring the GPU card
To configure the GPU card options, open the options dialog and select the GPU card from the detected GPU cards list in the “AI engine GPU
card” section:
Global configuration
Select the AI enhance tab to configure the global options and open the configuration dialog:
Configuration dialog
How to configure neoFootage AI video enhance and upscale
Acrovid,neoFootage AI and SuperBitdepth AI are trademarks or registered trademarks of Acrovid, Vulkan and the Vulkan logo are registered trademarks of the Khronos Group Inc. All other trademarks or registered trademarks like Adobe Premiere, After Effects, Dacinvi Resolve, Avid, Avid Media Composer,
ProRes, DNxHD, GrassValley, NVIDIA, CUDA,Intel, Quicksync or Cineform mentioned in this document are property of their respective owners. Reference herein does not constitute or imply affiliation with our company, endorsement, or recommendation. Some videos on this webpage are from Videezy.com
After changing the GPU card, restart the software.
GPU card test program
The GPU Test program check the GPU for compatibility and driver problems.
If you experience any GPU problem or the AI processing runs slower than it should
with your GPU card, please test your GPU and send us the test results.
In the AI global processing section, SuperBitdepth AI can be activated by default when a video processing needs bitdepth upsample.
For example, when processing 8-bit videos encoded with a 10-bit format, SuperBitdepth AI will be used automatically.
AI global processing
Configuring AI enhance
To configure AI enhance, select a video and press the button “Configure AI enhance”, this will open the configuration dialog.
After configuring, check the button “Activate”, this will activate the AI processing.
As with other features, AI enhance can be configured separately for each video in the media list, or configured for all videos in the list.
For configuring AI enhance for each video, set the configuration for selected videos mode.
If the configuration for all videos mode is selected, then the selected configuration will be applied to all videos when processing.
Options
In the AI enhance configuration dialog you can easily configure all the AI engine features.
After activating AI enhance or SuperBitdepth AI, the original and enhanced video can be compared by pressing the right mouse button over the
video, this will show the original video while pressed, or by activating the compare mode, this will show the videos side by side.
You can move the video frame using the mouse, zoom the video using the mouse wheel, or select a fixed zoom.
Previewing the enhanced video and comparing the results
If the video is HDR, the HDR display mode is activated by default, the HDR uses Windows HDR display mode if available, or a HDR simulation.
Check the HDR button to activate the HDR display mode.
HDR display mode
Press the button SnapShot to save a PNG image of the current processed video frame.
SnapShot
Configuring AI enhance and upscale
AI enhance includes 14 AI models trained for several video conditions.
The AI models are organized in two sections:
- Models for high/ medium quality videos.
- Models for low/ medium quality videos.
Check the button “AI enhance / upscale” to activate the processing with the
selected AI model.
After selecting a model, it is loaded in the GPU card, and the current video frame
is processed so the results between models can be easily compared.
The final videos can be improved keeping the source dimensions, or also can be
upscaled or downscaled.
The preview dimensions can be set to source video dimensions, upscale 2X and 4X,
HD or 4K.
The final video can be scaled to any video standard by configuring the video
resample in the transform panel.
TIP: If “Auto configure” upsample button is activated, then the selected preview
dimensions will be configured automatically in the transform panel.
Preview dimensions and final video dimensions
When processing HDR videos, selecting a HDR mode can improve the quality for
most videos.
TIP: when selecting a nits mode, all the values over the selected nits will be
removed.
SDR and HDR modes
Configuring SuperBitdepth AI
SuperBitdepth AI include 3 configuration modes:
Preset for 8 bit videos
This mode enhances the video bitdepth quality for average 8 bit
videos.
Preset for 10/12 bit videos
This mode enhances the video bitdepth quality for average 10 and 12-
bit videos.
Manual mode
The manual mode allows to adapt the filter strength to the video, from
adjusting the bitdepth enhancing strength, to removing posterization
and banding.
SuperBitdepth AI includes a powerful manual mode.
For selecting the manual mode, check activate in bitdepth enhance/restore.
Now you can configure the filter strength from 1 to 100 and check the
results.
Auto extra mode
The auto extra mode analyzes the video frame and adds a extra strength to
fix posterization and banding.
Manual mode
As with AI enhance, when processing HDR videos, selecting a HDR mode can
improve the quality for most videos.
TIP: when selecting a nits mode, all the values over the selected nits will be
removed.
SDR and HDR modes
Creating video previews
Video previews can be created using the current AI configuration.
To create a video preview, select the preview duration and press the button
“Create preview”.
The video previews filename include the AI model used.
Double click on a preview to play it, for deleting, select the preview and press the
button “Delete preview”.
Press the button “Open previews folder” to open the previews folder in the
Windows explorer.
TIP: the preview player is the internal player by default, you can configure other
external player in the options dialog.
TIP: the preview folder can be configured in the general options dialog.
Configuring color correction
The video can be easily color corrected using the color effects to get a perfect video.
For color correction, select the color correction tab, and apply the color effects, there are available basic and advanced color effects.
Color correction can be done in real time over the enhanced video frame, with AI processing activated or not activated.
TIP: color correction configuration is shared with the configuration in the color tab.
To show and hide the scopes check the buttons “Parade”, “Waveform” or
“Histogram” .
The scopes show the processed video, both with the AI enhance and
SuperBitdepth AI if activated, and the configured color correction.
Scopes
The AI models are organized in two groups, for high/medium quality videos, and
for low/medium quality videos.
Depending on the video quality, select the high/medium or low/medium AI
models, the results can be checked immediately in the current video frame after
selecting the AI model.
The base models for high/medium and low/medium quality produce balanced
results.
The models include a version with film grain, the grain models preserve more film
grain from the source videos and also create new film grain.
All models are trained for artifacts removal for most video compression formats
including HEVC, the low quality HEVC model is trained for low bitrate HEVC videos,
removing strong HEVC artifacts.
AI models tips and tricks
Configuration with other processing features
AI processing integrates seamlessly with all the program processing features,
After configuring the AI processing, any other processing setting can be added, for example the video can be converted to other video
standard, or converted to HDR.
The AI upsampling preview can be set to HD or 4K, to upscale the video, select
any upsample or standard conversion with aspect ratio conversion in the
transform tab.
For example, when converting SD 4:3 to HD or 4K 16:9, the aspect ratio
conversion can be selected for zoom or black bars, or configured manually in
the aspect ratio dialog.
TIP: if “Auto” button is checked, the selected preview upsample will be
configured automatically.
Upscaling and standards conversion
The AI configuration dialog deinterlace the videos for the preview.
When processing an interlaced video, the final video will be also interlaced, for
example, when AI enhance a 1080i video the final video will be 1080i.
To produce progressive videos, for example convert SD interlaced to HD 1080p
or 1080i to 4K progressive, simply select deinterlace in frame rate conversion
options.
Interlaced and progressive videos
For converting SDR videos to HDR, for example HD to 4K HDR, configure the AI
enhancing and then select color space and gamma conversion to HDR in the
color tab, or configure the HDR conversion manually in the color dialog.
Then select a 10-bit encoder format, for example high quality DNxHR for editing
or storing, or HEVC HDR10 for the final video.
SDR to HDR conversions
The video position slider can be dragged to a video frame, and when the
mouse button is released, the video frame will be processed with the selected
AI model.
GPU cards compatibility
AI engine uses Vulkan for GPU processing, the standard supported by Nvidia, AMD and Intel.
The Vulkan driver is included in your GPU card drivers, if there is any Vulkan error, update your GPU drivers to the latest version.
Vulkan is supported by Intel Skylake and higher, AMD GCN 2.0 and higher, and Nvidia Kepler and higher.
In this page you can find the GPU cards compatible with Vulkan:
https://vulkan.gpuinfo.org/listdevices.php?platform=windows
To test your GPU card, simply select it, and press the button “Test selected GPU”.
The full test is better to diagnostic problems, but it can freeze or crash if there is a GPU or driver problem, in this case uncheck the “Full test”
button.
The full test can take some time depending on the card model.
After the test is completed, click over “support@acrovid.com” to create an email for sending to us.
Alternatively, press the button “Copy to clipboard” and paste directly on an email to “support@acrovid.com”
Configuring temporal strength
AI engine temporal processing uses several previous and next frames with the
current frame.
The temporal strength can be configured manually, the minimum and default value
is 20, recommended for most videos, increase the temporal strength value to
remove flickering, interlacing line flickering, noise or pixels fluctuations.
It is recommended to create a preview to check the temporal strength value results.
DVD upscaled to HD - 100% crop
AI upscaled to HD, temporal strength 90 - 100% crop
