Topaz Video AI Early Access v3.0.0-0

Hello Everyone,

The long wait for 3.0 is almost over! Over the past few months we have rewritten the entire Video Enhance AI app from scratch. We are now pleased to announce that Topaz Video AI is ready to enter early access.

While Topaz Video AI is entering early access today, we’re not quite done yet. There may still be the occasional issue, so please don’t hesitate to provide your feedback.

Windows | Mac
Headers, Dylibs, & Libs for custom FFmpeg builds

Changes from 2.6.4

  • Completely new app from the ground up
  • Full frame stabilization
  • Models can be chained now
  • Support for various output formats
  • Command line support
  • Parallel execution support
  • Ability to keep previews
  • Color depth up to 16bit now supported
  • Video files are playable on crashing (recovery coming soon)
  • Colorspace and color inconsistency issues reduced significantly
  • Many many more things

Known Issues

  • Certain VOB files do not open correctly
  • Some videos can still have some seeking issues in Preview/Export
  • Color space / HDR issues
  • Keyboard navigation missing
  • Intel codec not working on some machines
  • Some quality mismatch between 2.6.4 and 3.0
  • App doesn’t clear the temp folder
  • App only shows resolution in DAR and not in SAR
  • Multiple selection for input videos is missing
  • Different view mode is still missing
  • Customized Output bitrate will not reset to Preferences bitrate when importing a new file
  • Some failures can happen with some image sequences
  • Cropping while deinterlacing can cause failures
  • Failure can happen when resizing to a big resolution like 4k or 8k without a warning if machine runs out of memory
  • Speed can be slower than 2.6.4 with some models on some machines. We’ll be working on speed optimization in the next weeks

Problems & Known Solutions

  • If the output doesn’t open in other applications, disable Allow Recovery before processing in Preferences > Output (note: Allow Recovery is required for previewing exported video in the app while processing)
  • Some Mac users who participated in the Alpha or Beta may experience failures with all models. If this happens make sure the data directory is pointing at the correct version of Video AI
  • H.264 encoders have a maximum resolution of 4096x4096, use a different encoder for 4k or 8k videos
  • Changing data folder path can cause issues. To resolve: manually create a folder and name it models in the new location and copy the files from old to new location

Proteus Auto Caveats
The adjustments will not be visible and the sliders will show values of “0” when in auto mode. However, each frame is being analyzed for optimal settings while processing. If you keep Auto enabled and modify the sliders, the settings you choose will be value-adjustment percentages. For example, if Auto estimates the Sharpen slider with a value of 50, and you manually select -50, then the actual value applied will be 25.

CLI Usage
CLI mode has returned for Topaz Video AI 3. Detailed usage instructions are available here, on our support website.

Frequently Asked Questions can be found here.

Please upload videos that are not working or having any type of conflicts to this DropBox Folder .


Seemed like there were a ton of issues continuing to be reported in the beta forum… and it’s already going to get a final release? :grimacing:


Is Apollo rolling in later?

If I have a video at 30FPS that I want to be 60FPS but using Apollo or Chronos to improve it, is it functionally the same if I set Chronos to 100% vs setting it to 200% (but then playing the video at 2x in the video player)?

[I guess what I am getting at is, does the “slow playback speed” effect what Chronos is doing? Or does it just change the metadata so the video player plays the video at a different pace?]

This isn’t a final release but an early-access one that Video Enhance AI users can access. We’ll continue to release beta versions to beta users before the official release.


I’m a little confused on Proteus Auto as it sounds like the sliders are controlling percentages not actual values. Let me give you a hypothetical test case. I have a Dehalo value of 26 and a Deblur value of -6 from estimations. I want to have a constant Dehalo of 6 and Deblur of 21. I can get the Dehalo very close by using -76 but I can’t get to 21 if these are percentage changes. Am I wrong?

New to this beta testing so hopefully this is where I post an error:
I just tried previewing before trying to render and the preview failed with error code:
Invalid chars ‘,veai_up=model=prob-3’ at the end of expression ‘1,veai_up=model=prob-3’

Unable to parse option value “1,veai_up=model=prob-3”

Error initializing filter ‘veai_stb’ with args ‘filename=C:/Users/morro/AppData/Local/Temp/tvai_C_Users_morro_Desktop_Video Editing_Videos Being Edited_4-27-01_scene20010427 22_27_23_mov_0_2116935_cpe.json:smoothness=6:rst=0:wst=0:cache=128:dof=1111:full=0:device=0:vram=0.5:threads=1,veai_up=model=prob-3:scale=0:w=1440:h=1080:preblur=0:noise=0:details=0:halo=0:blur=0:compression=0:estimate=20:device=0:vram=0.5:threads=1,scale=w=1440:h=1080:flags=lanczos’

Error reinitializing filters!

Failed to inject frame into filter network: Invalid argument

Error while processing the decoded data for stream #0:0

On auto, you cannot set any constants. Your settings will alter the auto determination by a percentage on each frame. If you want constant throughout the video, you can’t use auto.


I was under the impression when we set a slider only that setting that the slider affects is not in auto and we get a constant value through out the video. Anyway I’ll have to evaluate auto very closely and determine a way of using it that is repeatable and logical. Otherwise I’m back to having to look at many locations in a video to determine what percentage of change I want. I already had a working method for estimation that works well for me so I need to figure out auto and how to get the most out of it.

Without knowing what the settings values are during auto makes this very difficult. You almost need a spreadsheet file with the values per frame for a proper evaluation. At least something like that output during a preview. I’d love to see a graph of all the settings and the auto value changes during a preview and how each setting might interacts with others.

Oh well guess I’ll wait and see how this progresses.

No, the underlying “base” is still auto, with an “overlay” of your percentage that you want to turn it down or up.

Frame 1 - Auto=30, your setting=-50, Result=15
Frame 2 - Auto=25, your setting=-50, Result=12.5

1 Like

Ok this is going to take some time to evaluate but i just did a 1080p to 2160p using Proteus auto, H264 Nvidia mp4 Auto bitrate and audio on auto. Speed was very comparable to 2.6.4 but the quality of the video was the best I’ve had from VEAI. While the input video was good quality the output was very good. Good enough that I did not see any glaring issues while actually watching the video.


I think Proteus Auto needs some improvements.

Ability to switch between percentage and absolute numbers to adjust the estimated values.

Ability to lock one slider to a specific value throughout the video.


Similar to the Standard setting in Photo AI, my M1 MacBook Pro running Ventura beta 5 gives the same glitches with the Gaia model. I did not see this with the other models.

1 Like

I’m getting great results from this version! Better results from Proteus than from 2.6.4. Only a few things are missing for me. The ability to make presets, and also having the preview just show me the last processed frame. I don’t want a looping preview, I just want to see the last processed frame. Oh also I had an issue where having scale set to Original resulted in an error relating to “-1 width” so I had to change it to “Select Scale” and set it to “100%”

1 Like

Sorry, I’ve got Processing failed on preview and here is error:

And here is mediainfo:

Complete name                            : F:\Videók\2022-03-21002.avi
Format                                   : AVI
Format/Info                              : Audio Video Interleave
Commercial name                          : DVCAM
File size                                : 275 MiB
Duration                                 : 1 min 15 s
Overall bit rate mode                    : Constant
Overall bit rate                         : 30.5 Mb/s

ID                                       : 0
Format                                   : DV
Commercial name                          : DVCAM
Codec ID                                 : dvsd
Codec ID/Hint                            : Sony
Duration                                 : 1 min 15 s
Bit rate mode                            : Constant
Bit rate                                 : 24.4 Mb/s
Width                                    : 720 pixels
Height                                   : 576 pixels
Display aspect ratio                     : 16:9
Frame rate mode                          : Constant
Frame rate                               : 25.000 FPS
Standard                                 : PAL
Color space                              : YUV
Chroma subsampling                       : 4:2:0
Bit depth                                : 8 bits
Scan type                                : Interlaced
Scan order                               : Bottom Field First
Compression mode                         : Lossy
Bits/(Pixel*Frame)                       : 2.357
Time code of first frame                 : 00:00:00:00
Time code source                         : Subcode time code
Stream size                              : 259 MiB (94%)

ID                                       : 1
Format                                   : PCM
Format settings                          : Little / Signed
Codec ID                                 : 1
Duration                                 : 1 min 15 s
Bit rate mode                            : Constant
Bit rate                                 : 1 536 kb/s
Channel(s)                               : 2 channels
Sampling rate                            : 48.0 kHz
Bit depth                                : 16 bits
Stream size                              : 13.8 MiB (5%)
Alignment                                : Aligned on interleaves
Interleave, duration                     : 40  ms (1.00 video frame)
Interleave, preload duration             : 200  ms

maybe a temporary fix, (not sure), convert your video in an another format maybe will fix the issue. be careful to keep the interlaced signal, until someone at topaz would be ok to fix that. you seems to have a DV interlaced recording, if it’s the original file from your camera, maybe the issue is here. don’t know if pro res can keep an interlaced signal, but you can convert your file in mpeg2 with a very big bitrate.
maybe it’s related to the 4/3 file which seems to be in anamorphic/widescreen (seen as 1024x576 by the software). someone else, had this issue if i remember well.

1 Like

Issue : no live preview available when Auto in Proteus is set to “On”. allow recovery turned on / off doesn’t change that. tried on two differents videos. if i turn off the “auto” button on proteus, the live preview work.
anybody else ?


There should be an option to delete preview-files from within the application.

Settings should be kept when opening a new file as in 2.6.4

Why do AviSynth files do no longer work in 3.x? They did in 2.6.4 (even though .avs was not in the input file extension list).


Setting the default encoder to ProRes 422 HQ does not work, the default is still H264 High (NVENC).

1 Like

Clearing the complete temp folder should be optional in case someone wants to keep previews for future reference.

There also needs to be an option to set the pixel aspect ratio manually (so resizing beforehand in other software is not needed if the AR in the metadata is wrong or a input format without aspect ratio metadata like image sequences etc. are used).

1 Like

Oh no I hope they fix .avs support cause I use that all the time.

1 Like