Video Enhance AI v2.1.1

Download: Windows | Mac
Released March 30, 2021.



  • Fixed the garbled text issue on some Mac machines
  • Video Format dropdown should no longer clip at the bottom of the screen when the window height is small
  • Multiple animations have been removed during processing, which should speed up processing times slightly (your mileage may vary)
  • “Keep Audio” and “Crop to fill frame” should both stick correctly between different videos in a batch

Known Issues

  • Some mac machines may not be able to properly render “Split View”. If the output doesn’t seem to be any different than the input, switch to a different view mode.
  • When re-processing videos, old settings from the previous process may intermittently get applied.
  • There’s currently an issue with the command-line not selecting the correct model in the background. You can circumvent this by opening the GUI, selecting your model, hitting “Save Default”, then running your command line options. This is not ideal, and we’ll have a fix out for this in the April release

v2.1.0: Windows | Mac

Major Features

  • Artemis MQ/LQ have been updated, with dehaloing removed (introduced in v11). Artemis MQ/LQ v12 should not produce the hazy look found in v11, and is now more similar to v10 with quality improvements.
  • Two new Artemis models: Artemis Dehalo and Artemis Strong Dehalo v1. These models specifically address the oversharpening problem, and should have less of a hazy look than Artemis MQ/LQ v11.
  • Dione TV/TD have been updated, with dehaloing removed (introduced in v2). Dione TV/TD should not produce the hazy look found in v2, and has a similar look to v1 with quality improvements.
  • Two new Dione models: Dione Dehalo and Strong Dehalo v1. These models address the oversharpening problem with previous versions, and should introduce less haze than previous versions.
  • Playback toolbar has been overhauled to be more flexible, fully functional, and easy to use
  • More preview modes: Choose between single view, split view, and side-by-side view


  • When opening videos, only the first one is selected to prevent accidental editing of added videos
  • Continuous scrolling, with ability to input a specific zoom amount
  • Added ability to change the number of previewed frames in Preferences
  • Added ability to change the preview panel’s background color, chooseable from Preferences
  • Added ability to change the video playback toolbar’s docked position to above or below the preview panel
  • Preview panel now bounds the preview video correctly within it’s borders
  • Controls panel has been cleaned up and made more consistent as to not appear as cluttered
  • Redesigned video thumbnails to make the selection box and clear button more visible
  • Added filename to video thumbnail
  • Most actions have shortcuts now
  • Removed the “Are you sure?” quit dialog
  • Grain settings, Dione interlaced mode, MOV selection, and MP4 Compression Factor have all been added to the command-line interface
  • Dione interlaced mode has been added to the CLI
  • MOV selection has been added to the CLI
  • Compression Factor has been added to the CLI
  • Preferences dialog now shows a confirmation icon when a setting has been updated
  • Zoom Original has been removed, to reduce confusion in the previewing experience
  • “Reset default settings” has been added to the Process menu on Windows and Mac
  • Preview panel now remembers what portion of the video you were panned to between each video


  • Window resizing should now be much faster
  • Grain settings should stick when loading auto-save, and apply correctly to “Save Default”
  • Grain textbox input should work in 0.1 values
  • “Don’t Show Again” checkbox should work correctly in the System Compatibility dialog
  • Fixed a typo in the “Continue Processing” dialog

Excellent updates. Is there any loose timeline of when information about major audio updates to this program are able to be divulged?

Here’s a rough outline of our current gameplan (subject to change and not comprehensive, of course):

April Release: v2.2.0

  • Comparison View (multi-model preview)
  • Overhauled model selection (more flexible for advanced users, easier to use for new users)
  • FPS conversion model OR Artemis / Dione model updates

May Release: v2.3.0

  • Multiple video I/O fixes (frame skipping, frame cut-off on MKV, 4:2:0 chroma fix on interlaced source, plus others) and codec / output additions (MKV output, different ProRes options)
  • FPS conversion model OR Artemis / Dione model updates

June Release: v2.4.0

  • Better audio support (correctly copy with supported codecs / input formats, issues with audio and framerate conversion)

July Release: v2.5.0

  • Projects, separate Render Queue application, and Preset Manager

If you have any questions about these, let me know. From a product (non-model) standpoint, our top two priorities at the moment are Video I/O issues and then audio issues. We’ve been spending the majority of the time the past two months on interface updates to bring VEAI on par UI-wise with our other maintained products (GP / DN / Sharpen / Mask), which also serves as a nice ramp-up for me as I’ve migrated over to VEAI.


Wow, thank you very much for the super detailed answer. Sounds good.

Thanks, these look like great updates. I’m sure I’m not the only user who would like some information about RTX 3000 cards and why VEAI currently seems unable to take advantage of the extra resources these cards have. If you could provide some info that would be great, even if it’s just to say it won’t be anytime soon.


I would like to know too. Most other AI based processing apps i’ve used show a difference between RTX 2000 and 3000 cards except Video Enhance AI.

Thanks for the inside on future updates!

Could you please elaborate on this specific topic? Will there be an overhaul of those models of some sort or just further training? Also I would like to know what you mean by “FPS conversion” in context of Artemis, since as for now only Dione does FPS conversion.


Can you Add REC 2100 and Rec2020? H265 High Profile10.
This may a Big leap Process HDR10+.

I now Using PermierPro Along with Divinchi Resolve Studio.

1 Like

@taylor.bishop Has my VEAI been extended ?

Did my first few tests with the new Dione and Artemis versions.

I really see an improvement for the MQ and LQ versions, over the last. I also very much like the new approach to have a model for specific artifacts. I can also say that the Dione_TV model seems useful to me now. The old one had a tendency to introduce some weird “flakes” of artifacts with certain details. For example vegetation was a big problem. The new version handles it way better, pretty much like v1, but with more detail enhancement.

About the moire and dehalo models: Can someone post a good example where they should be applied? I always had trouble to distinguish between different video problems. I think I know what it should look like but I am not really sure.

BTW. @taylor.bishop Is there a new version of Artemis HQ in the pipeline as well for the next version?

Me GPU Setup

Need more options for Saving

Giga Pixel has more options when Saving File(Quality,Colorspace):slight_smile:

So I assume development team have no plan to improve RTX3000 series GPU performance in near future ? :thinking:

In my humble option, the biggest downside of VEAI is the processing speed.
I understand that AI process require lots of processing power, that is inevitable.

Recent high core count CPU and GPU have improve performance significantly, however VEAI can’t get the full benefit of those new hardware.

1 Like

Our engine development is decoupled from our product development at this point, seeing as the new engine is (almost) in all of our currently maintained applications. We have quite a few performance improvements in the works as we work closer with our hardware partners, but those get included into releases as they come down the development pipeline. We don’t slot those into our schedules on the product-side, if that makes sense.

We’ll be getting another engine update in VEAI for either the April / May release, but that depends on how much longer the engine integration into Gigapixel AI takes.


Will you add interpolation by AI as well? (FPS conversion model). For this task I use another free software, it works very well, so could be an improvement for your soft too.


In setup cannot change Models folder place. How can I move this folder to other drive from my SSD (C:\Users…)?

1 Like

Just thought I’d comment on 2.1.0 now that I’ve done a little testing. First My computer has a 9 3900X with 64GB Ram, GTX 1660 Super OC and a 970 EVO Plus 2TB NVMe. I’ll be glad when people can buy GPU’s again!:dizzy_face: I normally process 240-480P videos to 720P. Sometimes 1080 but 720 is fine my my use case. I had been running mostly 1.9.0 as I wasn’t fully happy with 2.0.0. I mostly used AMQ-10 and AHQ-10.

Before installing I installed the latest Nvidia Studio release driver 461.92. I was on 460.?? but I didn’t want any surprises. I tried all models and all sizing with no issues. Everything is working great. No crashes or conflicts. Everything is working so well I uninstalled both 1.9 and 2.0.

The latest Artemis models seem better for what they are supposed to work with. The Strong Dehalo works well for some of my high contrast /high sharpness videos.

The UI improvements are very nice also.

Thanks Topaz.


@taylor.bishop where’s in your plan for the sdr to hdr ? :o

I would like to know this as well. Thank you.

Would it happen to be this?

Testing this out this morning… wonderful improvements! CLI is working perfectly, and really happy it’s completely aligned with the GUI options now. The GUI definitely feels a lot snapper overall too. I love the “single preview” mode. Great release!