Topaz Video AI Early Access v3.0.0-4

Hello Everyone,

We have another update for Topaz Video AI 3.0 with lots of fixes and improvements. Please provide your feedback.

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

Changes from v3.0.0-3

  • Visual cropping tool has been added (Lock aspect ratio coming soon)
  • Added “gather log for support”
  • Fixed deinterlace and upscale with cropping not applied correctly
  • Misc. fixes and improvements

Known Issues

  • Preview mismatch for certain VOB files
  • 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
  • Multiple selection for source videos is missing
  • Output bitrate is not reset to Preferences bitrate when importing a new file
  • Different view mode is still missing
  • 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
  • Proteus with Auto + Stabilization may cause artifacts
  • Preview with image sequence doesn’t work or plays at 25fps
  • Aspect ratio crop missing

Problems and Solutions:

  • Make sure Max. Memory Usage in the Preferences is set to a high number
  • If the Gaia model fails on Mac, try re-installing the app using the DMG file
  • If export fails, try using Transcode/None/Auto in Audio Settings in Preferences

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 .

6 Likes

I have used 2.6.4 and previous versions for many months. Today I downloaded the 3.0 version and tested out some of the features. Here is my feedback for the new 3.0 version. Overall, I like the old 2.6.4 UI much better, but this version does have some UI improvements that I prefer.

Suggested Improvements

Mark Start and End Time Buttons: Bring back the Mark Start time and Mark End Time buttons. Using the mouse for this is not easy and I don’t want to enter an exact time code. Using the arrow keys and then pressing one of the mark start buttons is easy and efficient.

Mouse Wheel: Using the mouse wheel to scroll the right hand panel is way too slow. It needs to be sped up by about 4-5x.

Fit to Screen: This doesn’t seem to work properly. The larger you make the video panel, the less the video fits to the screen. If you make this top area very small, there is no gap above the top and below the bottom of the video, but if you pull the divider down to make the video area bigger, the empty wasted space above and below the video widens more and more as you pull down the divider.

Zoom and View Buttons: Please move these back to the same position that they are in, in version 2.6.4, which is slightly to the left of the right panel. It’s more convenient having these two options on the right hand side, which is where most of the work is being done. I don’t like these options being on the far left in a drop down. It’s inconvenient. (I realize that the “View” option is not in yet, but will be added later)

Zoom Options: Please bring back all of the view options that were in 2.6.4. Zoom to fit, zoom to 50%, zoom to 100%, 200% and 400%. I don’t like the 1:1 or 1:2 options. They are confusing. I do like the new button that is added that makes the video fill most of the screen. The icon confuses me though. Shouldn’t it say Fullscreen, or something like that?

Source Video in Bottom Left: The source video thumbnail in the bottom left is way too big. It takes up too much vertical space. Maybe there should be two lines underneath the video. Name and then the video info. This would allow for the width to be shrunk considerably. Or just use the same layout from 2.6.4 and show the current video details at the very bottom. I don’t need to see a giant thumbnail of the video I’m working with as I typically work on one video at a time. I want the video at the top to be taking up most of the real estate.

No Estimated Time to Finish: Maybe I missed it, but when I’m exporting a file, I no longer see how long it is going to take. This needs to be added back. It’s very helpful to know how long something is going to take to process, even if it’s an estimate.

Better Contrast: A lot of this UI is gray on black. It’s hard to read. I’d like more contrast, like in 2.6.4, where you have white on very dark blue. It’s much easier to read and looks nicer. I prefer the blue theme that 2.6.4 used. The black is kind of boring. I’d like to have UI theme options if possible so that I can go back to the blue.

Video Autoplays After Export and then Clicking on Enhance: I think this is a bug. The video autoplays after exporting and then clicking on the enhance button. I do not want the video to autoplay ever.

New Feature - Face Recovery: I’d like for face recovery to be added, like in gigapixel ai.

What I Like

Clip Preview Bar: I like that there is a visual showing you how much of the clip you are going to preview when you change from 2 seconds to 5, etc.

AI Filters Organization: I think that the new model UI is a major improvement. In 2.6.4 you had a huge list of models and you had to know which one did what. Now all of the models are organized based on what they do. I like this much better. It is more organized.

Saved Previews: I like being able to watch previews I generated in the past without generating them again. This is a nice feature.

3 Likes

I agree with almost everything you said, and have a few more suggestions:

  • Add frame numbers back and provide an option to show both the timecode and frame number. Displays are large these days—showing both values won’t take up much space (and what I propose is an option anyway).
  • Add back the ability to select the number of frames to preview. And make it a text box (in addition to the list)—I don’t want to be limited to a list of predetermined values.

Overall, I prefer Handbrake’s queue over VEAI’s source view, since Handbrake has the following features that make the user experience much better:

  • Selecting and reordering videos in the queue using either the mouse or the keyboard works like a file browser, which means that users don’t have to “relearn” those tasks in Handbrake.
  • Each video in the queue has a clearly visible icon that displays its status (in the queue, processing, etc.).
  • The queue uses a compact list view that works well with a large number of videos. (VEAI should keep icon view as an option.)
  • Pause/resume functionality.

The main disadvantage of Handbrake’s queue is that one has to “load” a video from the queue to edit it. VEAI shouldn’t have this limitation because its right pane is always active.

Please, please work on the timeline. Right and left arrows should be frame right and frame left. We should be able to enter specific timecodes, down to 100th of a second (hh:mm:ss:xx), OR specific frame numbers to jump to. Without this, finding what you need in a source or preview is TEDIOUS, and it’s so easily fixed.

5 Likes

Still the same dilemma: Only 1 video gets processed, 2nd video in parallel GUI crashes within seconds.

New Feature - Face Recovery: I’d like for face recovery to be added, like in gigapixel ai.
New Feature - Out of focus Recovery: I’d like for out of focus recovery to be added, like in shapen ai, especially repair out of focus, during autofocus deblurring.

11 Likes

Option to disable the preview when exporting is now an essential need as it slows down the process and can create bottlenecks in your hard drive if you are exporting large files or worse with images.

1 Like

The GUI of this version also crashed after a short period of processing. I add the crash logs here.
logs (2).zip (209.9 KB)

And up/down arrow could also be something like jumping/back 50-100 frames in the timeline so it would make much easier to get near a specific point and then adjust left-right for the right IN/OUT cut with I and O keyboard shortcuts.

1 Like

Crop function could even be more useful with a choice of fixed ratio selection (original; 16:9; 1:1; 4:3, 3:2). Actually one has to calculate separately the correct ratio when borders are moved by the mouse. Ok, that’s a minor issue, but could become perfect.

3 Likes

I didn’t see this mentioned, but it was my first thought when I went to install v3: “Can I install this without uninstalling my current v2 installation?” The answer seems to be yes… v3 is named “Topaz Video AI”, while v2 is named “Topaz Video Enhance AI”, and they get installed in folders that match those names.

Most video apps do those sorts of functions using modifier keys, like shift-left, alt-left, ctrl-left, etc. Up/Down arrows are typically used to control the sound volume during playback.

In/Out cuts typically use [ and ] to set from the keyboard.

But yes, it’s hard to understand why these things haven’t been put in place yet, since the GUI was the focus before we started these pre-releases, and that is part of the GUI.

This and Beta 6 sucks the space out of c: drive. (even if temp is set to another drive) I managed to get one video processed to 3%.

2022-09-14-09-12-16-Main.tzlog (12.8 KB)
2022-09-14-08-06-9-Main.tzlog (4.6 MB)
2022-09-14-09-24-1-Main.tzlog (4.6 MB)

edit: the app crashed again, but it took a bit longer. here are the log files. -.-
logs.zip (238.4 KB)

1 Like

Thought I would give this early access version a try. It appears I am still unable to get audio when running an interlaced file through the system using “Auto” settings. Has this shortcoming been fixed and I’m doing something wrong?

I can play the file fine in it’s unprocessed form. I am running Windows 11.

Make sure you have audio set to “copy”, or you’ll need to remux the audio in after processing.

1 Like

Thank you, I will give it a try. However, I note that when within the Topaz application if I chose just to play the original file (not the preview), I also get no audio which would lead me to think it’s a missing Codec, however as I can play the file outside of the application in the Windows Media Player and I get audio I am thinking it’s not a Codec issue but rather something else specific to Topaz.

Topaz doesn’t play audio on source files within the application.

Understood. Using export, I get the following error and the export fails:

Guessed Channel Layout for Input Stream #0.1 : stereo
Input #0, avi, from ‘M:/c2Videos/Ouroboros/RAW/Catalina 3.avi’:
Could not find tag for codec pcm_s16le in stream #1, codec not currently supported in container
Could not write header for output file #0 (incorrect codec parameters ?): Invalid argument
Error initializing output stream 0:0 –

As mentioned above, I am able to play this file (including audio) directly using Windows 11 media player.
Audio set to “Copy”

Any ideas?