Topaz Video AI v3.2.7

I also want to know. My guess right now is that it will be after they get the scene change detection implemented.

apollo works fine (30 fps to 60 fps)

But if there are duplicate frames, it really messes up the motion.

The output video is larger than 7680 × 4320. What format should I use? If the video is encoded with ProRes, the playback of the video is very stuck

Hi there, I did not answer your question about Proteus since @ForSerious did a wonderful job answering your question.

As for the results you are seeing with Topaz Video AI, there is a good chance that the correct model has not been applied or that your specific use cases are not a good fit for our current models. Without seeing samples of your file, I am unsure which is the correct assumption, however, I would recommend reaching out to our Support team for assistance finding the perfect model if our docs are not getting you the answers you are after.

1 Like

Hi there, I would take a look at our most recent road map post. This section of the forum will be the best source of information as far as what is next for Topaz Video AI :slight_smile:

Does anyone other than @Imo also have the advancement slider offset problem? I want to know if this is a software bug or an isolated problem on one machine so I can decide if it’s worth upgrading to this new version.

1 Like

In some cases, you may need to raise the sensitivity up to 100, this is video-specific and we see this most frequently with very noisy files, or highly compressed files, dusty/scratched, etc.

Nah. Not all files have this problem. It all depends on the source fps. I improved the source at 23.976 fps to 120 and it did not duplicate frames. But when I rendered the source file from 30 fps to 120, it started duplicating frames. Not always, but it happens.

I used a time remap in After effects and made a changing picture for each new frame. It should not have duplicated frames, because there are no duplicates in the source. You have to set the sensitivity to 100 when there are duplicate frames in the source, but I didn’t have that.

Basically, this problem occurs only in Apollo presets. I have not noticed it in others.

Yup, I just keep reporting it each time, really appreciate all your comments, they were much needed in this forum!

1 Like

Why on all videos processed by TopazAI, Mediainfo shows only 21,5 Mbs, if I have 180 Mbs in the settings ?

1 Like

Can someone say why the color space changes so much? After processing DaVinci, I process in TopazAi. And the maximum bit trade that TopazAi gives is 21.5Mbs

DNxHR444 _apo8_prob3 444 12.mov_snapshot_00.00.000
DNxHR444 10.mov_snapshot_00.00.000

3 Likes

Are you working with 8bit rec.709 in Davinci ?

Please advise whether it is better to import from DaVinci to TopazAi in the DNxHR mounting codec or into H255?

I don’t really understand it yet. Converted to Shutter Encoder to ProRes. 601 to 709. Then processed in Davinci 10 bit. Output 10 bit DNxHR444

Not sure what you are doing in Davinci regarding “Processed” but if you are working with rec.709 you need to setup colour space transform nodes so your working enviroment is within Davinci Wide Gamut and a final colour space transform node back to rec.709 Gamma 2.4

I finally installed the latest version to keep up to date. It seems to work fine for me. And I’ve checked that it starts at frame 0 as on previous versions and it’s fine. It starts at frame 0, so no problem. :slightly_smiling_face:

1 Like

I personally use Resolve after VAI to clean up and correct.