When do I need to use motion deblur and stabilizing
In other words what is the difference between camera movement and camera shaking…
what is the difference between themis 1 and themis 2 in motion deblur
I found out that I raise questions that others also have and are hesitant to ask which is not good
From what people have said on this forum, I think Themis 1 becomes unstable on clips longer than 10 to 15 seconds. Maybe it was once a big shake happens, it loses track and never gets back on track.
I think it’s only available in case you used it with success in the past, and cannot get the same results with Themis 2.
(Edit: I think I got this confused with the stabilization model.)
That is a really good question!
As you may already know, Themis v1 is a legacy model that was replaced by Themis v2 due to some issues that were reported V2 will give you better image quality and will work consistently. I would not recommend using v1.
You should use Themis when your file has motion blur which is the apparent streaking of moving objects in a file.
No one should feel hesitant to ask question on the forum but if that ever does come up, you can always message Support directly 
I see that When I do Iris 1.0 X 2 and Themis 2.0 the time is acceptible for conversion but if I choose 4 X the file is so hugh the time to go through the file is unacceptible long looks like for many features stay on HD or 2 X otherwise the file is hugh and any additional processing will take an incredible amount of time
Despite I use H265 the file is much larger than I would expect after 4 X increase. If compared to TV 4K recorded H265 coding
furthermore I did not see any improvements using motion deblur
Depending on the file, upscaling at 4x would create a very large file and would take much longer than 2x, etc. What you can do, is process the file in two passes, which may speed up your processing time.
It seems that many of your comments are targetted toward a specific video file. Each model was trained for a different use case and Themis may just not be needed for your file.
I would recommend opening a support ticket and sharing your file for testing.
Since we have no best practices or context sensitive help like a flowchart or even better a automated plan of action determined by the properties of the source file . I am learning it the hard way.
In my opinion support is for bugs and use of the applications must be at least supported by formalized documentation.
I worked in for several companies and it seemed some had their tests fully automated while others did not since we have every week a new version i belief automated testing is not possible any more
In the mean time I will use X2 or HD and use the new default encoder see the production version where h264 10 bits is the new standard.
And not use H265 anymore which I always assumed to be superior not only in compression but also better in quality.
Last but not least very happy with Iris 1.0 and the combination with blending set to 35
With X2 h 264 10 blending 35 the conversion is pretty fast and allows me to use more features. I will let my enigma 4 k box do the upscale to my 4 k lg tv which does its own AI magic to create hdr effects.
Hi Anne,
Great feedback! We actually have an amazing Docs page which you can find here. There are tips for getting started, in depths explanations of the models and settings as well as a few examples of workflows within the app.
We operate on an agile development cycle, meaning that we release timely updates to quickly patch issues and release new features. This way our users are not waiting months (or even years) for minor releases. Of course, some issues take a bit longer to resolve, however, we have great Roadmaps that serve as a map of what we are working on.
I would need to see your specific example and test this on my end.
Thanks for the Docs Page might it be an idea to add this link in the Topaz AI photo Topaz Video AI
I do not think it is there even worse I did not know it existed I love the examples . For now I have added the link to my favorites in Edge
The Roadmap link is a very nice to have and easy to read from my mobile as well.
Maybe it is me but I was a bit misguided with the existing Help Item cli Documentation which also sends me to a Topaz AI documentation but it is very rudimentary Not sure why The above is not added in the help. maybe i am overlooking something
Thanks for this it saves me a few you tube searches
Hi Anne, yes, this link is already in Topaz Photo AI & Topaz Video AI under Help > Support
We revamped all of our docs a couple of months ago and are slowly migrating all docs to the same place. There is a brand new CLI on the new docs page as well.
The CLI is clear but Did not expect that support would translate in a combined topaz AI photo Video page where I assumed it should be Manual or user guide and why is the link not made to AI video now it is somewhere down the line you have to scroll
Then you need to Find Docs Then Somewhere it tells the good things use artemis for HD to 4 K upscaling I can assure you no one has read this why i see everyone using gaja and artemis has many features options still not always clear but i like artemis but learned it the hard way I saw it is a bit slower hence must be doing the job better
I would revamp these selections they seem unimportant from a technical point of view but the goal must be to simplify complexity and this also means ensure that there are no misnomers such as support and no products added that obscure the navigation. anyway thanks for your kind reply
yes a revamp of this site looks necessary and needs to be more to the point when navigating and have for each and every AI Model and for each Selection have proper examples.
I am still sometimes surprised what is computer generated or high quality with Gaya what are we differentiating here, i won’t go that far to forbid pictures which I do in my work to ensure the text is clear but it can be done in very simple steps a bit better
thanks for your patience
Hi Anne,
I am not clear on what you are expressing in your first paragraph. This Doc’s page is linked in both Topaz Photo AI & Topaz Video AI under the Help menu as well as on our website at this time.
Our site was recently redesigned and the final product is currently live and has been for a while.
The docs section is under Support where you will also find quick and easy ways to connect with our Support team if the answers you are looking for are not found here.
Gaia CG is for computer-generated graphics/ typically art, illustrations, cartoons, etc. And HQ is for high-quality inputs. There are in-app tool tips for each model as well if you want a quick guide. Hover over the section to read the tips;
I think I was to the point if I have Topaz Video AI product I want to directly got it pointed to topaz Video AI and I am not interested in Photo AI.
I think support should be called Manuals and be as specific as typed out.
I like the slick web page but it is for marketing purposes only.
I have advised more on context sensitive help then others and this should be more then a short sentence not to speak about the way I have to hoover around it
Amen.
Amen
