It is true that since this change of team behind this software, they are too focused on changing the interface, and concerning the old models, for several years and months, there is no more update. They are too based on the interface but the rest we have the impression that they do not care. Because many of us want to see updates of all the models.
The interface is nothing more than eye candy, to entice first time buyers. It’s more about marketing and attracting a steady stream of buyers, than improving the models. It’s seems models have been relegated to the basement.
1 Like
It’s a shame that he doesn’t help me that in priority to update the old ones and even new models for me concerning the interface and secondary. Besides, like many Nobody says it here.
I can’t say anything else than agreeing with a too much focus on the GUI, but what I’ve noticed also is that they work with many samples that are allready of very good quality, to upgrade it to insane quality.
To me, that does not fit users needs. And such behaviour may lead model to be lazy in details recovering.
1 Like
Very well put. All the UI enhancement and changes are really just minor side grades at best. Concentrate on fixing the bugs, and making the models better for LOW RES video files. The models Iris and Reah need to be revamped to better handle wide shots and zoomed out content as that is where they both seem to struggle the most when working with low res VHS sources and trying to upscale them. I pretty much just stick with Iris as well, bouncing between LQ and MQ depending on my source. I also find I get better results by going from 480p to 720p, then 720p to 1080p. Not sure why but the results are always better doing it that way. It would be nice if you could create a model that would do that all in one pass.
One other thing that I think would be a UI improvement, is to allow for splitting the video into multiple segments so that we could apply a different model to different parts of the video. There are times where my video could use ne model and settings, but another section might change to where a different model or the same model with different settings will be better choice. So being able to “split” the video track in order to apply multiple models in order to output a single video would be a nice feature.
6 Likes
Applying different enhancement models to segments of videos is pretty much the whole point of making VAI a plugin for video editors.
it now ruins my exposure, gives my whole video this bloom look. smoothes out and reduces noise way too much even when noise reduction is set to -100. what happened? completely unusable now. At first I thought it was only a problem with Log footage… but tried it on a final render in rec709 and still trash… please help and or fix.
hey guys i dont normally complain about anything but this new update seems to remove manual adjustments for models that had the option. i cant find them.
also ever since Proteus version 4 wile it fixed the flickering issue. when ever something is moving around on the screen with a static background details are lost in the background. this didnt happen on v3 and im working with uncompressed raw video.
anyways ill be reverting back till you all can get this fixed with the manual control’s. and tbh i could care less about program interface updates i would prefer focus on the models improvements.
Check the “Enable parameters” box for Manual and Relative to auto (renamed “Dynamic”).
1 Like
oh… i must be blind lol. thanks. i just love layout changes.
why did you move my bug report to the Ideas section? all posts in the Ideas section are automatically ignored
I find myself needing more “clicks” to get stuff done.
I used to be able to select custom resolution from the drop down and then prioritize height (that should have been an optional user-defined default to begin with) and then type the pixel height I wanted.
If the goal is to preserve screen space, I don’t see why the other options are there, but greyed-out until we select “enhancement”.
Now, I must first click the enhancement toggle, and then do all the steps I did in the last version.
Agreed! Most don’t realize it, but they are actually paying to be Topaz software testers now… Rather then getting paid to do the job, we are actually paying them for the privilege of testing software for them! I’ve made this point crystal clear in one of the many tech support emails. I understand bugs happen, but this program keeps getting worse it seems and it has cost me a ton of time and real money because of them. The competition is everywhere now, it might be about that time!
1 Like
It finally occurred to me that this “release game” of fix a little and then introduce new bugs is pretty much a scheme. Much like games of chance, you keep putting in the money with the hopes of a payout. You’re simply hooked by the system. Software (some) seems to be the only industry that gets away with this “gaming” model. So, foolish me, here I sit waiting and waiting for better models, with a usable and intuitive GUI. How many $$$ must I payout?
1 Like
Here’s what I think would make the GUI better and more intuitive.
- Codec Settings should be near the bottom of the work flow (Right Panel) or part of the Export dialogue.
- There should be a greater distinction between the tools (Right Panel) so that they are easy to find and read. For instance, a bit more space or a horizontal line.
- Again, logical workflow should be from top to bottom.
- Presets and crop remain at the top of the right panel
- When you open an adjustment, the parameters should be offset (indented) slightly. Example:
Enhancement
_Pixel Type
_AI Model - Version of AI Model should be below the AI Model or drop down selection to the right.
- Parameters. Odd choice I think. How about Model Settings?
- Frame settings should higher up the right panel , I think.
- Carousel. Even though it can be toggled on/off, I don’t see how it helps much with just images. Dialogue and explanation would work better. Hover could do that
Gotta select them all, then hit cntrl e.
My update license ended in March. Since then, NYX 3 and Rhea have come out. There’s an SDR to HDR model in the works and even Rhea v2. So far, I have been able to use the trial mode in the Beta and Alpha versions to decide if it’s worth renewing my license. None of the new models have proven to be better for my needs yet. As for the GUI changes, they still have not made it so that we can define multiple sections, and apply per section the AI models and parameters. My script to generate commands for the CLI, and run them, remains superior for my needs—and has the added bonus of rarely needing any changes for all the updates.
My main point being: I (and therefore you [or anyone]) don’t have to buy the latest version of TVAI to figure out if it’s needed. With the public access to Beta and Alpha builds, you can get a look at what’s coming up. Money only needs to be spent on TVAI after the desired functionality has been proven to exist.
3 Likes
Thanks for the reply and your thoughts.
Q: What is the last version you’re using before the expiration?
Version 4.2.2.
My license don’t end, but I use 4.2.2 too - You are not alone 
