Roger that.
That’s unsettling. Give a model a new name (but really just updated) and your software will be absolute after a year.
I think this should turn into a subscription model now.
I probably have an odd sense of humor, but these forums remind me cryptomining forums.
No smoke and mirrors subscription “deals” please. (I have a very low opinion of that marketing scheme, but don’t wish to segway into a discussion of that just now.)
However, if you own a license, and are past the auto-upgrade period, you should be able to reload it. Including the last available models for that software version.
1 Like
Should is the keyword.
It is my position that 2.6.4 is a different program than 3.0.
And it appears the EOL for 2.6.4 has ended without warning.
They even changed the name.
Why are we not at version 1.0 - when it has been admitted that this is an entirely different program rewritten from scratch?
It is 4:43pm in Portland OR - Saturday afternoon.
Things evolve. This new version of Video AI has capabilities that the older ones lacked.
It still needs some adjustments and adding a few little niceties to make it into a handier tool. It has the ability to do multiple operations in one pass, which is important, especially where lossy CODECS cause image degradation at each pass. I am certain this evolution and future refinement will evolve into a better-fitting tool.
As for the current speed issues, some of these are caused by the new architecture and the new models. The speed and effectiveness of these will improve over time.
TVAI 3.x is simply the third generation of VEAI, but it has been rewritten and accordingly renamed.
1 Like
The model failed, repeated over and over is not a problem. That happens when it is trying to find a model that fits the settings and media. Eventually it will find an appropriate one and use it. - It is when it gets to (model) failed and just sits that there is a problem.
Proteus is evolving. Where did you ever get the idea that it won’t?
Agreed but more troubling is the old version of Topaz 4.7 also no longer works with audio either and it no longer weirdly has a play button only pause and forward and back it mega annoying. Though thank you for your reply
Lliam,
Yes, the audio did work in VEAI v2.x, but not reliably. - (I’m assuming your referral to Topaz 4.7 is just a typo. - I wouldn’t think v4.x is even on the drawing board yet. But there are many new areas that it might include; such as AI colorization, for one.
I just managed to finish my first complete video project since TVAI first rolled into Beta, and I’m rather pleased with the results. I hope that more of us users, come forward and suggest some simple features that might be changeg in the 3.x GUI to make it a bit friendlier to use. And that in the meantime, the crew writing the AI stuff get into streamlining their code so it will do the same quality work but faster, please…
1 Like
Switching from 3.0.6, exports always have an ffmpeg error
Last FFmpeg messages:
-vsync is deprecated. Use -fps_mode
Passing a number to -vsync is deprecated, use a string argument as described in the manual.
Input #0, matroska,webm, from ‘C:/Users/breez/Desktop/336ec494422113f0aaf5e879470603ac.webm’:
Unable to parse option value “0” as video rate
Stream mapping:
Where did you ever get the idea that I said Proteus won’t be evolving? 
All I said was, that should the name of the model change, then this, by no means, mean the actual model data will change. We may get a Proteus 4, one day, who knows; to which I then suggested that at some point newer models may arrive that no longer fit into the current versions of TVAI. And that doesn’t feel unreasonable to me: if your license were to epire today, then I don’t feel Topaz is beholden to let your current version of TVAI run all future models as well. The software will progess. I then proceeded to say, that even when that day arrives, such a fact will not render your current TVAI instance obsolete: it will still continue to run.
With this new update I have to login every time i start the software. As I have frequent internet dropouts this can get particularly frustrating. Also there was a period when your site was down for maintenance, which meant I couldn’t use the software during that period either. I’d advise you make the web login occur like once a week at maximum.
For me, it is currently only the Preview that needs urgent fixing. Unlike others, it seems, my TVAI (3.1.02.m) doesn’t crash, and I’ve processed literally dozens of full-length old BluRays with it. But the Preview is a disaster, frankly.
And I also don’t understand why Preview has to slow down everything so tremendously, when playing. I mean, we’re not talking about high-end gaming: it’s just what should have been a trivial Preview function. My i9 12900K isn’t bogged down for the full 100%, nor is I/O suffering outrageously: aka, Preview should work without any noticeable reduction in TVAI performance. To the powers that be, please, fix this!
I have only been using enhancements for any ‘real’ work (Proteus). Tried motion blur reduction and such too, for test purposes, but that slowed matters down significantly (as announced by Topaz, to be fair). To about 50% processing speed.
What is the difference between “Add Noise” and “Grain”?
I am no expert on these matters at all.
The way I alway understood things, though, is that ‘Add noise’ (like in Photoshop) tends to add monochromatic noise, to enhane contrast or create a type of dither effect for low-res material. ‘Grain’ should be [an estimation of] true, ‘multichromatic’ (that can’t be a real word?) movie grain.
I’m sure others can explain this better.
Hi, I’m running Windows 10 and trying to convert a DVD mkv movie to HD. I’ve done this successfully in the past. However, in you latest versions it only appears to work with setting Encoder ProRes 422 LT and container = mov. If I try H264 or H265 encoders and container mkv or mp4 it always says model failed and the Topaz Video AI app disappears. Is this a bug or am I doing something wrong?
