I see how to set In & Out points, but it appears this is for outputting only a set region of the video. To apply specific filters to specific regions of the video?
In my use case, I’m dealing with an 80s VHS import. Iris MQ works GREAT for close-ups…but in some longer shots, it’s disastrous (due to just bad video quality). In these long shots, I want to apply Artemis.
I guess the only way would be using the trim function on each scene and process each scene by itself as a separated process, then you stich all the rendered videos together with LosslesCut too or similar.
some people do it like this:
encode the the footage with different settings/models, then overlay them in a video editing program and determine desirable opacity% settings for each scene.
and if you use a professional program like premiere pro, it has extensive transparency/masking capabilities, so you can make single parts in a picture partially or entirely transparent, motion-track them throughout an entire scene, and let the video track underneath it (different encode) shine through. it is lots of work, but the results speak for themselves - plus, you can do all the other necessary editing in the same step of the production process.
This is good suggestion for troublesome repairs. Thank you for the suggestion, @DineshRepInt.
At this point in Topaz’s existence, it’s a bit ridiculous and tedious to have to go to such efforts to simply apply 2 different models to 2 different sections of a long video. I don’t think we’re asking to move mountains here in requesting it as a feature in the existing GUI. It’s an obvious improvement that I’m guessing many of us very quickly realized would be useful.
more than once i blended differently encoded materials with certain % of transparency, in order to get the “best of both” encodes as my final output. plus, this can change from scene to scene.
therefor, having multiple whole encodes is quite an advantage. you can easily come up short if you insist on doing a cutting job within topaz.
Ahhh…I need to clarify. I’m talking about putting 2 different models on 2 different sections. For example:
IrisMQ at 00:00:00-00:30:59
Artemis at 00:44:01-00:50:00
Should easy to do right in Topaz.
i understood what you said.
i too thought about having different sections for different encodes in a summed up job before.
it would mean to “simply” line up different encodes in a single job and join the accumulated partial encodes in the end. ffmpeg can do all of that, no problem, but i´d predict you won´t be truly happy with the outcome, since there is no guarantee whatsoever that the different models, spatial and temporal, will deliver appropriate transitions between the different parts. you might easily end up with an image transition which seems to “jump” a pixel or two to the right, or a movement that is cut short by a frame or two… and you wouldn´t know until the entire job has been finished.
weighing the time for 2 full encodes, which i know i can use, against the possibility of shortcomings/defects of a parted-and-joined job made me reconsider my initial wish.
I’d prefer to have the option and take my chances (and would just apply models around cuts in the video). I just don’t get why it hasn’t been implemented already. Multiple filters are already able to be used. 
I’ve used duplicate clips with masks, opacity, chromakey, etc. I can see this being a good option if users could use multiple in & out points in topaz, apply a different model to each, then have Topaz output just the clips that are changed as separate files. I guess the only work-around right now is doing the cuts, then sending each one as an independent job to Topaz.
maybe you can develop a good workflow with losslesscut, as @Akila said.
I speculate that the changes they made to the UI in TVAI 4 are to eventually add this kind of functionality.
Of course even if that is true, that doesn’t help you right now. And unless someone from Topaz confirms it, it could just be wishful thinking from me.
@ForSerious Hope you’re right! Not urgent, but would be really nice to see over the next few months maybe. 
