Why can't we set TCB/LINEAR per key? Last update for this was back in 2014?

Started by pixelsmack, May 06, 2022, 02:34:42 PM

Previous topic - Next topic

pixelsmack

I searched the forums for why the keys can't have separate LINEAR/TCB options and the last thread was from the dev saying...

"Yes, your assumptions about how things work are correct. We have improvements planned for the curve editing." 

That was back in 2014. Come on guys!

WAS

If only objects being enabled/visible etc was keyable, this could be overcome with separate objects on different modes. Too bad that's not possible either.

What I've been starting to notice about TG, is we kinda live within someone else's vision of how it should be done. Because it's done by one person, who doesn't differ to a team, who then looks at things from an end-user perspective as a collective. So things will be kinda hard-set to an opinion on how they think things should work, rather than approached with "What will the person want to do with this besides what I envision?" Kinda like the Transform Input shaders not having any inputs and relying on manual input, which then makes the whole of TG incapable of dynamic functional scaling, rotating and translating without re-inventing the wheel, which I found may not even be possible, I've tried to get help with translation, rotation, and scaling via functions but no one knows, and when I port logic from other node-based DCC, it doesn't work as epected (like my Rotation Matrix from Blender which only works on texture space, and useless for PFs etc)

I don't really mean to bash on Matt here, but some things confuse me. Like he once told me he didn't want to improve a shader, because he hadn't literally touched the code in over a decade. To me, that seems backwards. I've always been taught in various coding classes to always return to your code as you advance in your language, to make sure you aren't doing anything considered wrong, deprecated, etc, but also more importantly, to see how you can improve it.

Terragen's Atmosphere system can't be matched in any other DCC that I know of besides maybe now Unreal Engine 5, but I'm sure a lot of that is faked to screen-space. Maybe I'm wrong. But how long will this remain true? Already Blenders Adaptive Subdivision for stuff like terrains is sooooo much faster than Terragen's that it blows me away. All without the weird shadow tears from bad geometry or needing to up a ray detail multiplier that makes the scene even slower. It only works in cycles right now I believe, but it's pretty crazy how well it works. I feel like currently TG is aging dramatically faster than it's development, and at the same time, Matt is trying to look at other things now, like before Terragen Canvas, and now that that's gone silent, this Terragen Sky (which seems useless to me, there is a reason there is no such thing as a HDRI generator, element you may need may require a ground, or objects, etc.)