Terragen Sky?

Started by gao_jian11, January 21, 2022, 03:46:52 AM

Previous topic - Next topic

gao_jian11

I saw this on the official homepage, a separate new program, any more info to reveal?

WAS

Wonder what happened to Terragen Canvas. I can't see what a dedicated "Sky" software would bring that Terragen doesn't already do just fine. It's kinda my issue I had with the Canvas idea cause it was also a separate terrain generator program, when Terragen is a Terrain Generator already. It conflicted hard in my brain and seemed more something to save for a Terragen rewrite (new GUI, n' stuff that could handle it).

Still interested in what it is. Maybe it is super useful. Not sure.

Hannes

As far as I understood here:
https://planetside.co.uk/featured-projects/honda-the-origin-of-determination-vfx-by-a52/
it's more or less for lighting in other apps using TG's sun/atmosphere system.

WAS

#3
Quote from: Hannes on January 24, 2022, 02:00:13 AMAs far as I understood here:
https://planetside.co.uk/featured-projects/honda-the-origin-of-determination-vfx-by-a52/
it's more or less for lighting in other apps using TG's sun/atmosphere system.

It seems to just be Terragen's sky system without a planet, to export HDRI's, something you can do by simply disabling the planet, and already only a professional feature. Weird. Most people get TG for HDRI's so seems kinda obvious what's going on there... Kinda sad though, cause it doesn't force them to tinker with the terrain system and possibly expand their tool set.

It would be better off, like Terragen Canvas, to just be another tab "HDRI"; where TG simply disabled the planet and object systems for a HDRI system and volumetric.

Even then, will it truly be fully useful? What if you need objects in your HDRI? A generated or textured moon, spaceships in orbit, or the like? TG would be able to do it all fine, would Terragen Sky? And at that point wouldn't it be redundant?

mhall

Total conjecture ... but Matt did survey the user base some time back (couple years ago) and GPU accelerated sky/clouds were a hot item.

https://planetside.co.uk/forums/index.php/topic,28043.0.html

Perhaps this dedicated application features new code to move sky rendering onto GPUs and substantially accelerate it?

That would be worth a dedicated application, I think (guessing that trying to pull all of Terragen onto a GPU would be a massive undertaking and not likely given how much memory a large TG scene takes). Particularly for shops that are just looking for top notch skies to use in/light effects shots.

Is there a reason a dedicated HDRI couldn't be used as an environment light/background plate in TG itself? (I'm really asking - I don't know the answer to that).

Get faster renders in TG Sky, use results to light TG scene ... faster results because you don't have to render the sky every time on CPU, but can render it on GPU when you want to change it?

Again ... all of this is speculation on my part, I have no information.

WAS

You can definitely use HDRI in Terragen, and they work well.

And you are right, skies were a hot ticket, which means it substantially will impact TG sales, as that is what people get TG for. Besides us people on the forum, people get it, ask a couple questions about atmospheres, and are gone, cause they got the use they needed.

From a business stand point, it's a absolutely terrible decision which will ultimately run down Terragen which is already aging too fast and becoming antiquated.

And TGs framework is very flexible, there is no reason atmospherics couldn't be GPU accelerated within TG and let CPU chug at terrain or objects. In fact this would speed up Terragen significantly as atmospherics are the main hindering point (when not working with ridiculous lateral stuff).

Imo a dedicated HDRI app won't be as great in the end. They don't exist for the simple reason a lot of HDRI in film and such contain unique elements. Think Star Wars, Marvel, etc, etc. Whether planets, effects, or what not. Even HDRIs done in TG recently for film have elements that seem they won't be possible in Terragen Sky without Terragen to handle it.

I am also worried this is just the end of Terragen. Things have been to slow, and to hear about Terragen Canvas years ago, and then now this, while seeing nothing substantial done to TG, it's really shocking.

zaxxon

Actually, I'm very pleased to see that Planetside will offer such a product.  As WAS said TG is "aging too fast and becoming antiquated", while that may be a bit harsh, the large number of competing apps for "world creation' in the market today is amazing.  With all due respect to WAS I don't think TG Sky is a "terrible decision".  What the actual sales numbers are for TG and the financial position of Planetside is not public information: however the competition is fierce and several of those competitors offer their products for 'free'.  I, among others here, have invested a major time of our computing lives to one version or another of TG over the years.  Thanks again Matt, you have provided an opportunity and a remarkable program for creating marvelous imagery!  I doubt that Matt is ever going to stop developing TG, I would guess that it's as much his 'child' as any such long term investment of time and passion could be. The creation of a product that capitalizes on TG's strengths and can fit into production tool kits, serving unique and necessary parts of a commercial environment seems to me a sound and necessary step to continue the actual strengthening of Planetside, and the continued development of TG.  I've drifted away from regular TG use, but watch here regularly to see if the hinted at additions of GPU rendering and enhanced export capabilities are going to happen. I truly hope so. 
 

WAS

#7
Quote from: zaxxon on February 25, 2022, 01:08:08 PMActually, I'm very pleased to see that Planetside will offer such a product.  As WAS said TG is "aging too fast and becoming antiquated", while that may be a bit harsh, the large number of competing apps for "world creation' in the market today is amazing.  With all due respect to WAS I don't think TG Sky is a "terrible decision".  What the actual sales numbers are for TG and the financial position of Planetside is not public information: however the competition is fierce and several of those competitors offer their products for 'free'.  I, among others here, have invested a major time of our computing lives to one version or another of TG over the years.  Thanks again Matt, you have provided an opportunity and a remarkable program for creating marvelous imagery!  I doubt that Matt is ever going to stop developing TG, I would guess that it's as much his 'child' as any such long term investment of time and passion could be. The creation of a product that capitalizes on TG's strengths and can fit into production tool kits, serving unique and necessary parts of a commercial environment seems to me a sound and necessary step to continue the actual strengthening of Planetside, and the continued development of TG.  I've drifted away from regular TG use, but watch here regularly to see if the hinted at additions of GPU rendering and enhanced export capabilities are going to happen. I truly hope so.
 

Respectfully, I disagree. I also find it odd you talk about from a "commercial environment". HDRI generators are not a thing for a very clear reason. The opportunities in it will be immediately limited over any other DCC, which can already do skies plus everything else you'd commonly need to do in your HDRI. If you needed just a plain sky, we got the real world for that, which is still paramount over generated. Generated HDRI usually have some element that you can't get out of an empty sky. Planets, different stars, nebulas, whatever, as well as other static objects you may have. And then if it's just a stylized cloud you're after, that's what OpenVDB is for, and will be used before static clouds baked into a HDRI.

I've never heard about the need for a dedicated HDRI software, because you'll mostly need stuff beyond just a plain HDRI. Like even HDRI's made in Terragen for film many I have seen, even just for games, incorporate more than just the sky.

And as far as development goes, I mean, no offence to Matt, but it seems clear he is trying to come up with new avenues to generate income (Canvas, now Sky) and this whole time there has been actually no real additions to Terragen. Just bug fixes and imrpovement on existing things that had issues. GGX was great, but that honestly should have been a thing years ago. And eventually a developer will be so behind, what finical good is it to maintain it? Enter looking into new things. Which, I dunno, seems like exactly what's going on. I've been there myself. It's a common thing in developing stuff. Eventually you look at it and go "Meh, I wanna do something else".

it seems like a story told time and time again. As is, I don't see anything Sky will be doing that you can't already do in TG, and make it easy for yourself with a default project if all you're doing is Skies. It's inherently redundant right now unless it offers great aspects TG can't do. And as far as any DCC that already can do this, it's not like they are not optimized to do it better than TG as is, and relies on Matt implementing something competent against those GPU accelerated renderers. Which may be a process, not just an out of the box thing Matt writes first time round. Lots of risks here, without much reward. It doesn't make much sense from a pipeline perspective, business perspective, and community perspective waiting on TG to get with it.

pokoy

The thing I'm wondering about is why are there mentions of these products in demos/features but no specifics given anywhere else? Other than some mentions of 'cool stuff is about to come' for quite a while now.

I'd be really interested in a sky generator product but it doesn't make much sense without control over terrain in some form, and given that TG didn't get updates on terrain generation my hopes aren't high. But again, without information it's hard to tell.

Rather than a sky-only app, I'd be more happy to get better tools for generating clouds and terrains right in TG but TG's foundation and its procedural-only approach, plus the nature of a node-only workflow with the far-from-instant OpenGL tessellation preview would all need a major overhaul. Right now, I'm about to decline a project request on cloud/sky animation because I know exactly the client will ask for art direction and TG - while it can render nice clouds and terrain - just lacks any of the tools needed for a meaningful art direction and I don't want to frustrate the client and myself.

I agree with WAS, TG is aging really fast right now. I'm still waiting for VDB support (import, export and on Win platform without tinkering with Linux) and some 2-3 years later with next to none official news and little participation from Matt here, I am about to give up any hope.

And to think we can't even zoom in in the frame buffer to inspect the image, no live post processing options etc... yeah, so many places to improve.

zaxxon

Well said Pokoy.  Perhaps it's time for someone from Planetside to provide some more information regarding any new products and maybe some hints as to the future of Terragen.  You maybe right WAS as to the limited appeal that Terragen Sky would have in the market place. Without much information, as Pokoy says, it's hard to guess what the full capabilities of a possible Terragen Sky application would be.  My initial hope upon reading the Planetside note was that the app would have the ability to create and export VDB's (without "tinkering with Unix", thanks Pokoy!).  Whether Matt will continue to move towards some form of GPU render/preview remains to be seen; I hope that's the case, but some form of TG that provides aspects of TG that are complimentary with current pipeline tools would be greatly appreciated.  

WAS

#10
Yeah we have yet to see what TG Sky really is. I hope not what you'd think not just bare bones skies. Hopefully more artistic approach to clouds, and yeah GPU acceleration and import/export of VDB. At same time we are waiting for that alone in TG still. I thought Canvas would be a great idea. Just not a separate application. It feels like it should just literally be a tab "Canvas" within Terragen that has a toolset and larger RTP (and faster) to work in with tools. Similarly I think HDRI could also just be a tab within Terragen and easily work with what's already there, limiting some on-screen stuff like the planet etc. And again, if tools are involved for a artistic approach, a better RTP.

I think Planetside Software would really benefit from a development blog. Both for exposure and something for other entities to share, or construct news from, but also for the community here already. It'd be nice to know what's going on. I'm saving to update my licence (actually for the second time. Life is throwing curve balls), but at the same time I have a feeling of hesitation, like development is going to stay at a crawl, and fundamental assets put on the side burner for other projects and bug fixes.

WAS

I should add, I can't say enough how the new work on the documentation has helped. In fact we aren't getting the same routine questions we usually got before concerning some basic stuff cause there is adequate demonstration on the wiki. Bravo there, for sure.