[Micky C]

Trooper Dan, on 29 September 2018 - 04:58 PM, said:

Is there a particular size cutoff for what makes a texture eligible for the atlas treatment? A typical Eduke32 project will contain all the classic textures, plus thousands of new but classic-styled textures (usually 256x256 or smaller, but not necessarily), and then some higher resolution textures as well (e.g. for cutscenes, menu etc.)

The AMC TC has over 20000 tiles. I'm fairly confident that the vast majority of new textures would be power of two, but there may be some outliers.

In fact, I'd suggest that the AMC TC would make an ideal stress test case for PolymerRTX; several of the maps are very large, highly detailed, use a variety of unusual constructions, and of course draw from a huge pool of custom art assets, with countless custom code effects like 3D skyboxes and alternative pals.

Mark, on 29 September 2018 - 06:49 PM, said:

Thats got me wondering. With your best guess, will there be much difference, if any, between how a vanilla map in software or Polymost will look compared to in your new renderer? I would imagine the lighting and normal/spec/etc mapping will be an improvement over Polymer along with a speed increase. The point I'm attempting to make is if your renderer will not really offer much over software or Polymost rendering, don't hold it back simply to be compatible with them. Those users will have a perfectly good eduke32 to use.

A new renderer which has flawless TROR like polymer, but with a higher framerate than classic (not to mention correct perspective) will be very useful, even without dynamic lights and other fancy effects.

I'm sure that TX and Hendricks would want there to be the highest possible level of compatibility between the renderers if they're to be included in the main eduke32 branch. You don't want to automatically exclude a wide range of existing content.

Phredreeke, on 29 September 2018 - 07:33 PM, said:

How is Polymost only doing 282 FPS? I don't know the map in question but I typically get framerates in the thousands.

Clear the Coast is a fairly large map with reasonable detail level. I've been working on a large map which has a single-digit framerate in polymost, but is quite playable in classic.

[Mark]

That map he is testing with is known to drag down many computer systems. Oops, didn't see mickyC's post.

MickyC said: "I'm sure that TX and Hendricks would want there to be the highest possible level of compatibility between the renderers if they're to be included in the main eduke32 branch. You don't want to automatically exclude a wide range of existing content."

lUnless something has changed behind the scenes, I don't think our current devs will be able to handle future issues with the new renderer. I hope to be proven wrong.

[Forge]

[icecoldduke]

I'd like to start making a list of projects to test during development of PolymerRTX. If you guys can include a link to download. If your project has more then one or two maps, please include the names of one or two maps for me to try out.

[Mark]

Are you looking to start out slow and test vanilla maps with maybe some mirrors, ROR and TROR. Or are you looking to jump right in to a full fledged HRP project with lots of fancy SEs and custom content?

[Tea Monster]

Something, but only something, like Hollywood Holocaust Reloaded Rethinked.

ModDB link: https://www.moddb.co...caust-rethinked

Download: https://www.moddb.co...ked%2Fdownloads

Screens:

[icecoldduke]

Wow! I'm actually at a loss for words. I'm going to just leave a giant fucking "wow" here. This is beyond great. Downloading now...

[icecoldduke]

I've been working on lighting in PolymerRTX. I loaded in the Hollywood Holocaust Rethinked mod. Got a lot more work to do, but this is a definitely a great mod. Tea Monster/Mark/Spiker would it be ok if I used this mod in my various PolymerRTX vids?

[Mark]

I think I can speak for the team. Go for it. It was originally started for your Polymer NG project so I guess its fitting to try it in your newest project. If you have a preference, there is a version of that map without most of the NPCs and humor. Also I remade Red Light District and part of Pigsty. They are available somewhere in the HHR thread.

[icecoldduke]

Mark, on 01 October 2018 - 05:35 PM, said:

I think I can speak for the team. Go for it. It was originally started for your Polymer NG project so I guess its fitting to try it in your newest project. If you have a preference, there is a version of that map without most of the NPCs and humor. Also I remade Red Light District and part of Pigsty. They are available somewhere in the HHR thread.

Sweet! I really can't overstate how fantastic this stuff is. Seriously guys great work.

[Spiker]

I could probably update some of the assests but only for a "worthy and finished" engine.

[Tea Monster]

Cheers.

How are you handling glow maps? Are you going to turn them into emitters, ignore them and have people put lights in, or what? As an example, the light above the store under the clock is a glowmap and it's going to look weird if they don't cast shadows.

About lights, we don't use enough lights to properly show off the assets due to Polymer's lack of optimisation. On that subject, we used a light hack that disables lights that are deemed 'not in use'. Mark has the details on that. I imagine that the lights will have to be tweaked to properly take advantage of the raytracing renderer.

[Mark]

Oops. Looks like TM beat me to the punch.

Something I should mention. HHR has a light hack feature for better frame rate. The backspace key enables/disables it. When enabled it turns off Polymer lights not in the view or immediate area of the player. I mention it so you don't think you did something wrong on your end.

[Mblackwell]

It's also not quite as precise as it could be with more complex math (that is, checking if areas contains within the lights theoretical volume are visible in some kind of plot). More complicated math would increase overhead per light, and since it's from scripting I couldn't guarantee that it wouldn't in fact end up slower.

[Danukem]

The current Polymer seems to suffer from massive overdraw. In a big, complex map, fps will drop to single digits even when staring at a wall in an isolated room, simply because the rest of the map exists (the first level of DNE is a prime example of this). The ability of scripted light hacks to improve fps is more evidence of overdraw -- why should turning off lights in areas which aren't even visible improve fps? I could understand it if some of the light from those areas were visible, but iirc it helps fps even if the lights being turned off are completely off screen.

[Mark]

Yep. I remember a discussion with Plagman concerning some Polymer lights not showing in a HHR map. Turns out that some kind of calculation for light SE distance in the code made it think more lights were visible in a given area than actually were. So he adjusted that distance and all my lights started showing properly. So yes, even lights close by but behind a wall counted as visible lights in that calculation. So I had to bump up maxlights settings to a very high number before the fix. IIRC I had maybe 10 visible lights in a certain part of the map but Plagman was able to determine that something like double that were being detected as visible. Also, in a completely isolated area drawn outside the main map still slowed down because of rendering elsewhere in the map.Or something like that. I hope I was almost clear with my high tech explanation.

Framerate in area B was affected by area A.

Attached thumbnail(s)

• 

[Romulus]

You may want to update to Windows 10 version 1809 if you haven't done so already.

[Mark]

I'm running win7. If you mean someone else, never mind.

[icecoldduke]

Hey guys, just a quick update, right now im trying to figure out the fastest way to update the DXR acceleration structures for moving sectors. Right now I'm getting quite a bit of hitching and I don't know why. I reached out to a couple friends I know to see if they can help get past this issue. Docs on DXR suck right now.

[Micky C]

Why is the problem specifically with moving sectors? I thought one of the big advantages of ray tracing was its ability to handle dynamic situations in a more straightforward fashion?

[icecoldduke]

Micky C, on 20 October 2018 - 09:04 PM, said:

Why is the problem specifically with moving sectors? I thought one of the big advantages of ray tracing was its ability to handle dynamic situations in a more straightforward fashion?

Yes I shouldn't be having this problem. Basically because the geometry changes as sectors move I have to rebuild the acceleration structures, which shouldn't be a problem(since skeletal geo is used with DXR), I'm just doing something stupid its that simple, I just don't know were yet :/

[TON]

Do you have this doc?

https://github.com/N...t/14-Refit.docx

Download document


_____________________________________________________________________________________________

The DirectX Raytracing (DXR) specification v0.09 could be also interesting, especially the section 6.2.7:

http://intro-to-dxr....Spec_v0.09.docx

[icecoldduke]

TON, on 23 October 2018 - 10:07 AM, said:

Do you have this doc?

https://github.com/N...t/14-Refit.docx

Download document


_____________________________________________________________________________________________

The DirectX Raytracing (DXR) specification v0.09 could be also interesting, especially the section 6.2.7:

http://intro-to-dxr....Spec_v0.09.docx

I did not have the functional spec, thank you! .

[Romulus]

I am guessing another AAA title is being worked on since October.

[Mark]

who knows...sigh

[MusicallyInspired]

Don't get your hopes or expectations up and you won't be disappointed. When (or God forbid, if) he comes around again with a release it'll be a pleasant surprise.

[fgsfds]

Fool me once, I'm mad. Fool me twice, how could you. Fool me three times, you're officially that guy, okay? You know him, you know the one. You go up to Duke4 and he's like, "So I did not have time to work on it over the break, due to some work contracts I have I can't actually release any vr related code" Fuck you! I aaaaaain't havin' that shit!

[Forge]

[OpenMaw]

Forge, on 07 December 2018 - 07:18 AM, said:

I'm sorry.

Yeah. You better be.

[Tea Monster]

fgsfds, on 07 December 2018 - 07:16 AM, said:

Fool me once, I'm mad. Fool me twice, how could you. Fool me three times, you're officially that guy, okay? You know him, you know the one...

Yeah,I know the guy, he's the one that fooled us FIVE times. What does that make us?