Power users looking to profile on different GPUs: the controls and toggles, as well as a guide on how to get accurate results are now in: Shaders/ws30-ALS-landclass-search-functions.frag .
Changelog:
- Move landclass search functions in ws30-ALS-ultra.frag into a shader include: ws30-ALS-landclass-search-functions.frag . These controls should be removed when appropriate.
- 2 development controls to reduce haze and/or lighting are left at the top of ws30-ALS-ultra.frag .
(cherry picked from commit 5196dbd4ebd5f586a7261dc3ae46aa50218caf91)
commit e131ffaca7b2c934474f0e447b5852cb4c7c8d4a
Author: vs <vs2009@mail.com>
Date: Fri Oct 15 22:19:42 2021 +1000
WS30: landclass and texture lookups in fragment shader:
Changelog:
- Testing: Procedural random landclass function, get_random_landclass(). 1). Avoids a texture lookup for profiling. 2). Creates squares of configurable size with controllable random landclass ranges - to test landclass transition algorithms and quickly profile lots of landclasses in view. It's also possible to put pressure on GPU memory even with 1 lookup of ground textures per fragment, by reducing size to <20m and viewing from altitude.
- Add ALS ultra technique as no 5, move existing ALS technique to no. 6. Add WS30-ALS-ultra frag & vert in preparation for eventual ALS ultra port. Change comments/variables to match using a texture array instead of atlas. Sampler name: atlas -> textureArray.
- Testing/optimisation of de-tiling: WS30-ALS.frag: reorder perlin detiling to make noise source easier to switch. WS30-ALS-ultra.frag: Create alternative to perlin noise used for reducing tiling that avoids a texture lookup - to profile and optimise. This is commented out. The performance difference in the final version depends on memory pressure versus arithmetic load of the full shader. It varies by GPU and what's in view. In future, once a full ALS port is done, maybe it'll be more efficient to run a higher quality de-tiling algorithm all the time to avoid branching overhead due to effects like agriculture which need more de-tiling. De-tiling code in shaders should /eventually/ match CPU-side de-tiling of texture masks used for object and vegetation placement. De-tiling code in GPUs might be subject to precision and numerical artifacts, and if so maybe the texture lookup version would work better. The same applies to noise used for rendering WS3 water waves and collision queries. One option is to use exactly the same glsl code, with a C++ header to convert glsl syntax e.g. https://github.com/g-truc/glm .
The light sprite diminished to zero size as the FOV was decreased
(in sim zoom), and increased considerably as the FOV was increased.
This commit fixed this by taking the FOV into account to keep the light
sprite size roughly constant across the valid FOV range.
- model-combined-transparent is deprecated.
- model-transparent can be used to let some rendering pipelines know that
an object is transparent.
- In the case of Classic, model-transparent just inherits from model-default
and the object is rendered normally.
- The G-Buffer layout has been redesigned to be 96 bits per pixel. There are 24
unused bits that can be used for extra material parameters later (like
clearcoat).
- Add better debug views for the G-Buffer.
- Use octahedron normal encoding. This yields the same results as the previous
method but uses 16 bits less.
- Use rg11fb10f for the environment mapping cubemaps.
- Tweak the shadow mapping parameters and add a colored debug mode.
- Only render shadow maps for objects that inherit from model-default.eff or
model-pbr.eff instead of having a fallback Effect. Now transparent objects
should be ignored (if they are marked as such with model-transparent or
similar).
- Remove the separate occlusion texture. Now the PBR Effect expects a single
texture where R=occlusion, G=roughness and B=metallic.
- Add an utility linearizeDepth function to help visualize the depth buffer for debug purposes.
- Use 'color0' instead of 'color' for all color attachments.
- Do not write to the HDR buffer if a fragment doesn't need shading (depth = 1).
- water.eff: technique 2: air_pollution uniform is not used by water-ALS-base.frag
- water.eff: ocean_flag is now a parameter
- water-inland.eff now inherits from water.eff as it's a copy, except for ocean_flag value
- All non-ALS Effects have been removed (except generic ones from model-default and terrain-default).
- The rendering dialog has been reworked to accomodate the changes.
- All necessary properties have been added to graphics-properties.xml
- Five graphics presets have been added: Minimal, Low, Medium, High and Ultra.
- Some unused properties have been removed from defaults.xml
Canvas displays stop working with a technique index of 1000. There is no apparent reason for this, but changing it to 0 or a lower value fixes the issue. There doesn't seem to be any negative consequences from doing this.
- Remove all Compositor Effects and Shaders.
- Unify the low-spec and ALS pipelines in a single pipeline called 'Classic'.
- Readd shadow mapping.
- Move the WS30 Effect and fragment shader out of the Compositor-specific directories.
Previously we required 6 shader attributes to pass
in all the required information for the building shader.
By using packing techniques this has been reduced to 4.
This should improve support for integrated graphics cards.
Previously the front and side faces of random/OSM buildings
had identical texture coordinates. This resulted in the sides
of buildings texture mapping being squeezed or stretched.
This change generates a separate texture mapping for the sides
of the buildings.
Previously there was very limited texture variations
as a given texture index was used for both the wall
and ceiling. Now these can be specified separately,
allowing for more variation in both STG defined and
random buildings.
Requires simgear commit 053bda26a43314a91b01b08cd4617da82f7ab807
Replace random buildings with one using proper instancing.
Also implement more control over rendering of random buildings.
see README.scenery for details.
