Difference between revisions of "Model optimization"
m (→Update history: use {{undocumented}} template) |
|||
| Line 24: | Line 24: | ||
! class="header" width="4%" | LOD efficiency | ! class="header" width="4%" | LOD efficiency | ||
! class="header unsortable" width="10%" | Notes | ! class="header unsortable" width="10%" | Notes | ||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
|- | |- | ||
| [[Sydney Sleeper]] | | [[Sydney Sleeper]] | ||
| Line 39: | Line 34: | ||
| {{Chart bar|1478|max=7000}} | | {{Chart bar|1478|max=7000}} | ||
| 76.56% | | 76.56% | ||
| + | |- | ||
| + | | [[Fists of Steel]] | ||
| + | | {{Chart bar|5880|max=7000|hex=#93aecf}} | ||
| + | | | ||
| + | | 0% | ||
|- | |- | ||
| [[Black Box]] | | [[Black Box]] | ||
| Line 45: | Line 45: | ||
| 86.13% | | 86.13% | ||
|- | |- | ||
| − | | [[ | + | | [[Degreaser]] |
| − | | {{Chart bar| | + | | {{Chart bar|5485|max=7000}} |
| − | | | + | | {{Chart bar|998|max=7000}} |
| − | | | + | | 81.8% |
|- | |- | ||
| [[Brass Beast]] | | [[Brass Beast]] | ||
| Line 64: | Line 64: | ||
| {{Chart bar|205|max=7000}} | | {{Chart bar|205|max=7000}} | ||
| 96.15% | | 96.15% | ||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
|- | |- | ||
| [[Iron Curtain]] | | [[Iron Curtain]] | ||
| Line 79: | Line 74: | ||
| | | | ||
| 0% | | 0% | ||
| + | |- | ||
| + | | [[Powerjack]] | ||
| + | | {{Chart bar|4996|max=7000}} | ||
| + | | {{Chart bar|840|max=7000}} | ||
| + | | 83.19% | ||
|- | |- | ||
| [[Crusader's Crossbow]] | | [[Crusader's Crossbow]] | ||
| Line 119: | Line 119: | ||
| {{Chart bar|904|max=7000}} | | {{Chart bar|904|max=7000}} | ||
| 77.76% | | 77.76% | ||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
|- | |- | ||
| [[Eyelander]] | | [[Eyelander]] | ||
| Line 139: | Line 134: | ||
| {{Chart bar|430|max=7000}} | | {{Chart bar|430|max=7000}} | ||
| 88.60% | | 88.60% | ||
| + | |- | ||
| + | | [[Buff Banner]] <small>(bugle)</small> | ||
| + | | {{Chart bar|3760|max=7000}} | ||
| + | | {{Chart bar|843|max=7000}} | ||
| + | | 77.58% | ||
|- | |- | ||
| [[Big Kill]] | | [[Big Kill]] | ||
Revision as of 16:03, 27 March 2011
When discussing model optimization, LODs are often brought up. This article shows the effect of LOD on polygon count in models. The charts show which models have received optimization updates in the form of LODs, and which models are still unoptimized.
Contents
Background
LOD
Main article: LOD
To reduce the number of polygons that need to be rendered in a scene, the Source Engine will render alternate lower-quality models as the distance between the model and camera increase. LOD models are not created dynamically by the engine; instead all the levels of lower-quality models must be included in the game files for LOD to be enabled.
c_models
- Chart scale
- 0-7000
- Key
- Unoptimized
- Optimized
Notes
- Medi Gun and c_models
- The Medi Gun equipped on its own uses w_models and v_models. But when the Kritzkrieg attachment model is applied, a c_model variation of the Medi Gun is used.
- Unused c_model content
- Fire Axe, Revolver, Scattergun, Syringe Gun, Wrench, Stickybomb Launcher, Bat and Bonesaw.
w_models
- Chart scale
- 0-7000
- Key
- Unoptimized
- Optimized
| Item | Highest quality (polycount) | Lowest quality LOD (polycount) | LOD efficiency | Notes |
|---|---|---|---|---|
| Minigun | 89.73% | |||
| Syringe Gun | 85.99% | |||
| Medi Gun | 84.61% | |||
| Rocket Launcher | 85.04% | |||
| Sniper Rifle | 67.62% | |||
| Revolver | 80.77% | |||
| Scattergun | 64.92% | |||
| Shotgun | 76.15% | |||
| Shovel | 67.39% | |||
| Stickybomb Launcher | 61.00% | |||
| SMG | 65.37% | |||
| Fire Axe | 76.18% | |||
| Bottle | 74.00% | |||
| Electro Sapper | 37.76% | |||
| Wrench | 55.00% | |||
| Build Tool | 61.23% | |||
| Bat | 86.41% | |||
| Pistol | 59.10% | |||
| Toolbox | 48.04% | |||
| Disguise Kit | 62.03% | |||
| Bonesaw | 51.74% | |||
| Knife | 52.40% |
Notes
- Unused w_model content
- Grenade Launcher, Destroy Tool and Scottish Resistance. These items are currently being phased out.
Character models
- Chart scale
- 0-10,000
- Note: The lengths of the bars themselves are not to scale with the weapons chart. While the numbers at the end of the bars are comparable, the sizes are not.
- Key
- Unoptimized
- Optimized
| Class | Highest quality (polycount) | Lowest quality LOD (polycount) | LOD efficiency |
|---|---|---|---|
| Demoman | 93.34% | ||
| Soldier | 90.38% | ||
| Medic | 90.66% | ||
| Scout | 79.49% | ||
| Engineer | 83.18% | ||
| Heavy | 91.94% | ||
| Sniper | 84.76% | ||
| Spy | 89.46% | ||
| Pyro | 87.04% |
Hats and Miscellaneous items
- Chart scale
- 0-5,000
- Note: the chart scale is not the same as the previous lists! You can not compare the lengths of the bars in this table with bars in the other tables. You CAN compare the numbers though.
- Key
- Unoptimized
- Optimized
Update history
- Graphics optimizations for mid and low end hardware
- Added more particle optimizations for mid range and low end hardware
- Optimized some particle systems to improve performance on some hardware configurations
- Made some small changes to the Demoman viewmodel that were causing performance problems
- [Undocumented] Added LODs to the Huntsman, Direct Hit, Ambassador, Iron Curtain, Frontier Justice, German Gonzilla, Ghastlier Gibus, Tyrant's Helm, and Cheater's Lament
- [Undocumented] Added LODs to the Buff Banner (flag), Eyelander, Flamethrower, Flare Gun, Brass Beast, Equalizer, and Scottish Resistance
- [Undocumented] v_ and w_ model versions of the Flamethrower was removed from the game files and replaced with a newer c_model variant
- [Undocumented] Added LODs to the Black Box, Buff Banner (bugle), Degreaser, Blutsauger, Powerjack, Sandvich, Übersaw and Jarate
