No, that picture is completely wrong. I've got no idea what they did to it to make it so blurry:
Crytek's shot:
Actual in-game shot:
Last edited by a moderator:
Digital Foundry Article Technical Discussion Archive [2011]
- Thread starter user542745831
- Start date
-
- Tags
- digital foundry
- Status
Crytek's shot:
Actual in-game shot:
Yea the PS3 DF shots are blurrier than the actual game, makes me think just how sharp does the Xbox 360 version looks.
djskribbles
Legend
If you look at the rest of the DF comparison screens, there's hardly a difference in sharpness on most of the shots. It's weird how that shot in particular looks so soft. Does the PS3 version have more aggressive LOD?
The shot I posted was from Crytek's PostMSAA example in their paper, I think the blurriness might just be from scaling, though.
Last edited by a moderator:
I think depends more to a 'wrong' cap in motion during a cutscene, caused to a mix of motion blur, custom AA, more evident on the ps3 for the more subhd res, because in the same situation in the video comparison the ps3 version it isn't so much blurried but pretty similar to 360 when the image is 'completely' static.
There is pretty big difference in static shots imo.
I wonder how much it costs the ps3 to render 16xaf in crysis 2 in comparison to simply render at a higher resolution such as 1152 x 720? If the cost is roughly the same then which option would be the better for the overall iq?
The difference AF that makes is actually bigger than I thought as you can see the stark contrast here.
check the ceiling and wall tiles.
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/PS3_023.jpg.jpg
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/360_023.jpg.jpg
This is the biggest difference so far, that huge concrete on the right.
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/PS3_041.jpg.jpg
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/360_041.jpg.jpg
The difference AF that makes is actually bigger than I thought as you can see the stark contrast here.
check the ceiling and wall tiles.
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/PS3_023.jpg.jpg
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/360_023.jpg.jpg
This is the biggest difference so far, that huge concrete on the right.
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/PS3_041.jpg.jpg
http://images.eurogamer.net/assets/articles//a/1/3/4/2/6/0/5/360_041.jpg.jpg
Last edited by a moderator:
The PS3 16x AF is "dynamic", only means it can be up to 16x.
Even 16x AF can not save the loss of 12% pixels in IQ criteria,it would be RSX can do 16x AF from hardware design fundemental
,not Crytek take 16x things to trade for not using higher resolution.
I think the reason why 360 can do 4xAF is hardware design limitation with EDRam or something.
Even 16x AF can not save the loss of 12% pixels in IQ criteria,it would be RSX can do 16x AF from hardware design fundemental
,not Crytek take 16x things to trade for not using higher resolution.
I think the reason why 360 can do 4xAF is hardware design limitation with EDRam or something.
Most likely something as texture units read from RAM not EDRAM.
Running the game at a higher resolution would no doubt cost more than dynamic AF, if it can be turned down when the framerate takes a dive.
I think I'd go for the cleaner look and the slightly higher resolution. The games look near identical to me.
Rendering at a different resolution isn't that simple considering all the render targets/buffers involved - more memory, more pixel shading... more everything...
DF Crysis 2 Tech Comparison:
http://www.eurogamer.net/articles/digitalfoundry-tech-comparison-crysis2
Fairly puzzled that they didn't say anything specific about the AA in the PC version (FXAA or whatever it is) other than:
Easily close to the level of proper multi-sampling.
http://www.eurogamer.net/articles/digitalfoundry-tech-comparison-crysis2
Fairly puzzled that they didn't say anything specific about the AA in the PC version (FXAA or whatever it is) other than:
Just seems like a bit of a misguided statement since the FXAA seems to work incredibly well and cleans up the entire image, albeit with the loss of some sub-pixel detail, eg, what it looks like on Extreme:
Easily close to the level of proper multi-sampling.
Last edited by a moderator:
Where do you select FXAA? If it's not in the default option screen, then it's a no go. It wouldn't be his fault if Crytek hid away all the interesting variables. If you want these articles out sooner, there's just no bloody time to mess around tweaking. Let's see you try playing through the game three times over in a single week whilst having capture duties.
Besides, at this point, everyone's really just waiting for dx11.
Besides, at this point, everyone's really just waiting for dx11.
Last edited by a moderator:
FXAA is on by default on Extreme (AFAIK), but it's more about him complaining about the AA on max settings (even though it seems to work quite well, remarkably even), although I really agree that Crytek should have been more specific about it, which is probably the root cause of the issue.
EDIT: It makes me wonder why they've locked down everything. Encrypted .pak files, only 47 console commands, no advanced options, etc. I can understand that they're appealing to a broader community, but cutting off features that allow for tweaking is not the way to go about it at all.
Last edited by a moderator:
It's enabled by default when choosing "extreme" graphics setting according to document. There certainly is a difference in jaggy reduction between next highest and highest setting.
Might be this command: r_PostMSAAEdgeFilterNV
I'll bet it's vaporware.
Last edited by a moderator:
If interested here is retail games config file info. Dont know how correct they might be but this is what is present in retail PC game archives.
Global base console config (console.cfg).
Changes from beta leak config (beta leak value in brackets).
360 override (x360.cfg).
Changes from beta leak config (beta leak value in brackets).
PS3 override (PS3.cfg).
Changes from beta leak config (beta leak value in brackets).
base console multiplayer config (multiplayer_console.cfg).
Global base console config (console.cfg).
Changes from beta leak config (beta leak value in brackets).
note present (e_ShadowsOnAlphaBlend=0)
e_ShadowsConstBias = 1 (1,14)
e_ShadowsSlopeBias = 1 (3,4)
e_ParticlesQuality = 2 (1)
e_GIAmount = 5 (this cvar was missing in beta)
e_ShadowsConstBias = 1 (1,14)
e_ShadowsSlopeBias = 1 (3,4)
e_ParticlesQuality = 2 (1)
e_GIAmount = 5 (this cvar was missing in beta)
Code:
--sys_spec_ObjectDetail=2
e_LodRatio=4
e_VegetationMinSize=0.5
e_DecalsLifeTimeScale=1
e_MaxViewDistSpecLerp=0.5
es_DebrisLifetimeScale=0.6
e_CoverageBufferResolution=128
ca_AttachmentCullingRation=200
e_TerrainOcclusionCullingMaxDist=130
--sys_spec_Shading=2
q_ShaderGeneral=1
q_ShaderMetal=1
q_ShaderGlass=1
q_ShaderVegetation=1
q_ShaderIce=1
q_ShaderTerrain=1
q_ShaderShadow=1
q_ShaderFX=1
q_ShaderPostProcess=1
q_ShaderHDR=1
q_ShaderSky=1
q_Renderer=1
q_ShaderWater=1
e_TerrainNormalMap=0
r_EnvTexUpdateInterval=0.075
r_TexturesFilteringQuality=2
--sys_spec_VolumetricEffects=2
r_BeamsDistFactor=0.5
r_BeamsMaxSlices=32
--sys_spec_Shadows=2
r_ShadowsUseClipVolume=1
r_ShadowJittering=1
e_ShadowsCastViewDistRatio=0.45
e_GsmCache = 1
e_ShadowsConstBias = 1
e_ShadowsSlopeBias = 1
e_ShadowsMaxTexRes = "512"
--sys_spec_Texture=2
r_ImposterRatio=1.5
r_DynTexMaxSize = 20
r_DynTexAtlasCloudsMaxSize = 8
r_DynTexAtlasSpritesMaxSize = 8
--sys_spec_Physics=2
e_CullVegActivation=30
es_MaxPhysDist=100
es_MaxPhysDistInvisible=15
e_PhysOceanCell=1
e_FoliageWindActivationDist=10
p_gravity_z = -13
e_PhysOceanCell=0
e_OnDemandPhysics = 0
--sys_spec_PostProcessing=2
--sys_spec_Particles=1
e_ParticlesPoolSize = 768
e_ParticlesEmitterPoolSize = 1536
e_ParticlesQuality = 2
e_particlesMaxScreenFill = 16
e_particlesMinDrawPixels = 2
--sys_spec_GameEffects=2
g_corpseMinTime = 8.0
g_corpseMinDistance = 20.0
g_battleDust_enable = 0
p_max_substeps=2
-------------------
sys_flash_curve_tess_error = 4
r_HDRRendering = 1
r_HDRBlueShift=1.0
r_HDRBrightLevel=0.85
r_HDRBrightThreshold=6
r_HDRLevel=6
r_HDROffset=10
r_EyeAdaptationBase=0.1
r_EyeAdaptationFactor=0.85
r_EnvTexResolution = 1
r_TexSkyResolution = 1
r_DetailNumLayers = 2
r_DetailDistance = 4
r_TexBumpResolution=0
r_TexResolution=0
r_TexturesStreaming = 1
r_texpreallocateAtlases=0
r_TexturesStreamPoolSize = 90
r_ShadersUserFolder = 1
r_GeomInstancing = "0"
r_refraction = 1
r_ShadowBlur = "0"
r_SSAOQuality=0
r_SSAODownscale=1
e_CoverageBuffer = 1
e_Coveragebufferrotationsafecheck=2
r_WaterReflections=1
r_WaterUpdateFactor=0.25
e_WaterTesselationAmount=1
e_WaterTesselationSwathWidth=10
r_WaterUpdateDistance=1.0
r_WaterReflectionsQuality=4
e_WaterOcean_fft=1
r_WaterReflectionsMinVisiblePixelsUpdate = 0.05
e_HwOcclusionCullingWater = "1"
r_WaterRefractions = 1
r_WaterCausticsDistance = 16
r_WaterUpdateFactor = 0.15
r_WaterReflections_Quality = 4
r_Width = 1280
r_Height = 720
e_LevelAutoPrecacheTexturesAndShaders = "1"
e_VegetationSpritesMinDistance = 32
-- 0.3 is fastest value in game04 cxp_benchmark
e_VegetationSpritesDistanceCustomRatioMin=0.3
e_StatObjMerge = "1"
e_GsmRange = 3
e_GsmRangeStep = 2.25
e_GsmLodsNum = 5 // shadows fix for elevator scene
e_DynamicLightsMaxEntityLights = 5
e_ParticlesPreload = 0
e_VegetationBending = 1
e_Clouds=1
e_StreamCgf=1
e_streamcgfpoolsize=24
e_LevelAutoPrecacheCgf = 1
e_TerrainAo=0
e_Dissolve=2
r_UsePOM = 0
e_SkyUpdateRate=0.2
e_ProcVegetationMaxSectorsInCache = 8
e_ProcVegetationMaxObjectsInChunk = 256
e_ProcVegetation=0//not needed for city
e_LodMinTris=150
;for performance reasons disabled, saved 1% frametime, decided with MartinM to disable it
e_VegetationUseTerrainColor = 0
e_ViewDistRatio=25
e_ViewDistRatioDetail=19
e_ViewDistRatioCustom=50
e_ViewDistMin=10
e_ViewDistRatioVegetation=12
e_ObjQuality=2
e_LodMin=0
e_VegetationSpritesDistanceRatio=1
e_ShadowsLodBiasFixed=1
e_GI = 1
e_GIAmount = 5
r_PostProcessEffects="1"
r_UseEdgeAA = 1
r_MotionBlur = 2
r_SunShafts = 2
r_Glow = 1
r_beams = 0
r_coronas = 1 // unsuported on consoles atm
r_useSRGB = 1
--r_DepthOfField = 3
r_ColorGrading = 2
r_ColorGradingCharts = 1
r_rain=2
r_RainAmount = 1.2
r_RainMaxViewDist_Deferred=40
r_CustomVisions = 1
ca_MemoryUsageLog=0
ca_physicsProcessImpact=0
sv_gamerules = "SinglePlayer"
hud_startpaused="0"
sys_StreamCallbackTimeBudget=20000
r_ShadersAsyncCompiling = 1
--r_ShadersRemoteCompiler=1
r_ShadersPreactivate = 1
e_DynamicLightsFrameIdVisTest=2
e_DecalsMerge=0 // 0 = less memory allocations and stalls, but more drawcals
e_DecalsHitCache=1
e_DecalsNeighborMaxLifeTime=1
e_DecalsOverlapping=0
e_TerrainLodRatio = 2
e_TerrainTextureStreamingPoolItemsNum=16
e_LevelAutoPrecacheCameraJumpDist=0 // precache disabled for less loading time
e_ShadowsCastViewDistRatio=1 // less shadow popping in crysis and game04 areas
e_GsmCacheLodOffset=2 // shadow maps update frequentcy
r_DeferredShadingLightVolumes=1
r_DrawNearZRange = 0.12
r_DrawNearFarPlane = 80
p_max_mc_iters = 2000
r_DisplayInfo = 0
g_breakage_particles_limit = 40
p_joint_damage_accum=2
p_joint_damage_accum_threshold=0.2
p_max_entity_cells=10000
log_AllowDirectLoggingFromAnyThread = 1360 override (x360.cfg).
Changes from beta leak config (beta leak value in brackets).
Missing in final (s_VariationLimiter = 0.3)
New values:
e_coveragebuffer =0
e_statobjbufferrendertasks=1
e_coveragebufferreproj=2
e_coveragebufferbias=0.01
e_coveragebufferAABBExpand=0.007
New values:
e_coveragebuffer =0
e_statobjbufferrendertasks=1
e_coveragebufferreproj=2
e_coveragebufferbias=0.01
e_coveragebufferAABBExpand=0.007
Code:
r_multithreaded=2
e_LodRatio=5
e_CoverageBufferResolution=128
e_lodmin = 0
e_streamcgfpoolsize = 24
r_TexturesStreamPoolSize = 90
s_Obstruction = 1
s_ObstructionAccuracy = 1
s_ObstructionUpdate = 0.5
s_SoundMoodsDSP=1
s_ReverbType=2
p_num_threads=2
d3d9_pip_buff_size = 24
g_breakage_mem_limit = 5000
r_DeferredShadingSortLights=0
r_VSync = 2
sys_maxfps = 30
d3d9_rb_verts=49152
d3d9_rb_Tris=131072
e_particlesThread=4
sys_streaming_cpu=1
sys_physics_CPU=5
sys_dump_format = 2
e_CharLodMin = 1
e_coveragebuffer =0
e_statobjbufferrendertasks=1
e_coveragebufferreproj=2
e_coveragebufferbias=0.01
e_coveragebufferAABBExpand=0.007
e_gi=0PS3 override (PS3.cfg).
Changes from beta leak config (beta leak value in brackets).
Not present in final (s_VariationLimiter = 0.3)
p_async_rwi_use_spu = 0 (1)
r_MeshPoolSize = 1300 (1024)
New values:
e_coveragebuffer =0
e_statobjbufferrendertasks=1
e_coveragebufferreproj=2
e_coveragebufferbias=0.01
e_coveragebufferAABBExpand=0.007
p_async_rwi_use_spu = 0 (1)
r_MeshPoolSize = 1300 (1024)
New values:
e_coveragebuffer =0
e_statobjbufferrendertasks=1
e_coveragebufferreproj=2
e_coveragebufferbias=0.01
e_coveragebufferAABBExpand=0.007
Code:
sys_streaming_memory_budget = 2
//disabled due to devtrackID 7445
e_VegetationSprites = 0
e_VegetationSpritesDistanceRatio=0.95
r_VegetationSpritesTexRes = 128
e_LodRatio=5
r_PS3ConstBufferAdjustment=128
e_PhysMinCellSize = 4
r_multithreaded = 1
r_VSync = 1
sys_maxfps = 30
i_forcefeedback = 1
i_lowFreqMotorScale = 0.4
i_digitalMotorActivationScale = 0.4
i_motorDeadZone = 0.25
e_GsmCache = 1
e_GsmRange = 3
e_CoverageBufferResolution = 128
e_CoverageBufferVersion = 3
r_ConditionalRendering = 2
sys_physics_CPU = 1
spu_enable = 1
r_UseGSParticles = 0
e_OcclusionCullingViewDistRatio = 1.0
r_deferredshadingstencilprepass = 1
r_ShadowsDepthBoundNV = 1
r_UseShadowsPool = 1
r_DeferredShadingDepthBoundsTest = 1
e_WaterTesselationAmount = 6
spu_flipmode = 1
r_PS3HalfResRendering = 1
s_Obstruction = 1
s_ObstructionAccuracy = 1
s_ObstructionUpdate = 0.5
s_SoundMoodsDSP=1
s_ReverbType=2
e_ProfileLevelLoading=0
p_async_rwi_use_spu = 0
p_max_spu_rwi_cells = 5
g_no_secondary_breaking = 1
g_tree_cut_reuse_dist = 1
g_breakage_mem_limit = 2000
es_maxphysdistcloth=300
r_Refraction = 2 // enable reads from destination render targets - avoid copies
r_texturesstreampoolsize=60
r_MeshPoolSize = 1300
r_MeshVolatilePoolSize = 512
e_DecalsMerge=0
e_coveragebufferaccurateobbtest = 0
e_shadowsocclusionculling = 0
e_coveragebufferbias=3
e_AutoPrecacheCgfMaxTasks=2
e_DecalsMaxUpdatesPerFrame=2 // limit number of static decal mesh updates per frame
ps3_virt_mem_sys_size_kb=0
ps3_virt_mem_rsx_size_kb=0
r_AnisoBlendWeight=1
r_AnisoMax=8
r_AnisoMin=0
r_AnisoPasses=1
r_AnisoRefDistance=4
r_AnisoScale=8
//officially approved, saves 14 mb and 4 ms rsx
r_ConsoleBackbufferWidth = 1024
e_CharLodMin = 1
e_coveragebuffer =0
e_statobjbufferrendertasks=1
e_coveragebufferreproj=2
e_coveragebufferbias=0.01
e_coveragebufferAABBExpand=0.007
e_gi=0base console multiplayer config (multiplayer_console.cfg).
Code:
e_GI=0
r_DeferredShadingIndexedAmbient=0
e_ParticlesObjectCollisions=0
g_distanceForceNoLegRaycasts=0.00001
g_telemetryDisplaySessionId=1
g_breakageNoDebrisCollisions=1
; Breakage throttling
; These are explained in ActionGame.cpp
;
g_glassAutoShatterOnExplosions=1
g_glassNoDecals=1
g_glassMaxPanesToBreakPerFrame=2
g_breakageTreeMax=100
g_breakageTreeInc=101
g_breakageTreeDec=25
g_breakageTreeIncGlass=51
Last edited by a moderator:
I definitely think it's weird for a 2011 AAA PC game to not offer any sort of MSAA whatsoever (not that they would be able to implement it in D3D9, anyway)
- Status
Similar threads
- Locked
- Deleted member 11852
- Gaming Technology
- Locked