Graphics settings and what they do

Discussion in 'Technical Support' started by Pandalishus, Aug 23, 2013.

  1. Pandalishus

    Pandalishus Active Member

    Likes Received:
    245
    Log in to remove this ad
    One of the things I've been unclear on has to do with all of the various advanced graphics settings. Some of them I recognize, others are complete mysteries to me. As both a point of personal education and, more importantly, as a service to the community, I wonder if we could turn this thread into a kind of "this is what setting x does, how it impacts FPS, and whether it's particularly important." Hopefully, it'll turn into a good reference for those trying to squeeze a few more frames out of older cards, or wanting to keep newer cards running smooth and cool.

    *If you'll give a description, I'll edit this post with those descriptions. In the meantime, I've offered my (hopefully correct) observations based on mucking around with the benchmark. I'm beginning the process of testing these one-by-one now. My base score was 7264 on "Maximum" @ 1920x1080. As I go along, I will convert the new score to a percentage so others can get a feel for how noticeable the impact was. This does not mean I got xx% more FPS; just that my benchmark score went up/down xx%. This is only for my card, of course, but should give others a sense of where they'll get performance gains.

    General
    • Enable HDR rendering and improve overall graphic quality
      • @Krozjin mentioned this is likely the same thing as HDR in photography and film (i.e. High Dynamic Range) which makes shadows, highlights, reflections more realistic.
      • Disabled: +15%
    • Disable rendering of objects when not visible. (Occlusion Culling)
      • Affects the rendering of items off-screen that would affect lighting on-screen.
      • Enabled: +2%
    • Use low-detail models on distant objects to increase performance. (LOD)
      • Self-explanatory
      • Enabled" +3%
    • Cache LOD data only when necessary. (LOD Streaming)
      • ?
      • Enabled: 0%
    • Real-time Reflections
      • Whether shiny reflect moving objects around them.
      • Normal: 0%, Off: +1%
    • Edge Smoothing (Anti-aliasing)
      • Reduces the "jaggies" on edges (e.g. straight lines viewed at angles have fewer "stair steps" along the edge).
      • Disabled: +4%
    • Transparent Lighting Quality
      • @Krozjin mentions this is the effect of light shining through transparent objects (e.g. looking through a waterfall).
      • Normal: +8%
    • Grass Quality
      • Affects the amount/animation for grass?
      • Norm: 0%, Low: 0%, Off: +2%
    Shadows
    • Self
      • Self-explanatory.
      • Disabled: 0%
    • Other NPCs
      • Self-explanatory.
      • Disabled: +3%
    Shadow Quality
    • Use low-detail models on shadows(LOD)
      • Determines whether shadows are blobs or are shaped like the object casting them.
      • Disabling can increase performance.
    • Shadow Resolution
      • How any pixels comprise the shadow (whether the shadow is blocky or not).
      • 1024p: +1%, 512p: +2%
    • Shadow Cascading
      • @Krozjin suggests this the distance at which shadows are drawn at higher resolutions. Shadows farther away look better?
      • Normal: 2%, Off: 3%
    • Shadow Softening
      • How "contrasty" shadows are.
      • Weak: +1%
    Texture Detail
    • Texture Filtering - [Anisotropic/Bilinear/Trilinear]
      • Increases sharpness of all textures.
      • Bilinear: +1%
    • Anisotropic Filtering - [x16/x8/x4]
      • How much filtering occurs.
      • Generally speaking, higher values to do not significantly impact performance, with the difference between 4x and 16x being less than 1 fps in many applications (according to the interwebz).
    Movement Physics
    • Self
      • Affects the fluidity of movements of player models.
      • Lower values can increase performance in certain situations.
    • Other NPCs
      • Affects the fluidity of movements of NPC models (in live client, this is split further to friendly NPCs, enemies, other players). Lower values can noticeably increase performance in certain situations.
    Effects
    • Naturally darken the edges of the screen. (Limb Darkening)
      • Self-explanatory (slight darkening of screen edges), and a largely personal preference.
      • Disabled: +1%
    • Blur the graphics around an object in motion. (Radial Blur)
      • Self-explanatory.
      • Largely personal preference.
    • Screen Space Ambient Occlusion
      • Affects how objects reflect light, resulting in more detailed textures and "depth" at higher settings.
      • Can significantly improve the quality on items rendered on screen
      • Weak: +12%, Off: +16%
    • Glare
      • Effects such as lens flare when looking at the sun, etc.
      • Effect on performance is situational, with the difference between low and normal being noticeable.
      • Low: 0%, Off: 3%

    Cinematic Cutscenes
    • Enable depth of field
      • Blurs objects further away during cutscenes.
      • Disabling can have a significant impact on fps, but will likely not impact normal gameplay since it's (presumably) limited only to cutscenes.

    A few observations:
    • It is critical to understand that these figures represent averages. For instance, the benchmark has several dungeon scenes, where shadows push the GPU less. Where I got a negligible boost by disabling shadows, you might get a tremendous boost if, say, you were running around outside. The benchmark does not give a score based on one type of scenes, but on the average performance across a wide variety of scenes.
    • My card has some of the newest tech, so things that don't make much difference to it might kill older cards (e.g. shadows tend to be a HUGE strain on older cards).
    • These figures are relative. Whereas a +10 fps jump for me represents may only represent 15%, for older cards it could be 50% or even 100%. What the above hopes to accomplish is point out the big offenders, so you can disable large impact item first.
    • I am uncertain whether changing multiple settings is additive, multiplicative, etc. I never changed more than one at a time, and always reset back to "Maximum" after a benchmark.
    • None of the figures above takes into account the visual impact. While disabling SSAO is a big boost, it also makes the game look considerably worse. It may be worth changing a few small things to maintain higher settings on the big things.
    In the end, I hope the above is helpful, but it's no substitute for your own experimentation. SqEx has done a good job of giving a variety of presets, so try them out as well, referring to this post just to see what extra changes you can make. I will also be happy to add other scores if people want to run the game on Max, then disable this or that and give me the two scores.
    Last edited: Aug 24, 2013
  2. Fascinate

    Fascinate Member

    Likes Received:
    29
    Hey its cool that you are making this. I was lazy in beta and just did standard desktop, then enabled anti aliasing and max shadows. To me game looked nearly as good as max and i got 60 fps everywhere.

    Ill keep and eye on this thread to fine tune my settings to get max performance/looks : )
    xusha and Pandalishus like this.
  3. Krozjin

    Krozjin Active Member

    Likes Received:
    87
    I would test all of these for you but we can't. I'm going to do my best to describe it, since I have worked in the film industry and have made game environments/levels.

    HDR - Usually, for film at least, HDR lighting refers to using High Dynamic Range images to simulate real lighting scenarios and causes a huge jump in render time. For film, it is used for shadows and reflections. For example, if you have a red ball next to a white wall you will see some of that red ball on the wall. Just a bit of red at least depending on the wall texture. That is what HDR does, and is probably pretty intensive in game too. Games simulate all of that in real time, so disabling it makes everything look bland, but it really is a big performance boost I believe.

    Transparent Lighting Quality - When light shines onto a solid object it creates a shadow, but when it shines through something that has transparency, such as leaves or grass, then the quality usually needs to be higher for better results. So if you lower the settings then the shadows from things like leaves, grass, etc will be lower. No clue as to the performance boost.

    Shadow Cascading - The way I understand it, and the simplest way of explaining it I think is, the further away the shadow is the less detailed and accurate it is, it doesn't try to use the high density detailed shadow on far away objects since you can't see them anyways. This means the computer basically renders multiple shadows for better looking results as far as I understand. No clue as to the performance boost.

    Hope that helps at least a little with the ones you were uncertain about.
  4. Pandalishus

    Pandalishus Active Member

    Likes Received:
    245
    Thanks Krozjin. I've added your notes to the above.

    Tonight I think I'm going to start running the benchmark with carious settings changed and start reporting back with scores.
  5. carlos8787

    carlos8787 Member

    Likes Received:
    14
    Thanks for the info man. I was wondering what some of these graphic settings do. At least now i know
  6. Lokiie1984

    Lokiie1984 New Member

    Likes Received:
    7
    On my lower end computer, turning off shadows and grass made the biggest difference in my score. I will have to wait until i can actually play the game to really tinker with it but I'm happy with the score i have managed to pull out. The game looks pretty decent on lower settings. Generally as long as i can keep the character textures up, i don't care too much about the rest.

    For the record i got a 2500 score on a pretty dated computer. It's no 9000 like the OP got but it should be more then playable.

    windows xp
    amd athlon x2 dual core 5000+
    ati radeon hd 4600 (1gig vram) (I probably need to look for driver updates too)

    Edit: there is also about a 100 point difference between windowed mode and full screen. With windowed mode being the lower.
    Last edited: Aug 23, 2013
  7. Pandalishus

    Pandalishus Active Member

    Likes Received:
    245
    Yeah. Fullscreen will always give higher FPS than windowed.
    --- Double Post Merged, Aug 24, 2013, Original Post Date: Aug 24, 2013 ---
    OK, that's all for tonight. Time to get to bed so i can wake up bright and early and get to leveling my SCH.
  8. carlos8787

    carlos8787 Member

    Likes Received:
    14
    The biggest thing that can make a drag on youre video card is Shadows. If your having a hard time with frames and its bugging you. Turn off shadows. it helps tremendously
  9. Pandalishus

    Pandalishus Active Member

    Likes Received:
    245
    It depends on the card, based on the above. Disabling shadows had no real impact on my score. Only about a 3-4% increase in score. On the other hand, if it was my old Radeon 4850, it would ave been a massive change.
  10. Caitline

    Caitline Newbie

    Likes Received:
    10
    This is a VERY good explanation of the graphics. I, myself, seem to notice minimal difference between High and Max settings. I must say though, Square Enix knows how to make a very beautiful MMO world.
    Pandalishus likes this.
  11. carlos8787

    carlos8787 Member

    Likes Received:
    14
    If you have an amazing video card than it wont see much improvement. But chances are lowering shadows helps tremendously on many video cards. Not all, but its defiantly worth a try
  12. DarkEmpyrean

    DarkEmpyrean Member

    Likes Received:
    95
    Also at when we finally get in-game, might not be a bad idea to turn a few of the things you normally keep on higher settings a step or two lower than usual.

    Tons of people will be everywhere, especially in the starting areas, and even if you have an awesome video card/computer/whatever you might lag out a bit when youre actively rendering 200+ characters in your screen all yelling and emoting and fighting things. Once you get away from the main areas just switch it back up to usual settings. This of course isnt true for everyone, but wont hurt to try if youre having problems.
    xusha likes this.
  13. carlos8787

    carlos8787 Member

    Likes Received:
    14
    Yes that is very true. I'm not too worried because im running at 47fps, so even if theres alot i think im safe. I really dont get pissy untill i get around 25fps. But to those who are on that threshold its important to note what DarkEmpyrean said
  14. DarkEmpyrean

    DarkEmpyrean Member

    Likes Received:
    95
    Also something else I thought about is the weather. If it starts raining this could mess with you a bit too if youre lagging out, especially in an area with lots of people. Unfortunately if youre in an area you gotta do stuff in theres not much you can do about it, but it certainly can affect framerate. Ducking inside to a building/cave will stop the game rendering of the rain around you as long as your cam isnt pointed at it and allow you to do stuff if you find rain to be a problem.

    I recently got a new video card (gtx 650ti), and having played phase 3 on my old card (9500gt) I certainly noticed performance drops during rain. Havent experienced too much other weather, but by the time you get to the more exotic stuff you wont be in towns full of new players lol.
  15. Lokiie1984

    Lokiie1984 New Member

    Likes Received:
    7
    Anyone know if the benchmarks visuals are a good representation of the graphic settings? I have noticed that other then shadows, and grass i cant really see any difference between max settings and low. Even changing my resolution doesn't seem to change what i see. My score changes dramatically of course but the video it plays doesn't seem to.

    Lowering my resolution to 1028x786 in full screen bumped my score up from 2500 to 4500. I'm curious how that would look in game though.
  16. DarkEmpyrean

    DarkEmpyrean Member

    Likes Received:
    95
    The benchmark visuals ARE the graphic representation...thats why the benchmark thing is like 2gigs, all the little scenes that play out are within the actual preset game environment.

    The majority of what will affect you are things to do with lighting...like you said shadows are a good part of the render cost, but what may not be apparently changing in the benchmark is the lighting from the sky. In the game unless you cam around a lot from different angles its not something you really notice right off, but HDR settings are apparent in things like shafts of light through the trees, fog/atmospheric distortion and the like. Otherwise pretty much what you see in benchmark is what youll see in game.
  17. Lokiie1984

    Lokiie1984 New Member

    Likes Received:
    7
    Hmm, my eyes cant tell the difference so i think i will be fine with the lower settings. Makes me wonder if i need new glasses lol. Definitely like the graphic options for ARR vs 1.0 though, my computer could hardly run that version. One of these days i need to invest in a new card, but for now i think this should work well.
  18. carlos8787

    carlos8787 Member

    Likes Received:
    14
    Im happily running 47fps on max resolution yay!
    Pandalishus likes this.
  19. Vuxxy

    Vuxxy New Member

    Likes Received:
    8
    I made a video showing the optimal settings in FFXIV: ARR.

  20. Elva

    Elva Newbie

    Likes Received:
    0
    Thanks for this!

    But you forgot to mention that this game heavily relies on CPU. i have i7-3770k @ 4.5

    Using MSI afterburner, i can determine that this uses more than 4 threads. about 6 i think. so this game is well threaded.
    Having to calculate all the scripts during a busy battle is CPU intensive.

    People may benefit more from getting a decent CPU and just maybe a mid end GPU.
    --- Double Post Merged, Nov 11, 2013 ---
    Shadow cascading is how many times a shadow source will fall upon an object.
    Say you have a stack of 10 logs. the higher the shadow cascading is, the more logs the shadow would fall on to mimic how it might be in real world. but if it is set to minimum, it might only fall on 2 logs and not reach further even though the angle of the shadow source should be casting a shadow on more logs.

    this can be noticed a lot on many of tall mountain sides, tall rocks and generally things that have a tall side surface.
    Last edited: Nov 11, 2013

Share This Page