Tested: Graphics settings and how they affect performance

Game settings can seem complex at times. Today we dig in to some of those settings and take a look at how they affect performance and graphics quality!

○○○○○○ MSI GS65 Thin available at Amazon ○○○○○○

WATCH MY MOST POPULAR VIDEOS!: (link)

••• Follow me on your favorite Social Media! •••

About JayzTwoCents

Your one stop place to shop for fun and educational videos on all things tech! Oh, and don't forget gaming! Gotta make time for gaming!

Tested: Graphics settings and how they affect performance | JayzTwoCents

JAYZTWOCENTS

[jrmaty]

I thought I'd give a bit more explanation to the settings, for those interested on why these have the effect they do (source: I'm a technical artist & graphics programmer)

Anti-Aliasing:
Most modern AAA titles use deferred rendering of some sort.
This means that, yes, anti-aliasing is either a post effect (applied across the screen after it has rendered, typically using edge detection techniques), or its a super-sample effect - which hits performance hard. Post effects are driven by screen resolution, so upping your game's resolution (or conversely, dropping it) will also affect how long it takes for the anti-aliasing computations to occur. As mentioned with shadows further down, doubling resolution isn't a linear hit on computations, but a square one (so 1280x720 is less than 1 million pixels, 1920x1080 = 2 million+ pixels.

Texture Quality:
Game textures are compressed in a way that they generate & store lower resolution versions (known as MIP maps). As an object is further is away, the game will sample a lower resolution texture version to save on graphics memory. Setting the texture quality lower simply forces the game to use one of these lower resolution textures (MIP levels) as the maximum resolution - so you end up with lower resolution textures closer to the camera. This saves on video memory usage.

Anisotropic filtering:
A cheap technique to reducing banding, always max out.

Shadows!
There are a lot to shadows. You typically have two types: baked and realtime
Baked shadows cost almost nothing, but they don't work with dynamic lights (lights that move, change color, etc.). Some games use this more than others.
Realtime shadows are expensive. Its a trade-off between shadow map resolution (i.e. how detailed the shadows are) and performance. Shadows done like this usually have 4 passes (or "cascades") - this works similar to MIP mapping in that there are 4 shadow maps that are projected on to objects, with objects further away getting less density from the shadow map (the resolution is the same for all "cascades", but cascades further away have more surface area to cover, perspective and all that).
Usually realtime shadows can get you most of the way there on medium because the resolution is "good enough", and because each quality setting (i.e. medium, high, ultra) is twice the resolution of the previous. To explain why that hurts performance so much: this is because resolution is 2-dimensional, twice the resolution is 4x the pixels.
That means, if 'Medium' was 1024^2, and 'High' was 2048^^2 and 'Ultra' was 4096^2, ultra requires 16 times more pixels to calculate. Realtime shadows need calculating each frame, so in that example, that's 16x more pixels on each cascade, and it does this EVERY FRAME. Hence, it hits performance hard.

Ambient Occlusion
This is usually a post effect, and not much more expensive than most other post effects. It can add nice depth to the scene. All post effects are done after geometry and shading is complete, so the bottleneck here is resolution (and is what the "Quality" setting controls). You can apply the ambient occlusion at a lower resolution than your game renders at, and get the effect with better performance than running at full resolution, but you'll get a blurrier result.

Tessellation:
This works by subdividing the geometry that the game artist's authored at runtime, and displacing it - unlike older approaches that don't change the geometry but "simulate" it. How much is tessellates can be drive by proximity to the camera, and a local quality setting. This can add a lot of detail, but its costly on older hardware & consoles.

[jujuProductions]

imagine having a pc that lets you "drop" all the way down to *60 FPS*

[Grif]

I’m so glad I watched this. Now I can squeeze more frames out of Minesweeper.

[TheLukemcdaniel]

This makes me want to mod the heaven benchmark so that tesselation is constantly adjusted in relation to currently playing music....

[Yama_]

“Oh, it looks like I’m motorboating her.. I didn’t do that on purpose, I swear.” Gotta love that man!

[SidTheGeek]

My graphics settings are always set to Low. there should be a graphics settings preset like "No Graphics Card" lol ✔️

[DawidDoesTechStuff]

Next do a video on how much an RGB power cable effects performance. I bet its a load.

[passionfly]

Probably one of the most and useful videos ever to be made by JayzTwoCents and it is too short! MORE of this content please!

[Lamentxthe]

"We just lost about 15 fps" I would have 0 fps after that.

[riteu]

The "bass test" was by far my favorite part of the video.
Freaking love Jay!

[dragonknight]

Great examples and explanations Jay! This is the "voodoo" most gamer's have zero clue about when setting up for each game they play. This video is on my short list when selling my rigs, it saves me a ton of talking/teaching after the sale, thank you 👍🍻

[villads]

2 years later and still learning from these vids. Great job

[johnm]

Actually, for the smoothest experience, you turn on V-sync, in game, with G-sync enabled in the Nvidia control panel. G-sync controls the monitors refresh rate when your graphics card is unable to meet the maximum refresh rate of the monitor, providing a smooth experience. Simply put, with G-sync enabled, if your graphics card is pushing 52fps, then your monitor refreshes at 52fps preventing tearing and stuttering. V-sync sets the absolute fps cap the card is allowed to push. So, if you can push 200fps on a 144hz panel, V-sync will stop you at 144fps preventing tearing. Your monitor cannot display faster than its maximum refresh rate, and neither G-sync or Freesync can change that. Freesync and G-sync are only designed to handle the fps when you cannot meet the monitors maximum refresh rate. Both technologies auto disable when you hit the monitors maximum refresh rate and re-enable when you drop below. With both V-sync and G-sync turned on, V-sync will have absolutely no effect on your fps or input lag until you reach the maximum refresh rate of the monitor. Be sure you have set the correct refresh rate in the Nvidia control panel. Now, if you don't mind some tearing when above your monitors refresh rate and you absolutely despise input lag, then yes, disable V-sync.

Side note;
There are two settings in the set up G-sync menu. They are "Enable G-sync for full screen mode" and "Enable G-sync for windowed and full screen mode". If you are running into stuttering or low fps issues, set this to "Enable G-sync for full screen mode", this should resolve the issue.

[cestarianinhabitant]

The 2 most performance impacting settings are AntiAliasing and Shadows. (Some forms of AA have little to no impact on performance, but in general these options look significantly worse than when AA is just plain off)

[helthuismartin]

The best video i have ever seen for my MSI GTX 970 4GB videocard playing Battlefield4 on a 27 Inch 1440P 144HZ 1MS monitor in high settings making 162 FRPS.Thanks Jayzz.

[suleymanbatuhanaylak]

Dude how are you so fun to listen but so on point every video? I love your videos :)

[kennethralcock]

Jay, I really appreciate your: in-depth, "keepin' it real", and somewhat nutz approach to your videos. Of all your videos I have watched so far, this one was the one I personally needed to watch the most. So a very heartfelt thank you for taking the time. And I believe you explained it at the correct depth-of-understanding for the masses (something you once lamented in another video you had to remind yourself to do).

I have a topic suggestion, one I have been looking for a video on from any PC Tech YouTuber that I cannot seem to find. I've seen you cover PC cooling solutions as crazy as hooking up a portable AC unit to a PC water cooling loop, as well as your 4 x 480 panel fan radiator solution (Brillant, I want one!). But how about covering a much more practical topic that I have seen a lot of Internet debate over (more so in the past than nowadays)— Measuring the effects of using a shroud or a silicone/rubber gasket on radiator fans and CPU heatsink air cooling fans? There are two main points of contention. First, that all fans (of any type: AF or SP) must have a dead spot in front of their motor hub where air flow cannot pass through an adjoining radiator or heat sink; thus, using a shroud to move the fan away from the fins of the radiator or heat sink should theoretically increase the amount of fins surface area making contact with the air flowing through them. The second point is that a rubber or silicone gasket, might offer a smaller shroud-like effect, but also reduce or eliminate any airflow leakage that might escape between the edges of the fan and the radiator or heat sink. I would really love to see you break this topic down JayzTwoCents Style, and also use your incense smoke (go all zen on us) to show us the difference in where and how the air flows: A) no shroud or fan, B) Shrouds only, C) gaskets only, D) Shrouds and Gaskets.

Again, thanks for providing informational and entertaining content!

[tkout]

You forget to mention Fast Sync and AMD Enhanced Sync for displays that don't support adaptive sync technologies! Probably could dedicate a video for AMD & Nvidia control panel settings too.

[saltonaburger]

Last week I posted a comment for the first time on this channel requesting this exact topic to be discussed. Jay delivers 100%.

[vasilije]

Once you try ambient oclusion and tesselation, you can't really go without them. It really adds to the graphical fidelity of the game.