Do capture cards cause INPUT LAG? | Understanding Input Latency Testing & HDMI Passthrough

are capture cards adding input latency to your games are they making you worse at your games slowing down your shots messing with your accuracy any of that that's the premise of today's video whether or not capture cards add input latency to your game when running through hdmi passthrough and the simple answer is no but also sometimes yes we're going to be discussing the results in today's video as i have a few different methods of testing input latency and we're going to talk about those and how they actually you know impact what results you get which tool is the right tool for the job things like that i think you're going to learn a lot in today's video and i'm stoked to finally share it with you especially because my entire desk has been a disaster despite the rest of the studio mostly hasn't even finished up and it's because i have all sorts of latency testing hardware that i've been switching between and all of that so i'm glad to finally get that done this video is brought to you by nerd or die and their new vibrance stream overlay package whether you stream on twitch youtube or facebook gaming this clean concise but very colorful stream package might be for you if you like hard clean edges sharp corners and really being able to customize your color scheme to pretty much anything you can imagine with a really sick stinger transition this is the stream pack for you as always you have customizable widgets for recent sub follow tip alerts you got webcam frames you got chat boxes you've got borders you've got alerts you've got everything you can imagine and if you get the full complete package with the source files then you have the entire after effects file to tweak to your heart's content head on over to eposox dot gg slash nerd or die and use coupon code eposvox to save 15 percent at checkout i'm eposvox your stream professor and this is about to be a super nerdy video first and foremost what is latency or specifically input latency or input lag this is literally just a delay in the case of gaming it would be the delay from when you click on a mouse when you when you click someone's head and the photons firing on your monitor at the end result of that chain to actually display that happening on screen latency is affected by quite literally everything involved in your gaming setup be it the monitor itself taking the input and processing it and displaying it in front of your eyes technically even the air gap between the monitor and your eyes the actual processing speed and usb polling rates and things like that of your gaming hardware such as your gaming mouse or your keyboard uh your the video hardware you're using for example if you're running it through some sort of upscaler box that would probably add latency and then cpu and gpu performance of course actually being able to render your game fast enough if you're doing online multiplayer gaming there's a whole extra side of that with network latency and then of course you have the software side of things which is primarily the game engine actually rendering and taking in all of your inputs and rendering them to the screen but there's also you know background processes windows updates happening things like that can also affect your game's ability to render those things and affect your input latency there as well even things like graphic settings can affect your input latency vsync adds input latency that's something that's been known for a very long time and why adaptive sync is great because it doesn't add as much which we can start to test now and so basically everything involved in gaming affects your input latency which is why this conversation gets a little weird there are two types of ways to actually test input latency that i have available at least that i use for testing capture cards as well as other things and those two ways are the actual raw video input signal to output latency which you are seeing back here using the time sleuth there's also a leo bodnar device that i have as well that does virtually the same thing and this is literally testing the change of your monitor from white to black back and forth based on the timing of the output sent by the latency tester so it has a signal generator inside of it just like a little microcomputer that generates this black to white signal it spits it out over hdmi and has a photon detector to detect the light change between black and white based on how long it takes your monitor to display that this is great for testing monitors and video hardware that you're specifically testing the hdmi passthrough of which is kind of what we're focusing on here today capture cards in their hdmi passthrough but it's not useful for testing software tweak settings and it doesn't usually support arbitrary formats such as 1440p or greater than 60 hertz so it's a little limited in what it can do but it is a great way of getting a direct read on what you know your monitor or your video hardware is capable of the second type is a full click to photon latency that tests the entire chain and for that we're using this tiny little device here called ldat from nvidia this is a similar concept to the video signal latency tester only it's hooked up to a mouse and then you connect it to a second computer because it takes the immediate click the electrical you know signal showing that a mouse clicks it takes that signal and shoots it up here as well as to the software to then measure how long it takes before after you click for there to be a significant luminance change on your monitor basically this great you know white to black test here and then determines how much latency that implied and so you have to set up some very rigid structured testing to make sure every time you click it causes the right you know change on screen in order to be detected but that's what this is for now for any concerns because i you know the name nvidia is on it uh gamers nexus as well as i believe someone else battle nonsense maybe have both put up videos about comparing it to their existing latency testing which all of us did at one point and i immediately as soon as i tried it out i just said nope and didn't want to go back to it but it involves setting up a very high speed camera synced to either the exact or double the refresh rate of your monitor and counting the time between an led hooked up to a monitor just like this does and the time something changes on screen it's the exact same process this just allows us to do it more efficiently faster and with a little bit more accuracy because we're not having to fumble around in video editors to make it work so super shout out to nvidia for sending out this hardware because it's not something you can just go buy this is kind of a hacked together indie project within the company and it is freaking awesome overall i'm super stoked to combine the ldat into my existing latency testing suite to test for uh streaming settings and things like that because i'm also putting it together with new budget streaming rigs that i've been working on and getting together so that i can test how various software tweaks optimizations that people recommend that i usually i'm like about um as well as obs settings and things like that how they affect the input latency of your gaming as well as your streaming with a variety of hardware and we're testing one of the budget streaming rigs today along with a high-end rig so back to the initial question of do capture cards impact your input latency when you're gaming and the answer is yes and no because some capture cards can and they're kind of designed that way for example most blackmagic design capture cards with hdmi passthrough usually have a frame buffer on board that allows them to alter the output most intensity products allow you to kind of output a different signal entirely or at least a scaled or frame rate converted or what have you signal to the output it's not just a straight pass-through you can actually control it and tweak it a little bit which seems desirable except the ability to do that requires a frame buffer which means it stores the frame kind of on the card before spitting it out to make sure all the changes are done which is at minimum one frame of delay which is 16.67 milliseconds or 33.33 depending on if you're 60 or 30 fps respectively but it can vary up to four ish frames per second of delay as well as it's the case with the blackmagic intensity pro 4k so that's a little iffy there that being said some of the newer blackmagic products such as the atem mini and that whole line that i covered a while back uh you can actually disable the frame buffer for the output and set that the first input is real time pass through to the output which is pretty neat theoretically some cheaper capture cards with pass-through could introduce delay as well they they are using cheaper components making clones of the same product with updated hdmi revision boards and they could throw on a cheap splitter or some other form of splitting the signal that adds delay and you may not know it that is why i always test pass-through latency in my reviews as well but generally speaking capture cards that actually claim to have lag-free or real-time hdmi pass-through for your games are true to those claims that's just kind of how they're built such as from elgato avermedia and other companies they include hardware that is basically the same thing you would find on an hdmi splitter for that uh that way this hdmi signal is split both to the onboard processing chips to digitize it and whatever for the capture card itself and just directly spit back out to the hdmi output based on some tear downs that are good buddy arsenio dev pointed me too hdmi splitters act as a tv accepting hdmi standard signals including hdcp and then the split signal is a dvi hdmi signal that isn't encrypted this is how many of the old ones bypass htcp in the case of capture cards however the hdcp integrity is maintained to the processing chips which act as the final tv receiving end which maintained compliance with the hdmi standard but then your pass-through is spit out in the hd hdmi dvi kind of format which doesn't have hdcp necessarily maintained this is how you're able to pass through hdcp signals and see them on your display but the capture card is like now you can't capture that that being said the hdmi splitting process unlike analog which is electrically literally just boosting the signal in two directions isn't perfect and isn't 100 real time as nothing in the digital realm is because it's digital it's not analog it's not just an electric signal it is a digital conversation that literally involves a handshake between devices to verify that the signal is compatible so there's always a little bit of fluctuation so the most accurate way to verify that an hdmi passthrough device is actually real time is using a signal latency tester such as the time sleuth or the leo bodnar latency tester as these are literally testing just what the hdmi signal is doing and so with this monitor when this is currently running through the capture card but when it's not running through the capture card the monitor's latency is 2.41 milliseconds at the top of the screen 9.44 milliseconds at the middle of the screen and then 17.38 milliseconds at the bottom of the screen keep in mind screens running at 60 hertz have a minimum of 16.67 milliseconds that they can be at the bottom of the screen because even though it's not quite the same as crts most monitors are still rendering line by line or you know displaying line by line in order to render out the signal and so for 60 frames per second in order to get the full frame rendered out at the bottom would be a minimum of 16.67 milliseconds now there are some editing monitors that do page flipping where they wait until a full frame is done before flipping to the next frame to avoid screen tearing and things like that but those would be abysmal for gaming now running it through a capture card such as the elgato 4k60 pro we can verify that the input latency to the same monitor just running through pass-through now is 2.88 milliseconds at the top 9.75 milliseconds at the middle and then 17.86 milliseconds at the bottom of the screen so you can see there is technically 0.3 ish 0.2 to 0.3 milliseconds of input latency added there but that's not something that would ever impact you in any way that you would ever be able to tell or feel we talk a lot with input latency especially with frame rates and things like that that you you can't see it but you can feel it less than one millisecond is not something any human will ever be able to feel ever and it will not affect their performance but of course to verify this we used our new toy the nvidia ldat to run some tests i've tested many times you know many different testing scenarios with 200 data points per test and repeated the test multiple times to make sure my results were consistent and right off the bat i do want to note here that since this is also measuring you know the latency of the game engine and all of that that so many things happening on your pc can impact the latency of your game and suddenly i can just restart the computer and suddenly the game is rendering three to five milliseconds different you know slower than it was the first the previous reboot and so this took quite a while for me to get you know consistent results in a way that were isolated from potential software changes that might be impacting the results so [Music] this was an experience my test cases were as followed no capture card in the system at all so we're just measuring the game to the monitor output rendering latency plugging a usb capture card into the system but not actually using it you know no signal passed through yet then running pass through through that usb capture card and then running pass through the usb capture card but the capture card is plugged into a second pc so we're using a dual pc setup then i also tested the same thing where i plugged in a 4k60 pro pcie capture card into the system didn't use it used it for passthrough and then used it for pass-through on a second pc so those are my test cases and after testing on both a high-end rig with an intel core i9 9900k at five gigahertz and rtx 3090 to just go basically as high as it gets and then a low end rig which is running a ryzen 3300x with a gtx 1660 graphics card i found that specifically the hdmi passthrough of capture cards does not add any meaningful amount of input latency it's 1 milliseconds or less which is completely insignificant however added input latency from these tests did occur and that can happen mainly on the lower end pc simply by adding the capture card to the system which was my theory when i started seeing other videos suddenly claiming contradicting previous videos that capture cards were adding latency to your game and things like that in that my theory was it's not the capture card hdmi passthrough that's adding latency but the capture card being added to the system be it using additional pcie resources the chipset taking away resources from the graphics card the driver doing something in the background just by the system being there you know there's a lot that can happen when you plug a new device into a system that uses up resources that could be adding the signal and this is kind of what i'm seeing here now i want to note that generally speaking this is bad practice adding a capture card to a single pc setup is not useful in any practical way other than for literally three or four games i can think of that it could really benefit in terms of making it easier to hook the game depending on your streaming or capture software and not not all software even struggle with this in the first place but in terms of performance there are no performance gains to be had by adding a capture card to a single pc setup this is bad practice and it's a waste of money if you're using a capture card it should be plugged into a separate pc where you would avoid this issue entirely however with my budget streaming rig i did see an added of two to four milliseconds of input latency by adding a pcie capture card into the system versus just gaming on it normally this is of course then reflected in the hdmi passthrough using that capture card being equal to that two to three to four milliseconds of added input latency over normal but not added input latency over just putting the card in so if we go from gaming to plugging the card in that is a jump in the input latency but going from plugging the card into using it for hdmi pass-through there is no jump for input latency now there are some fluctuations text tests to test but anything one milliseconds or under as far as i'm concerned is zero lag even in terms of processing things on a crt which is analog and as lag free as it gets less than one milliseconds is considered to be lag free still so while there are specific capture cards that can add additional input latency to your hdmi passthrough because they're designed to do so the gaming capture cards that specifically are built with real-time hdmi pass through in mind do not have added input latency that is my conclusion they don't of course this will vary on a card by card basis which is why i test this in every single one of my capture card reviews because it's something that needs to be tested and confirmed before you go buy a capture card and then find out that it is adding input latency to your stream now keep in mind that this is different than the input latency that i also test for your obs preview in all of my capture card reviews i also test obs preview latency to determine how you know how far off your video signal is going to be from real-time audio running on your signal that is utilizing obs and a whole separate conversation we'll dive in into in the next part here but i did also want to mention that you can see in my results that the difference between running a capture card in your system versus running obs on your system which does add at least a few milliseconds of latency is pretty significant depending on how powerful your rig is now on the higher end rig with the 3090 obs didn't really have any significant impact but on the lower end rig we did start to see a little bit of impact so the difference between running just your game on your pc and running in your game running through the hdmi passthrough of a capture card of ones that are built for real time no difference now that i have new testing methodology and i've been getting through all of this i will have a new presentation kind of format for all of my capture card testing data because the way i was able to release it before wasn't ideal so i've been working on new ways to present this data so that it's more accessible to you and i'm looking forward to getting that put up and things like that so be on the lookout for that subscribe for that if you're interested uh this is just kind of validating my claims that uh capture most you know which capture cards have real time pass through or don't based on my previous video signal testing and ldap was able to verify those results and validate that that previous testing methodology was correct as well as give you all an introduction to this wonderful piece of kit that nvidia sent over that we'll be using more in the future to test things like obs preview delay syncing audio whether game mode or some other windows you know setting might impact your game delay and things like that as well so hit the like button if you enjoyed subscribe for more tech education and stream guides i'm your stream professor eposvox and i will see you next time

Channel: EposVox

Views: 56,699

Keywords: input lag, 4k passthrough capture card, cheap capture card, video game lag, input lag fix, capture card passthrough, input latency test, input latency tester, capture card input latency, capture card input lag, do capture cards add input lag, testing input lag with capture cards, nvidia ldat, nvidia ldat tool, nvidia ldat capture cards, hdmi passthrough, hdmi passthrough capture card, hdmi passthrough input lag, capture card, game capture obs, capture card delay

Id: gcQCuu86GN8

Length: 17min 21sec (1041 seconds)

Published: Mon Mar 15 2021