Gadgetory

we recently published an exhaustive study and benchmark on Apex legends performance across a wide range of GPUs for gaming PCs but now it's time to turn our in-house software loose on the Xbox version of apex legends the game is no doubt popular and given its requirements maintain consistently high frame rates we thought it about time to see how well the game runs on our Xbox one X went under various load conditions throughout the game including fov testing all the way up to 110 degrees today we're looking at apex legends at frame time performance and frame rate on the Xbox one X before that this video is brought to you by the gigabyte ARS ad 27 QT gaming monitor the ad 27 QT is a 27-inch 1440p gaming display with 95% of dcpip 3 color saturation for high color accuracy accompanied by a 1 millisecond response time 10 bit IPS panel and display HDR visa certification additional features include fluid adjustment and slide RGB LEDs for personal Flair and firmware features like cooldown counters at reticles and adaptive noise reduction learn more at the link below we can pull some information from our earlier PC testing to inform our Xbox testing of apex legends in some of our tests we found that performance varied heavily from region to region and apex legends we can show some PC footage of that with the initial drop into the game exhibiting the worst example of performance we mostly wrote this off as inconsequential seeing as it only happens once per game but it's still important to be aware of your performance floor the same is true with the Xbox as we'll show you throughout this video and the reason for that is because of the tremendous draw distance in apex legends framerate primarily drops more as you look out toward the horizon at a slight isometric angle downward to observe the heavy geometric complexity of some of the mountains as opposed to looking straight down and seeing a much more limited scope of that terrain geometry some commoners have postulated that the game is rendering everything constantly and it's full quality but this is extremely inaccurate it is false plainly the game still calls objects like nearly any game does it's just that the draw distances are long enough and there's a lot of complexity in the distance environments that you do see the heavier impact the framerate where other games might not show it there's clearly LOD scaling some of our PC footage illustrates this as does the Xbox footage as there's it means to optimize for performance without sacrificing too much of a competitive edge there's also object : and poppin and it's visible most clearly during the drop in sequence when you can heat some of the containers disappearing and reappearing on the camera depending on how close the player is to the ground apex legends on Xbox one X runs at the near constant 60fps respawn has chosen to be aggressive with dynamic resolution scaling on consoles so although the game doesn't run at native 4k the responsiveness and smooth frame rate make it preferable to our experience with pub G on the Xbox now we can pull some of our old footage as a reminder of that overall poor experience before we proceed with performance numbers here's a quick refresher on our software this is sort of a tutorial of the software well you're seeing on the screen is an example of some of our software capture from a previous review maybe the PlayStation classic frame times are presented on the bottom rather than a frame rate as frame times are the most accurate representation for live frame to frame playback the playhead in the center is the current frame and ideally the line is perfectly flat to indicate a steady framerate incoming frames are on the right side of the line pass frames are on the left side of the playhead and the axis is in seconds the current frame time is displayed in the top left rather than the frame rate because frame rate is inherently an average over time frames per second and so you sort of if you're looking at the frame rate of one frame that you're looking at now it doesn't make a whole lot of sense so we show the frame time for that and frame time if you don't know is a way to describe the frame to frame interval or the period of time required to render the next frame presented on the screen it was present being another keyword in frame delivery a frame time of sixteen point six six seven milliseconds is 60 fps thirty two point three three is 30fps and fifty milliseconds is 20 fps if you do prefer to use frame rate because it's a bit more relatable this is determined by just dividing a thousand for the number of milliseconds in a second by the frame time which results in your frame rate the trouble is that frame rates are inherently misleading and DBA from the base metric of time they're still really useful though they're incredibly useful we use them everyday they're really easy to relate to but it's not the perfect metric and so it's important to look at something like frame time for the fluidity of the output let's look at Apex legends now I'll show one of our initial combats with relatively high system load we've probably been showing one of them up until now as well there are definitely some frame time spikes to be seen in some areas despite an overall relatively smooth sixteen point seven millisecond or 60fps experience the most stressful and most repeatable these spikes is in the initial flight or the initial fight in high conflict areas like high tier loot zones and in that drop where you're flying down towards the ground even the Xbox one X which is the high tier of the Xbox and I think that's the plural struggles one locate at distant areas on the ground spiking to thirty three point three milliseconds constantly and occasionally even 50 milliseconds are about 20 FPS if you wanted to convert it looking straight down or at the sky is of course less demanding as fewer objects are drawn and the GPU gets a break from the extreme draw distances as does the CPU because it's getting hit with fewer draw calls and the heavier drop load reduces as the players View approaches the ground and fewer new objects need to be streamed in with objects exiting view getting cold from the render pipeline the drop is the worst case scenario in the game or one of them we do have one instance that's a bit worse than this with the creeping barrage will show that later but it's streaming new objects and textures rapidly into memory crunching geometric data and filling the screen with environmental data as you come down towards the Earth and because the draw distances are lawn even on the loot crate's which are critical to game from a gameplay standpoint everything is just it's the render pipeline it's getting filled with a whole lot of data that's not really cold until you're extremely far away from it so this is mostly interesting as a demonstrations it's having a low frame rate before combat begins doesn't necessarily hurt gameplay still during the drop sequence those frame drops can be noticeable to players more accustomed to higher frame rates then say 20 to 40 FPS average in actual gameplay the most frequent spikes we saw weren't always tied to water or smoke as our PC benchmarks would lead us to expect in PC benchmarking and we can show one of those GPU benchmark charts now maybe 1080p or something with a lot of cards on it we've that the river zones had some of the heaviest GPU load when splitting the view frustum between the river and the distant mountainous geometry instead of this being the heavier loader on the Xbox there were occasional frame time increases while simply running around the environment didn't require anything special like water and part of this gets complicated because it's dynamic resolution scaling with a backs of legends so the resolution of the game is changing constantly as it tries to account for those more intensive scenes these often appear to be tied to loading in objects and new areas we have some footage of that and it continued to happen even indoors we believe that much of the frame timing consistency here has to do with dumping old data and rapidly pulling in primitives and textures for the upcoming scenes but it's more noticeable in this game than others due to those extreme view distances feeding into this there were multiple spikes while climbing ropes which would reinforce the hypothesis that rapidly fetching new data for wider view is most heavily impacting the pipeline although water didn't inherently mean a lower frame rate like it did in PC testing it didn't help once the going got tough and one of Patrick's fight scenes and I'll credit it to him because neither one of us is proud of our fighting abilities we can see grenades detonating water splashing from bullet impacts post-process filters as we heal and shield allies and fire from explosives and none of these is singularly impressive as a graphical effect but combining all of them together and then shoving it into an Xbox is difficult to process all of this together made for a fight sequence that stuttered between 33 points 3 milliseconds and the more desirable 16 millisecond frame times it constantly never quite finding consistency and that inconsistency is what's the most noticeable not just hitting 33 constantly or hitting 16 constantly you're kind of ok neither one of those especially 16 but inconsistency once you deviate beyond them the mean frame time of more than 8 to 12 milliseconds it does start to become noticeable to the player especially frame the frame but nothing was that bad overall truthfully it's still playable it was overall fine despite the intensity for a few seconds we get hit by another round of stutters when we're downed by a smoke launcher we have time stamped clip of that too for the editors whose smoke then causes more frame time variants as it thickens finally ending the scene frame time spiked erratically when we get hit by a creeping barrage which is something you can see coming on the frame time plot before it even happens if we pause the frame or maybe replay it a few times you'll see it coming up on the timeline ahead of the playhead and you know it's coming at this point so if you could actually see this in gameplay well one you'd be able to see the future but two would be really useful so there's a few times where if we look at this in slower motion you can see exactly when the frames are dropped frame time spikes all the way up to 67 milliseconds briefly and holds that 50 milliseconds for about one and a half seconds this is noticeable in gameplay and becomes stuttery although the stutter is tactically brief at one point five seconds a sixty seven millisecond frame time roughly calculates out to 15 fps and it's surrounded by a bunch of twenty FPS frames over that time period so because fps is a inherently a rate over time it helps to know the time that we're looking at 1.5 seconds here so with our average FPS for that 1.5 second sequence we're at about 28 fps and obviously sometimes a bit worse than others the frame the high spikes don't mean up the biggest drop seems to be from the heavy filters applied to the screen as combined with the smoke and fire effects being rendered just outside of the view frustum or slightly coming into it not everything is like this though we're just highlighting the most exciting scenarios by which of course we mean the most intensive because honestly when we're benchmarking consoles for the most part what you want is a flat line that's not exciting but we do get that and that's a good experience in a lot of instances with apex legends so as we can show on random gameplay clips that Patrick took performance overall is generally closer to the 60 FPS average target the game is absolutely playable on the Xbox 1x but the previous scenes that raised some concerns for combat on the lower end consoles we might branch out into testing these if it's of interest but we need to hear it in the comments below so please sound off if you have specific units you want us to look at changing fov is the only real graphics option on the console minimum and default is 70 and maximum is 110 increasing our fov to the max just to stress it didn't instantly tank performance but frame rate during the initial drop was a bit during gameplay there were some sustained periods of 33 milliseconds spikes that we didn't see in our default fov testing resultant of the expanded viewable terrain some of the poorer frame times we saw were during a fight where an incendiary grenade was detonated for example but the game still didn't freeze up or a spike beyond 33 milliseconds except instances like this where I'll just spike really hard and suddenly and you have a very visible frame drop for a fraction of a second and that fraction is noticeable in play as you can see in this clip repeated respawn has been careful enough with its resolution scaling that increasing F of V M Xbox one X is a better experience without a noticeable performance hit in most instances capturing footage while spectating didn't work out as well as we'd hope we have some examples of that so none of it was used for actual analysis nothing we've talked about thus far has been from spectating it's been from gameplay spectate mode seems it laggy and that's network lag not frame lags that video out lag so we get some of the classic source engine glitching out of the map with falling under the world or the appearance of the void and if spectating is important to you it's buggy to a point that the frame time analysis isn't all that relevant we very rarely saw any frame time spikes in gameplay exiting spectator now above 33 point 3 milliseconds even the most graphics intensive scenarios except for some of those extremely heavy ones that we showed so in those like the one with the creeping barrage they're rare enough that we can sort of write them off but it's still important to be aware that the heaviest firefights will still tank framerate it's just that hopefully they're not too common and as long as everyone's on the same hardware that's sort of fair from a competitive standpoint but it's still undesirable from an experienced and point and that's bad since we record at 60fps to analyze console frame times that 33 point 3 millisecond frame time translates to one skipped frame the smallest hitch that we can measure considering that many modern console titles are locked to 30fps and deliver every frame thirty two point three milliseconds apart apex Legends is doing pretty well overall on average as you can see in some of our less targeted or intensive footage the game sticks pretty close to sixteen point six seven milliseconds or 60 FPS with occasional spikes to thirty three point three and those occasional spikes when it's just one on that whole time bar at the bottom you can ignore it that is not something you will genuinely notice or at least not most of the time it's typically when you start seeing more frequent deviations or excursions from the mean that it becomes noticeable so it appears that the dynamic resolution scaling is largely responsible for the frame rate recovery and heavier scenes and that will wrap our apex legends xbox one I expense mark for now part of our team is flying out to Taiwan in China this week but we'll be running more console benchmarks at home base it's important that you post the console testing requests you have below and if you'd like us to compare to some specific PC configuration let us know that as well we'll look at that in the coming weeks otherwise get subscribe to the channel to make sure you can catch our Taiwan and China factory tours when we're looking at some of the PC hardware manufacturers thank you for watching go to stored on cameras access net to support efforts like this directly that is the best way to support us and I'll see you all next time