Gadgetory

Andy has finally killed its rapture and gaming evolved integration the incumbent utility that dismayed us on a number of occasions today with the AMD relive updates to the Crimson brand drivers and these implemented its own solution to software capture for gameplay and retroactive gameplay capture this is a direct competitor to the shadow play software from Nvidia's GeForce experience suite and performs many of the same functions and has the same end objective years ago we did this comparison with shadow play versus fraps versus GVR which is and these previous utility before it was subsumed by raptor and the gaming evolves platform and we're back now that there's a new tool again we live under an these latest Crimson brand update and this tool we have side-by-side comparisons for we've also got some benchmarks of the framerate impact from doing live capture with shadow play and with the new relive tool and then finally we have some storage capacity requirement benchmarks between the two to look at the per minute data consumption of each software solution and these announcements for today contain a whole lot more than just the relive update but that's all we're focusing on here because really it deserves the most direct focus check back tomorrow for more analysis and discussion of oak at the overlay and benchmarking tool that couples with presentment and for discussion on other tools within the AMD suite like an update to the chill platform but again today's focuses on relive let's start with walkthrough the interface which should be immediately apparent that the utility is in fact capable of desktop capture Andy is using an overlay similar to shadow play for capture of screenshots and video relying on even the very same hotkey that shadow play uses alt Z by default though that's changeable the hotkey pops up the overlay through which you can easily modify capture settings toggle capture without a hotkey or begin retroactive gameplay capture and broadcast and the relive at solution plugs into twitch directly as well so that offers customizable streaming settings for the platform onto which you may be streaming your gameplay in many ways this is similar to shadow play identical really and we've highlighted a few of the key differences between the quality settings and other options in the interface in this quick table map so I'm supported bitrate is where we're starting this was checked using an rx 480 gaming X for the relive option and shadowplay was checked with a GTX 960 gaming X because we figured those cards would be the best for creating scenarios where will later see any potential performance loss and better represent the current end-user market with shadowplay were able to output a level of quality that rapidly enters into sort of placebo territory maxing out at 130 megabit per second data rates and again that is limited primarily by your upload platform like YouTube GFE used to be limited to around 50 megabits per second depending on how you configured it previously but that's been expanded Reliv seems to max out at about 50 megabits per second again and that's at least with this card the RX 480 it uses similar low medium high quality presets to shadowplay but with different tuning for the bitrate and for convenience we've listed the data rate for each preset in this table for each device or software though it's not an indicator which is better just the preset encoders for Reliv can be switched between ABC and HEV see while nvidia uses its own encoding solution both relive and shadowplay our hardware accelerated encoders with designated GPU components responsible for handling all that encoding on the card this moves the workload off of the CPU and is something that was really emphasized in our original video on shadowplay because it's critical to performance and retaining that performance while capturing video moving down the list we've got a default audio bitrate of about 184 kilobits per second when captured with shadowplay without options for customization so that's a fixed data rate and it was inspected when we looked at the files and he gets points here as the tool allows customization of bit rates up to 320 kilobits per second for the audio there's minimal Delta between the higher bitrate audio captures when using Reliv in terms of the file size so running with higher audio bitrate isn't really going to impact your performance in an exponentially negative fashion both tools enable custom audio input device selection and choice between push-to-talk are always on microphones and in-game audio recording and as a few interesting side points Reliv has multiple overlay options when recording so one of them is overlaying your own image on the captured screen or game or what ever and that could be something like a watermark so this is a pretty simple feature that I'm actually happy to see as content creators because this means if you were a gaming channel and you wanted zero editing involved in terms of post process stuff like that and you just want to record your game and upload it this will allow you to still watermark that footage to make sure you've got your branding on there and that it's not easily lifted so that's added in relive and they've also gotten overlay that's just kind of interesting but not necessarily that functional and in the bottom-right or whatever corner you select there's a little slide out when you start recording that list the system information that so that'd be the CPU and the GPU basically if you wanted that present in your video so it's not quite as advanced as something like xsplit or OBS but it's also not trying to be for example there's no way to do scenes or overlaying multiple of your own images and graphics with relive or with shadow play it's pretty basic you've basically got webcam overlays and then the other options that I just listed let's get into the benchmarks between the two first here's a side-by-side captured comparison of an identically executed benchmark scenario both captures were at 50 megabits per second despite being able to go higher on shadow play and they had roughly the same audio bitrate though shadow play was slightly lower video playback basically looks the same it's really up to you to compare that though and granted YouTube will compress this a bit but it should look about the same because they're both using Hardware encoders and they're both capturing at the same bitrate when we look at the file size comparison per minute of the capture this chart is what we end up with for reference our old data from fraps benchmarking is also present here since that software hasn't changed in years or even been updated and we're looking at file sizes of nearly 5 gigabytes with fraps for one minute of recording or about 4,700 megabytes which records effectively losslessly and without GPU acceleration that's a bit rate somewhere in the range of 638 megabytes per second weigh in to placebo territory and that's assuming no compression plays which with fraps there probably are none shadowplay the newest version anyway is capturing at about 366 megabytes for the entire one-minute run and Reliv is capturing at about 370 6 megabytes with the same pass that's a 10 megabyte difference and this is a repeated benchmark so we know that the data is about the same for each pass though we do have a bit rate difference in audio that we can't control for regardless relive and shadowplay are effectively identical in storage requirements and should it really be a heavy point between the two of them next we'll be looking at the impact on gaming performance when using either shadow play or relive this is to say an analysis of the framerate loss when using either tool because the way these work you're inherently going to have some loss now that could be i/o bound it could be in the software side it could be GPU or CPU bound it really depends on the solution you're using it with fraps given that bitrate obviously you run into a storage bottleneck pretty quickly it's way faster than what most hard drives can reasonably handle and fraps also doesn't really do GPU acceleration at least not like these do where you rely on individual custom-designed encoders on either AMD Hardware or Nvidia Hardware to perform your encoding of the live captured video so both of those solutions for these two cards natively will be far better than what something like fraps can do but OBS things like that do have integrations with encoders that work with hardware we're just not looking at them today so as stated the full testing methodology is defined it for this in the article below if you're curious about what hardware was used or how we tested a particular game or captured the gameplay starting with dirt rally at 1440p with ultra settings using an RX 480 gaming x4 AMD or gtx 1060 gaming x4 nvidia these are the results were seen note first that this is all relative performance scaling since raw FPS means nothing to us given the difference in video cards used for testing the GTX 1060 using shadow play is losing approximately ten point nine percent of its performance with dirt rally when capturing gameplay and that means it's moving from 98 FPS average to 87 point 3 FPS average and with compare ibly reduced 1% in 0.1% lows the RX for 80 using Reliv is losing approximately 3.9 percent of its performance with the game and the same capture settings or a shift from about eighty six point seven to 83.3 FPS average and note here that the performance on and video with shadow play is a bit worse than what we see in other games for example moving on to GTA 5 we see that shadow play is now performing better than relive and also obviously better than it did in dirt this is certainly something we saw previously in our lab set of tests fraps and GB are and in those tests gbr was doing worse in some games than in others and a lot of that varies based on how a particular game interacts with the overlay provided by the vendor so it's not necessarily a hardware level thing anyway the performance differences peg shadowplay and the gtx 1060 at a performance loss of 3.3 2% moving from ninety nine point three to 96 FPS average basically negligible and the arch 480 with Reliv loses three point six five percent of its performance I moved from 93 to 87 FPS average and this is all really good news it didn't used to be the case that you could use these hardware accelerated software solutions for recording desktop or gameplay footage and you had to rely on things like fraps which was either well both CPU intensive and storage intensive in ways that would have your frame rate sometimes depend BER you had underline so that is where we're moving and where we've been moving for a while it's all going towards GPU acceleration it makes a lot of sense that's what these are built for and they have individual components as I said for the encoding process of the video so you do always lose some performance that's the nature of this it's just a matter of how bad it is and with both of these tools it's really negligible it's not that bad at all now dirt was an interesting scenario where we saw a worse performance with the 10 60 and shadowplay but that's not reflected in every game and the same will be true for this if we expand its test scenarios again depends on how the software the game interacts with the overlay the biggest takeaway here is that where shadow play was already really good at what it does Andy now has a competitor in the form of relives so you don't need to rely on third-party solutions like OBS and xsplit to get the same functionality that Nvidia offers natively now both vendors offer native gameplay and desktop capture solutions again a good thing and both of the tools are low overhead so you're not losing too many frames again dirt rally a bit of an outlier here but not losing a lot of frames either way and then in terms of storage you're not really consuming much of that either 300 something megabytes or less in some cases depending on what you're capturing for about a minute of footage really not something to complain about especially when we look at where we came from which would be raps at 4 gigabytes per minute so that's good news overall the compression is good and not that lossy you obviously lose some of the data you always do or I should say you lose some of the quality of that data but it's not bad and YouTube will compress it more anyway so if that's your end goal who cares both perform well they produce about the same visual quality now it's just no longer really a point of comparison if you're considering buying a card that's something you can mark off the list because they both have a solution that's similar so we're happy that Andy finally has its own software solution for this task because that was kind of a driving reason to get an Nvidia card if you wanted really easy to use software capture for whatever you're doing and now they're level on that playing field and in the future we'll be looking at okat which is basically an interface that Andy built for the open source present amount projects which is what we use for dx12 and vulcan testing it's an important thing to have but it's not easy to use it's all command line driven so we'll talk about that probably tomorrow subscribe for that content links in the description below for more on this particular content as always and I'll see you all next time