Nvidia Geforce ShadowPlay Vs Fraps Vs Dxtory – Review and Benchmark Milos Milosevic November 15, 2013 Benchmarks ShadowPlay Introduction When competition is breathing down your neck in terms of innovation, the only adequate response is providing even better quality , especially in the field where competitors are lacking full potential. Nvidia is bringing to the table something that every gamer and youtuber has ever wanted. A piece of software (part of Nvidia’s GeForce Experience package) that will enable gamers to record their best (or worst) moments with a minimum impact on their systems performance. All that at 60 fps! Nvidia also claims that 1 minute of gameplay recorded at those 60 frames per second will take only 370 Mb on your hard drive, which is something crazy even to think about how big chunks of files are created with other pieces of recording software, such as Fraps and Dxtory. One more very nice feature from Shadowplay that was announced by Nvidia is called Shadow Mode, the ability to record the last 20 minutes of your gameplay if you have the option enabled. When we thought about how we are going to proceed with testing this new craze, we determined that comparing it with other popular software is a must. We also wanted to investigate how big an impact on performance Nvidia’s Shadowplay will have with slower CPU’s and since we had finished that part , we tested Fraps and Dxtory with those slower alternatives as well. This was all done in order to provide the most accurate testing and performance comparison in the game recording sphere. System Requirements If you want to use Nvidia’s Shadowplay you will need an Nvidia card , and not just any one but a 6xx or 7xx series. Nvidia states that you will need at least a GTX 650 if you want to record your epic moments using this new Geforce Experience feature. Those of you with notebooks that have embedded Nvidia Mobile GPU’s, will not be able to record your adventures at this point in time. Nvidia does not mention any other hardware requirements such as CPU or even Ram and that’s one of the main reasons we decided to test all of the CPUs in order to to solve this bit of the mystery. Nvidia says that most (if not all) of the recording work will be done by Kepler’s dedicated hardware H.264 video encoder which might be the reason Nvidia decided to skip any other hardware specifications. Test System Setup In order to prevent any bottleneck we used 2 Intel 520 Series 240GB SSD’s . One was for the OS, Software and games and the other one for storage. However even this caution will be proved wrong down the road, but more about that later. The testing was done on 6 modern CPU’s with one GTX 770 4GB. We set to record gameplay at 60FPS with all of the testing software: Fraps, Dxtory and Shadowplay of course. [komper pid=34,35 category=9] Software Overview As you can see, Shadowplay at first and even at second glance doesn’t look like such a big deal. There aren’t many options for users who want to to customize their recording, whether it comes to sound or codecs. You can choose three modes: Shadow and Manual, Shadow only and Manual only. As we already mentioned Shadow mode runs in the background while you are playing and after pressing the Alt+F10 buttons saves the previous 10 minutes on your computer. With manual mode you can turn on Shadowplay manually. Shadowplay & Manual are a nice combination of those two previously explained abilities. The next option is Shadow time where you can choose how many minutes of your last gameplay will be saved. The other options are pretty generic, you can choose the settings when it comes to quality of your recording and whether you want to record your in game sound or not. If you play at 1440p or higher resolutions, we must note that your game play will be recorded only at 1080p. 1920 x 1080, Shadowplay Vs Fraps Vs Dxtory In order to make this benchmark as accurate as possible we first measured FPS without any recording software and than recorded FPS while the different software solutions were doing their thing. All this was tested at BF4 Multiplayer on the Map Siege of Shanghai after one minute of recording which was more than enough to determine performance impact while the recording software was running. As you can see what Nvida promised, they delivered. In all scenarios Shadowplay only affects performance by 10% but the same could be said for Dxtory as well, however the results with Fraps are more than disappointing. Fraps kills the frame rate by 40% on average and there is a legit question about why anyone would want to use it. When it comes to CPU performance we see that all of the processors (more or less) take an equal amount of drop in terms of performance. Although we do see that the the lower rang I3 and FX 6300 performance is dropping slightly more than 10 percent with Shadowplay. However lets not forget that Shadowplay is still in its Beta state and minor glitches and optimization issues are highly expected. Compared to other pieces of software Shadowplay does excel, especially if you consider that Shadowplay is free which is not the case for Fraps and Dxtory. Although they do have free version as well, those versions are more or less useless when it comes to any serious attempt of recording your gameplay. Frap’s free version records only 30 seconds of gameplay while Dxtory’s free version carries one big watermark on top of the screen. Average Storage usage Now this is an important part, maybe the most important part of all the tests.Whether you have just one HDD dedicated or piles of HDD’s with 2-3 TB each, you need to know how much space is required for a certain amount of recorded gameplay. The results are interesting to say the least. Shadowplay does take only 370 MB of storage on your HDD or SSD for just one minute of in-game recording, which is truly amazing, especially if it’s recorded at 60 fps. For the same amount of time Fraps will eat a bit more than 5GB, and if you think that’s a lot, Dxtory needs *only* 17,6 GB to store one minute of gameplay. We tested this more than once, so any mistake in testing it is out of the question. Recorded FPS Of course we also wanted to make sure whether the end product is truly recorded at 60 FPS, low impact on performance and low requirements in terms of HDD space does not mean much if you don’t get what you have been promised right? Shadowplay shines again, giving the best results and far more consistency as well. Fraps only managed to record at 46 fps on average while Dxtory only 30! We redid the tests for Dxtory at least a dozen times and the end result was always more or less the same. We think that our SSD’s, Intel’s 520 Series 240 GB (one of the finer SSD’s you can get on the market) was a bottleneck for Dxtory and this is the reason why we got the end result of only 30 fps with 17, 6 GB of file size. If you want to record with Dxtory at 60 FPs you will probably need 2 SSD’s in RAID O, in order to write those enormous amounts of data We do not dare to imagine how big a 10 minute file recorded at 60 fps under Dxtory would be though. Image Quality Comparison But all that we set aside. having videos with much bigger file sizes ensures that those clips would be top quality. I mean how good can Shadowplay’s image quality be when it takes 15-60 times less space than the other recording software? A picture says more than a thousand words. It is noticeable that image quality with Shadowplay isn’t as good as it is with Fraps or Dxtory. The image is a bit blurry and the colors are kind of watered down. If you consider the size of the clip, the results are more than impressive. At the end of the day what you see and your audience of course, is the most important thing of all. It all goes down to personal preference, but we truly doubt that anyone would complain about the image quality of the clip recorded by Shadowplay. When you record your gameplay and present it to the public, the core attention is turned to what you do in the game and not to the actual image quality of you clip, unless its really really bad of course. But we believe that Shadowplay does if nothing else fulfill minimum standards when it comes to image quality. Conclusion This thing works! This is probably the most useful feature that Nvidia has ever created for gamers, far more useful than that thing called PhysX (which isnt Nvidia’s invention in the first place). Nvidia has done an amazing job in giving gamers a very well made, although in its Beta stage, product. This company has recognized the trouble the gaming community is having when it comes to recording their games and we must say that we are blown away with the performance of this little thing called Shadowplay. The software is quite simple, however bear in mind that most of the people don’t know how to adjust recording software properly and that additional options might confuse them even more. Shadowplay with all of its simplicity and a $0 price tag is an nuclear bomb on the recording software market. Developers who are selling recording software will need to think about what their future steps are going to be. When you look at how Shadowplay works with a marginal impact to performance, a small clip size and relatively good image quality with all that at 60 fps, those people really need to wonder why would anyone buy their software? PS : For those of you who have already read our BF4 benchmark you will see certain discrepancies with results in this test. The Battlefield 4 Benchmark was made after 5 patches and this one has been made prior to new driver release and the big update. We think that DICE has managed to kill the framerate (again). However we hear that the frame rate with the last patch and update has been improved, but we do not know to which extent. They also managed to ruin something again in the true DICE fashion. If you have any questions about the review please ask in the comment section below and I will be glad to answer. cpy Agreed, i tried shadowplay, and it beats software encoders with both hands tied behind back. Seabottom I tried shadowplay and it didn’t work. I went to the Nvidia Forums and they told me that shadowplay doesn’t work in surround mode. Like WTF? They make an awesome piece of software but it only works with single monitor usage LOL http://www.totallydubbed.net/ TotallydubbedHD I love shadowplay – but the claims of 40% by FRAPS and 10% by Shadowplay are ridiculous. I did all my benchmarks of shadowplay vs fraps in 64p servers, running ultra on my i7 3770k @ 4.5ghz with my MSI GTX680OC and FRAPS lost less than 5% and shadowplay lost less than 1% in performance. In perspective for you: 10-15FPS hit on your current FPS whilst running FRAPS and less than 5FPS whilst running Shadowplay. guniu Was dxtory used with default codec? You know the biggest advantage of dxtory is ability to choose codec of your likings? I’m using dxtory with x264vfw codec and my 1min files are around 350MB. I know, shadowplay is still better cause it’s 60fps, but your 17,6GB is ridiculous. You definitely did not use dxtory as it’s supposed to be used – with decent codec. Milos M We tried x264vfw and couldnt get it running we got a black screen. From what we were able to find out it is some error with W8. blueman24 Then you could have tried some different codec like lagarith instead of the crap stock codec. Chris Agreed- with lagarith a fifteen minute video at 60 fps is around 10 gigs for me. Not ideal, but certainly not the 200+ gbs you’d expect with the results seen here. Desa I really don’t know what to do, to get 14GB file size with Dxtory 😀 My file size is much smaller. Tip of the day – Open Broadcaster is king.