On the bright side, I can confirm that the recorded video is buttery smooth with the new hard drives, especially when motion blur is enabled in-game. No hiccups at all. Subjectively, the output looks a bit less smooth when the fps drops below 60, but it could be placebo... It makes sense though that the recording looks best when the in-game framerate is an integer multiple of the to-file framerate. In any case, playability suffers a little bit, which is why I'd like to keep framerate as close to 60 as possible Perhaps overclocking my CPU further might help. I've ran it at stable 4.5GHz before, that's another 7% extra performance, and if we're assuming a fully CPU limited scenario, it should increase a framerate of 56 back to smooth 60.
Recording at 60 fps instead of 30 fps drops my framerate by another 10 fps outputting Lagarith to rawcap, and by 15 if outputting raw RGB to rawcap. What does this tell you? Also noticed that recording in a CPU limited scenario with the game set to 1600x900 gives me a considerably higher framerate. (I know it was CPU limited because the framerate when not recording was the same for both 1600x900 and 1080p.) On the other hand, playing at 1080p but setting Dxtory to record at 1600x900 gave the same framerate as when recording 1080p at 100% scaling. This suggests to me that Dxtory taxes the system based on the actual resolution being recorded or downscaled. What I don't understand why there is an FPS drop at all, because recording drops the framerate also when PS2 tells me the framerate is GPU limited, as it definitely is when hanging about at the warp gate. As I posted above, the second hard drive did nothing to improve the actual gameplay framerate when capturing to rawcap, it only affected the ability of the hard drives to handle the framerate to file.
First, bench them and don't screw with it. There's no reason to lower anything. I assume we're talking about recording planetside. Play around with the number of cores you use: processing threads. How many cores you use for the best record/gaming performance is really up to the game. You can set it up to 8 because dxtory takes advantage of hyper threading. You might find you only need 2 or 4. You definitely want more than 1.
Okay, I'll set it to the exact value the bench gives me. I'll try different core setups and report back. Do you think setting thread affinity in task manager for Planetside 2 and Dxtory might help, i.e. separate threads for each application? Also edited my post above, FYI.
You have to keep in mind that you're attempting to record a game that's supremely cpu intensive with a cpu that's not overlocked a lot and with a 7950. Typically guys who record raw at 60fps are sli guys with multiple raid drives for storage. I think considering the game and the hardware you're doing really well. What's your fps like in big bio lab fights?
No, just use the processing threads section in dxtory. You have to play around with it more; you'll come on the right/best settings after a day or so. It's not easy or straightforward, but it is the best solution when you consider what you're doing. The alternative, if you want to approach the quality you're getting now, is Fraps, and with Fraps it's a guarantee that you'll be getting 30 fps. I'd take 70 fps over 30 any day; so would you. You didn't waste money on the other drive, if that's what you're worried about. You're going to find that other games (any and every other game) is going to behave very differently (read: much better). As fun as it is, planetside is a sloppily written unoptimized game; nothing we can do about that but accept this reality. Downscaling requires cpu; that sometimes it seems to do the same/worse is not a surprise; it all depends on what's happening at the time. Raw caps will always require the least amount of cpu. There's no scaling, compressing or writing to avi between the frame buffer and the hard drive(s).
I don't believe it goes much under 60 (when not recording), I'll note it down next time there's a big bio lab fight. Yeah. I'm going to get some excellent footage with Dxtory, that's for sure. I just want to find the limits of what I can do with this setup and get a good grip on what it is I'd need to improve if I wanted to do something more. I'm currently doing some very methodical testing in a GPU limited scenario, trying different settings and writing down the framerates and whether the hard disk performance can keep up. I definitely didn't waste any money as I got the drives with employee discount and needed extra space anyway. I'm also happy they're not too noisy (was a bit afraid of that, being used to 5400RPM drives), although they are still just barely then noisiest components in my rig. To be exact, it seems to me that it doesn't matter if you're recording to .avi or .rawcap, what matters is the codec being used. Recording Lagarith to .rawcap seems to give the exact same fps (and the exact same compression, judging by file size), as recording Lagarith to .avi. On the other hand, recording Dxtory Video Codec with no compression to .rawcap improves gameplay framerate.
It seems that way only because you haven't recorded in enough situations. The app is really good, so most of the time it won't matter whether you chose raw or avi; however, the mere creation of an avi itself requires some compression (albeit not much), which requires cpu, whereas there is no compression (or creation of a complete avi, which also requires cpu) saving raw caps. It will always require less cpu than saving to an avi. How that plays out ingame is another matter; sometimes you won't notice the hit, sometimes you will. It's been more than a year since I've tested the default dxtory codecs. The general consensus still seems to be that lagarith is faster and cleaner; however, if you find that your system responds better to the dxtory codec, then by all means use it.
I've never made a lagarith video that small. Perhaps it's because at that size there's so little compression it's hard to notice? It might be more discernible when your capture length is a half hour or more? It's not the size that matters anyway; what matters is that you, ideally, want to keep your cpu as free as possible to drive the card and run the game, which is why raw caps are always better in terms of saving cpu. Like I said, a lot of times it doesn't matter, some of the time though it does, a lot... it really all depends on the game and what's happening at that time.
For ten times the cost per GB. Two cheap HDDs will record video just as well as a single good SSD, but will let you record for much longer, and will cost the same or less. Depending on how much recording you do, the HDDs may well last longer as well. There is always going to be some performance hit, no real way around this. A ~15% drop is not that bad, better than I would have expected, to be honest. Software video capture polling the card 30 to 60 times per second and copying the contents of the frame buffer out over the PCI-E bus is not "free", there is going to be some sort of performance hit, even if you aren't encoding it on the fly. This is typical. Being able to write to the disk faster isn't going to make the load on anything any less, it's just letting you meet the minimum requirements for consistently smooth video on the disk's end.
Performance seems to become CPU limited whenever recording, even in the most GPU limited scenarios like idling way out in the outskirts of VR with nothing happening - according to the in-game framerate meter. I should be able to see an increase in performance while recording if I overclock my CPU further, I'll probably play around with that this weekend.
Just to update this thread and bring it to a close... I decided not to mess around with overclocking more. The reason I originally settled to 4.2GHz was that it was the best compromise between performance and quiet, cool operation... With my current cooling setup, I'm not really comfortable with the heat and noise that comes with higher clocks. Besides, higher clocks wouldn't come close to fixing the problem of sub-optimal fps while recording. The problem is more with PS2 being poorly optimized. I would need perhaps 40-50% more performance per core to get a solid 60 fps at all times. That's not a realistic solution because it's probably at least 3 CPU generations away before we get that sort of performance; I'd be better off switching to NVIDIA Shadowplay if I badly needed to keep my framerate up. In any case, the video looks fine even if recorded with an in-game framerate well below 60. If I delidded my 3770K and replaced the crappy stock paste, I might get good enough temps to comfortably overclock to 4.5-4.6GHz or so; perhaps I'll do that in the summer as a little project, who knows. As for the recording settings, I ended up recording directly to .AVI with Lagarith. The performance increase when recording to .rawcap is so minimal (if there at all) that it's just not worth having to wait for Dxtory to build the files, not with the large amount of video that I record. My "workflow" is as follows: 1. record 1080p @ 30fps to a single partition, preferably one file per game session, but new file every time I change continents or characters 2. extract audio files from recordings; I record two tracks - one for in-game audio, the other for microphone 3. cut and reencode to Lagarith .AVI interesting parts of the recording, using the microphone audio track's waveform to identify parts worth keeping 4. store reencoded clips in another partition for later use This way, I can pretty effortlessly document my gaming and keep all the stuff that's interesting, in a lossless format but without wasting too much space. Having a good naming scheme for the clips also helps when deciding what to use in an editing project.
This looks overly complicated... Just buy Elgado Game Capture and use obs, and or the gamecapture program that you get from their website. Setup a dedicated computer, or just use the same one on it. Give it maximum performance on the cores, your cpu is basically the same as my laptops cpu. http://www.cpu-world.com/Compare/82...tion_i7-3940XM_vs_Intel_Core_i7_i7-3770K.html My cpu is the i7-3940XM. I OC it to 4.2ghz, but usually never need to, i allocate 6 or 7 cores, sometimes all 8 to elgado game capture device process, or obs process, i record to a 3.0 2tb Toshiba HD. A 3hour 51min 14second video comes out to 17,383,267kb or 17.3gb in file size. Its in mp4 format. Recorded at 60fps. Smooth as butter video. I tend to record all of my gameplay for the moments when theirs that random special ****** 4th factioner or hack/exploit user to make em famous and file a report. But altogether more so to record my awesome epic kill runs. Anyway... yeah, for audio muxing, demuxing, editing, i use SUPER@ for demuxing the audio, and remuxing audio including multiple audio streams. For video editing i use VirtualDubMod 1.5.10.3 and for audio editing i use Super@ & Audacity. All programs i've listed are free, the only thing that isn't free is the Elgado game capture device, thats about it. I usually put in 5 to 6 different audio streams, game audio, and 2 input sources which Elgado game capture supports (it supports 1, but theirs a bypass to get 3 to work utilizing analog audio, component/line in audio, mic audio). the first audio source is the game itself, the second is through hdmi (which is what i use as the trick for getting analog audio to pass through a 3.5mm audio jack to the computers line in/component line in) and then the 3rd source is the Mic itself. The other 3 sources are music (itunes/whatever im playing at the time) utilizing line in, into my computer (my computers the Frigate Series). 4th source is music program, 5th source is audio chat coming in from Ventrillo/teamspeak, 6th source is through ventrillo/teamspeak (which ever source i decide) it all operates under my computer. Might sound like a crazy cluster fk of audio but actually the audios streamed through bluetooth headset, and i have different audio settings for everything so i hear it differently where as the viewer hears it at full blast. Difference? Tons of tweaking on audio sources, and editing to get good audio quality. End result, a nice finish product, but i usually dont upload my gameplay, more of a show it off to friends and to show my kickass moments ingame. I usually use 2tb and 4tb hard drives that clock at 6400rpm, 5400rpm, or a SSD 256gb strictly for quick file video capture.
Hi, I've been trying to record with Dxtory (since FRAPS gives too much impacts on my frames) with this guide: https://forums.station.sony.com/ps2...ing-in-planetside-2-dxtory-ffsplit-obs.80232/ I'm using the x264 codecs, and followed the guide exactly. The issue is that the recorded files are just audio with no video; a black screen - I googled the issue and a few people have had it but, like me, don't have a solution. If the problem persists I'll try out bandicam - does anyone have a ball park figure how much it affects your frames compared to Dxtory and FRAPS?
Are you sure you have the correct audio input selected? It should be whatever your playback device is for PS2, most likely your default playback device in Windows Sound options. Here's mine For audio codec and format I use PCM 44100 hz 16 bit Stereo, and encode it to AAC once I export to h264 .mp4. I record with Lagarith Lossless codec, it's better quality than x264. Also, are you sure the problem is with the video file not containing audio, rather than being a playback issue? Try extracting the audio track and playing it on its own (right click on file -> Extract Audio Stream).
