安装 Steam
登录
|
语言
繁體中文（繁体中文）
日本語（日语）
한국어（韩语）
ไทย（泰语）
български（保加利亚语）
Čeština（捷克语）
Dansk（丹麦语）
Deutsch（德语）
English（英语）
Español-España（西班牙语 - 西班牙）
Español - Latinoamérica（西班牙语 - 拉丁美洲）
Ελληνικά（希腊语）
Français（法语）
Italiano（意大利语）
Bahasa Indonesia（印度尼西亚语）
Magyar（匈牙利语）
Nederlands（荷兰语）
Norsk（挪威语）
Polski（波兰语）
Português（葡萄牙语 - 葡萄牙）
Português-Brasil（葡萄牙语 - 巴西）
Română（罗马尼亚语）
Русский（俄语）
Suomi（芬兰语）
Svenska（瑞典语）
Türkçe（土耳其语）
Tiếng Việt（越南语）
Українська（乌克兰语）
报告翻译问题
讨论规则及指引
举报此帖
dont know why this guy has such a bad frametime but i will show you how this looks like on my "top pc". frametime is in every game a strait line without any stutter at all and this with always maxed out gpu !
edit: uploading some random games i play
https://youtu.be/94QbgqdV10g
And yeah, most 2070 reviews I've seen can't get clocks that high, but yours can. I guess you're just lucky, as your exact 2070 shows awesome results.
You don't expect the difference between 4ms and 7ms to be clearly visible on the default 0-100 graph, do you? You frametime graph is nothing like a straight line, all you need is to either reduce the graph limits or just use simple maths.
Let's analyze some measurements from the very beginning of your video.
1st measure - 190fps at 4.4ms frametime, should've been, 5.3ms.
2nd measure - 175fps at 5.0ms, should've been 5.7ms.
3rd measure - 168 fps at 5.4ms, should've been 6ms.
4th measure - 171fps at 5.7, should've been 5.8, so far the closest to what you should get.
5h measure - 169 at 5.8, 5.9ms.
6th - 164 at 6.1, this one is actually good.
7th - 162 at 4.7 vs 6.2, and now it's really bad.
8th - 171 at 6.5 and not 5.8, this time frame took longer than it should - stutter and added input latency.
9th - still 171fps but 6.4ms frametime, once again stutter and extra input latency.
10th - 166 at 6.7 vs correct 6.0, same old crap.
So, at measurements 1-5 and 7 we see PC compensating for late frames by pushing some frames faster than it should, 5 and 6 are close to perfect, and at measurements 8-10 we see PC making frames too late. Those frametimes are absolutely not right, but there are two things to note here.
First - you've got around 170 frames per second for the most part during those measurements, and it took 4 seconds to get 10 measurements. We only know the results of 10 separate frames, while 4 seconds at 170fps is almost 700 frames. We can possibly not truly analyse this video unless we analyse every and single frame, but that takes too much time and some kind of log analysing software that will do this job way faster than we would, hopefully before we get too old to stop seeing numbers on the screen. Still we can get the general idea of what frame inconsistency is - we've got some fine frames, we've got some late frames, and we've got some early frames compensating for those late ones.
Second - yet again, around 170 frames per second. During the measurements 2-10 (I exclude 1 because it's not close to 170) the highest difference between frametimes you had was 1.7ms. You're compensating for frametime inconsistency by pushing more frames, and it works indeed - I doubt I personally would even notice the 1.7ms difference in input lantecy. That said, if you're running some kind of VSync (GSync and FreeSync are also just VSync variations) - each time a frame takes longer than it should, front buffer doesn't get a new image and screen shows old image once again, resulting in stutter. Higher framerates and higher refresh rates make stutters and input latency inconsistency less and less noticeable, so when it comes close to 200fps - well you might get awesome results with graphics card at 99%. But difference between 5.0ms and 6.7 is still 34% difference. If you would be getting, say, around 60fps like I often do - it's as good as difference between 16.67ms and 22.34ms. Now THAT kind of difference would be clear, as it's almost 6ms input latency difference and it will look like it's running at 45fps.
tl;dr - while you don't get consistent frametimes, at framerate that high you're unlikely to tell the difference anyway.