You get 10,000 people to reply that thread and you might have a case for "common".
There are a couple of million people on their forums?
No, but if you want to play that game, there are over 100,000. 139 is not common.
I was talking about game owners.
Their forums are there for a decade.. and they are for all of Eidos' games.. There aren't 100k people registered to tall about Deus Ex and not all of them own it on PC.
100+ people in less than 24 hours is common enough to be considered an issue to investigate because chances are you'll be able to reproduce it.
i've beaten this game a couple of days ago. it took several days playing off and on. i'm about 10h into my 2nd playthrough.
the graphic is excellent with dx11 and 2560x1600 resolution support. although, they could work on the character faces some more even though they're shooting for stylized not realistic look.
the gui/hud scales correctly to 16:10 as well as allowing us to tweak the field of view (fov of 80 is just perfect for 16:10).
i love the grid-based inventory, hacking game, conversation game, weapon modding, and multiple outcome of your decision. all this makes for a great experience.
perfect xb360 pad support rounds out an excellent package. i love the quick inventory select radial wheel. it makes using pad to select items that much easier.
i wished we could go back and forth between levels instead of waiting for the story to decide. forced killing bosses instead of non-lethal method of takedown is another thing i wished they did not implement.
i would like to see more breakable or more interactive environments. right now, the environment looks very static even though there are a lot of trash about. it kills immersion when you shoot at something and it doesn't destroy it.
looks like i'll be replaying this game for some time.
It could, potentially, be a graphics setting - as I haven't tested thoroughly for each setting, but I don't think it is. Since by turning off VSYNC - the issue goes away entirely, I doubt it has anything to do with FPS drops or demanding settings. I think it could have something to do with how movement works with keyboard as opposed to a game controller, as console versions tend to implement "gradient" speeds for movement. If they've not fully taken that into account with the PC version, that could be one reason.
Micro stuttering is not a drop in FPS. It's a rise in render time. A short example: 60 FPS = 16.67 ms per frame. Obviously, not every frame will render in exactly 16.67 ms, but on average the time required to render a frame varies by +-10%. Micro stuttering occurs when a *single* frame takes like 20ms all of a sudden. That can still mean that it's 60 FPS overall, but that one slow frame is what you will notice. Once per second is something most people overlook, but have it happen twice in a second and it's annoying.
Also the reason why FPS as a measure is pointless.
Casus wrote:
The thing to look for, is how the movement "skips" a little, which is most evident if you focus on the car in the video. I realise this isn't the ideal demonstration - but at least it's the right issue, and not the "stuttering" most people seem to talk about.
I'll check this myself later, but it looks more like input lag than anything else - that could indicate a problem with scheduling. Not noticed this either though.
There are a couple of million people on their forums?
No, but if you want to play that game, there are over 100,000. 139 is not common.
I was talking about game owners.
Their forums are there for a decade.. and they are for all of Eidos' games.. There aren't 100k people registered to tall about Deus Ex and not all of them own it on PC.
100+ people in less than 24 hours is common enough to be considered an issue to investigate because chances are you'll be able to reproduce it.
No but we know that the PC version sold significantly more than 139 copies, and with internet coverage being what it is these days, if it was serious they'd be there. Given the fact that this kind of poll produces biased results (which is why no serious statistician would ever use them) 139 is a biased high number to begin with.
It's not remotely indicative that this is a "common" problem.
As it is there are 186,000 posts in the Deus EX forums. That's a non-trivial amount, and you're probably looking at 20,000 or more users being responsible for that. While some posters are prolific, a lot of people will only post a couple of times.
Micro stuttering is not a drop in FPS. It's a rise in render time. A short example: 60 FPS = 16.67 ms per frame. Obviously, not every frame will render in exactly 16.67 ms, but on average the time required to render a frame varies by +-10%. Micro stuttering occurs when a *single* frame takes like 20ms all of a sudden. That can still mean that it's 60 FPS overall, but that one slow frame is what you will notice. Once per second is something most people overlook, but have it happen twice in a second and it's annoying.
Also the reason why FPS as a measure is pointless.
Actually, you're not quite right.
A stutter IS an FPS drop - because the average frame rate will inevitably be lower when the "render time" is increased - even for just a single frame. However, without an established frame rate or let's call it "performance under non-stuttering circumstances" the concept of an FPS drop is a bit abstract.
The reason behind the render time increase is, of course, what's interesting - and in the case of stuttering, it's most often about how assets are loaded. Streaming games tend to suffer much more from this issue, obviously, since they dynamically load content.
That said, you're right - FPS is not always the most useful measurement of performance, and it's most definitely not the ultimate way to determine how a game truly behaves.
A stutter IS an FPS drop - because the average frame rate will inevitably be lower when the "render time" is increased - even for just a single frame.
FPS = Frames per second = Average framerate
FPS will be the same if, for example 60 frames each take 16.6ms, or if 59 frames take 8.3ms, but every 60th frame takes 500ms. This is an extreme example of course, in real world the difference is much smaller, but often noticeable. It can be caused by systems that don't update every single frame, like physics or AI, although there could be many other reasons for it.
In any case I don't think this is what Deus Ex players are experiencing.
A stutter IS an FPS drop - because the average frame rate will inevitably be lower when the "render time" is increased. However, without an established frame rate or let's call it "performance under non-stuttering circumstances" the concept of an FPS drop is a bit abstract.
No, because you can still hit the exact same framerate over a period of time depending on how long that period is (with FPS, one second). Even with these increased render times, that is exactly how games running in SLI/CFX at 120 FPS can still lag, just because a few frames in between were slower. Here's an example with numbers (render time in ms) for a target 10 FPS (heavily exaggerated, but that's not the point):
- Ideal: 100 100 100 100 100 100 100 100 100 100 = 1000ms
- Reality: 90 100 110 90 90 90 90 110 110 100 = 980ms
- Microstuttering: 100 90 120 90 100 90 90 100 100 100= 980ms
Now when measuring #2 and #3, that will still result in 10 FPS, because that 11th frame will not be counted. Therefor, FPS or framerate are both completely pointless metrics. Average render time (and corresponding min/max) would be a far better method
Don't look at micro stuttering as "a stutter", it's a phenomenon that's quite unique in its own. This is also never, ever caused by loading assets or something similar, as hitches introduced by that will be bigger than "micro". With SLI/CFX they're caused by the two GPUs being out of sync and I suspect that this is related to the game trying to match input to display.
And that just gave me an idea. What kind of mouse do you have? This is an easy one to test for me (assuming I'll have some time tonight), as I can simply turn the DPI on my G500 way down so that it's inaccurate as hell. That'll make it harder for the game to match its input up to the frames if that's the problem; although I must also say that my poll rate is set to 1000 Hz (default in Windows is 125 Hz I believe), again something I can very easily adjust in Logitech Setpoint.
A stutter IS an FPS drop - because the average frame rate will inevitably be lower when the "render time" is increased - even for just a single frame.
FPS = Frames per second = Average framerate
FPS will be the same if, for example 60 frames each take 16.6ms, or if 59 frames take 8.3ms, but every 60th frame takes 500ms. This is an extreme example of course, in real world the difference is much smaller, but often noticeable. It can be caused by systems that don't update every single frame, like physics or AI, although there could be many other reasons for it.
It's not an "FPS drop" in itself when a game stutters. But it's an "FPS drop" compared to the game not stuttering.
To put it another way:
The game doesn't stutter for a while, and you run around with an average of 60 FPS. Suddenly, you enter a room and the game starts to load new assets from the slow harddrive, causing some frames to have an increased render time because they can't render what's not available. THAT will be an "FPS drop" compared to what it was before you entered the room - all things being equal. So, if you measure gameplay for a similar amount of time with and without stutter - you'll se an "FPS drop".
True it is semantic issue. Werelds is talking about micro-stuttering which is not related to data streaming. Streaming usually causes proper stuttering over multiple frames that does indeed cause a FPS drop.
True it is semantic issue. Werelds is talking about micro-stuttering which is not related to data streaming. Streaming usually causes proper stuttering over multiple frames that does indeed cause a FPS drop.
Which is what I experience: streaming+stuttering+FPS drop for 1-2 seconds.
No, because you can still hit the exact same framerate over a period of time depending on how long that period is (with FPS, one second). Even with these increased render times, that is exactly how games running in SLI/CFX at 120 FPS can still lag, just because a few frames in between were slower. Here's an example with numbers (render time in ms) for a target 10 FPS (heavily exaggerated, but that's not the point):
- Ideal: 100 100 100 100 100 100 100 100 100 100 = 1000ms
- Reality: 90 100 110 90 90 90 90 110 110 100 = 980ms
- Microstuttering: 100 90 120 90 100 90 90 100 100 100= 980ms
Let me put it another way:
Logically, it's 100% impossible to reach the same FPS in the SAME period of time if the game that has just 1 frame with increased render time due to stuttering (not microstuttering - whatever that means). Mathematically, that will result in a LITERAL FPS drop over that period of time. All other things being equal, of course. That last part is essential, obviously - because it's quite plain that a similar FPS can theoretically be reached if the non-stutter frame renders are correspondingly faster.
Unless we're talking about two different things - which I think we are. I don't even know what "microstuttering" is
Quote:
Don't look at micro stuttering as "a stutter", it's a phenomenon that's quite unique in its own. This is also never, ever caused by loading assets or something similar, as hitches introduced by that will be bigger than "micro". With SLI/CFX they're caused by the two GPUs being out of sync and I suspect that this is related to the game trying to match input to display.
I wasn't aware of this "micro stuttering" having an actual objective definition. Do you have a source for this? I'm asking, because I'm not sure we're talking about the same thing.
Also, I think it's becoming about something completely different - and it smells like a "who has the bigger knowledge" dick contest.
Not really the kind of thing I'm interested in, and I'm quite ready to concede you know more about the technical stuff than I do. It's not really my strong point.
Quote:
And that just gave me an idea. What kind of mouse do you have? This is an easy one to test for me (assuming I'll have some time tonight), as I can simply turn the DPI on my G500 way down so that it's inaccurate as hell. That'll make it harder for the game to match its input up to the frames if that's the problem; although I must also say that my poll rate is set to 1000 Hz (default in Windows is 125 Hz I believe), again something I can very easily adjust in Logitech Setpoint.
I have a Razer Naga - but I'm pretty sure it's not input lag, as it happens even without mouse input.
Logically, it's 100% impossible to reach the same FPS in the SAME period of time if the game that has just 1 frame with increased render time due to stuttering (not microstuttering - whatever that means). Mathematically, that will result in a LITERAL FPS drop over that period of time. All other things being equal, of course.
Unless we're talking about two different things - which I think we are. I don't even know what "microstuttering" is
That's in an ideal situation. In reality, no two seconds will have the same average render time at all. So 10 FPS from 0->1 can be 10x100, but like Bearish points out so subtle (), the next second it might very well be 1x910 + 9x10. This stuff doesn't give a rat's ass about mathematical consistency. Hence my statement that FPS is just a worthless metric as the reality is that every second is completely different from the one before, and the one after it
I'll see if I can find a good articles that explains micro stuttering in depth, but yes, it does have a very objective definition. That is because with multi GPU setups it's just a very real problem unfortunately. The main thing to realise is that with micro stuttering you can experience "jittering", "stuttering", "hitching" or whatever you call it while still maintaining a constant FPS overall. Once you got that, you'll see how it's a much more annoying problem than FPS drops (where you get less frames overall in the same period of time).
This is not a dick showoff though, I'm quite serious. That video displayed something that is closer to microstuttering than FPS drops. As you're describing it as a VSync problem, the first thing that's synced is the input (which is why input lag matters), audio would be next.
Well it seems win7 thinks I have a 3800 card and win7 x64 knows I have a 5800 card so that's why dx11 didn't work, well it says 3800 in the deus ex menu but 5800 in the ati menu, weird, 11.8 drivers on both windows.
That's in an ideal situation. In reality, no two seconds will have the same average render time at all. So 10 FPS from 0->1 can be 10x100, but like Bearish points out so subtle (), the next second it might very well be 1x910 + 9x10. This stuff doesn't give a rat's ass about mathematical consistency. Hence my statement that FPS is just a worthless metric as the reality is that every second is completely different from the one before, and the one after it
I don't really feel like retreading the same thing over and over. I'm not talking about the ideal situation as a realistic one, but for purposes of explanation. I will remain quite firm that a "traditional" stutter will result in a very real FPS drop overall, even if the FPS drop in itself isn't the problem.
But enough about that, as the problem I'm talking about hasn't got anything to do with stuttering. At least, not the kind of stuttering I'm familiar with.
Quote:
I'll see if I can find a good articles that explains micro stuttering in depth, but yes, it does have a very objective definition. That is because with multi GPU setups it's just a very real problem unfortunately. The main thing to realise is that with micro stuttering you can experience "jittering", "stuttering", "hitching" or whatever you call it while still maintaining a constant FPS overall. Once you got that, you'll see how it's a much more annoying problem than FPS drops (where you get less frames overall in the same period of time).
I looked it up myself
As you say, it's specific to multi-GPU setups, so while it's interesting enough - I'm not sure why you're bringing it up in response to traditional stuttering. I haven't used the term micro stuttering once when I've made my point.
But, just to be clear, "my" issue isn't a multi-GPU one - as I don't have a multi-GPU video card, and I've tested it on several setups. I'm not sure if any of them had multi-GPU cards, but I've tested both Nvidia and ATI cards, with a multitude of drivers and settings. These include 5870 and GTX260 - which are my own cards. Not sure what cards my friends have - but they're probably in the same ballpark.
My conclusion is that it's related to VSYNC with almost absolute certainty, but I'm not sure of the technical details.
Quote:
This is not a dick showoff though, I'm quite serious. That video displayed something that is closer to microstuttering than FPS drops. As you're describing it as a VSync problem, the first thing that's synced is the input (which is why input lag matters), audio would be next.
I don't know if the guy in the video has a multi-GPU setup, but as it's identical to my single-GPU issue - I think we can rule out "microstuttering".
About input - I'm not sure how the keyboard input gets synced - but my assumption is that "key pressed down" is registered and unless that changes, I don't see why it would interrupt the rendering process. The problem persists with just one key pressed down indefinitely.
AFAIK, VSYNC is related to rendering only - and specifically it's about stalling the video card until the monitor has refreshed its cycle.
I'm not sure why audio would even interfere with that, but since you're the techie - I'll take your word for it. I'll look into that possibility.
No, the engine has to match up the audio and your input to what's going on. At 120 FPS it obviously has to poll your mouse more to get a reliable input. Same for the keyboard; correct me if I'm wrong, but just strafing doesn't reproduce it right? It's the combination of strafing and turning? That's why I'm pointing at the mouse input; if that's just a few ms out of sync that fucks up mouse input, something plenty of games suffer from and the reason why almost everyone I know has their USB polling rate set to 500 or even 1000 Hz, that way there's always fresh data available. Your input is part of what it has to render
That's also why TFT monitors were so hated for gaming for so long: the input lag was more than 1 frame (i.e. your mouse input would come through 30+ms after you actually doing it).
And no, micro stuttering isn't specific to multi GPU setups; to describe it correctly it's a syncing issue (and with AFR, a multi GPU setup doesn't sync, hence the problem), it's just most apparent and most widely known with multi GPU setups. But let's drop that for now
No, the engine has to match up the audio and your input to what's going on. At 120 FPS it obviously has to poll your mouse more to get a reliable input. Same for the keyboard; correct me if I'm wrong, but just strafing doesn't reproduce it right? It's the combination of strafing and turning? That's why I'm pointing at the mouse input; if that's just a few ms out of sync that fucks up mouse input, something plenty of games suffer from and the reason why almost everyone I know has their USB polling rate set to 500 or even 1000 Hz, that way there's always fresh data available. Your input is part of what it has to render
That's also why TFT monitors were so hated for gaming for so long: the input lag was more than 1 frame (i.e. your mouse input would come through 30+ms after you actually doing it).
And no, micro stuttering isn't specific to multi GPU setups; to describe it correctly it's a syncing issue (and with AFR, a multi GPU setup doesn't sync, hence the problem), it's just most apparent and most widely known with multi GPU setups. But let's drop that for now
I realise that audio, input, and video need to match up
What I didn't realise is that audio syncing is related to VSYNC. I assume that's what you're saying, as the issue only exists with VSYNC enabled.
Yes, it occurs when only strafing - and also when only moving forward. It has nothing to do with mouse input - unless just having the mouse plugged in counts. However, since I've also tested the Xbox 360 version - I think we can rule out mouse input. The problem is there also, only much less pronounced. I doubt most people even notice it there.
However, that doesn't mean that input in itself doesn't count - and I'm suddenly uncertain if I actually tested thoroughly without touching the mouse. But if it IS a mouse issue, it's one that the 4 setups I've tested all share - and we all have different mice.
I uninstalled the game - but I'll test it on my replay. It would be interesting to know what causes it exactly, regardless of whether it's universal or not.
TWIN PEAKS is "something of a miracle."
"...like nothing else on television."
"a phenomenon."
"A tangled tale of sex, violence, power, junk food..."
"Like Nothing On Earth"
And I'm done, 77 hours according to Steam. Amazing game, sure it has some problems, but who would have expected even this these days. As I said before, this is my GOTY and nothing that will get released later this year can top it for me. There could have been a chance to be topped if Risen 2 would be released this year, but that's not the case.
NZXTS340ELITE : EVGA Z370 FTW / [ Intel i7 8086k @4.0Ghz ][ ASUS TUF RTX 3060 Ti 8GB ][ 16GB G.Skill Trident Z @3200mhz CL16 ][ 128GB Intel760p Series + 1TB Crucial MX500 + 3TB WD RED ][ Thermaltake Toughpower PF1 650W ]
Signature/Avatar nuking: none (can be changed in your profile)
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
although, they could work on the character faces some more even though they're shooting for stylized not realistic look. 
