So you can see the “STAT UNITGRAPH” with “STAT RAW” applied in this video. Each one of them spikes is a really bad stutter (especially anything that takes over double the time the previous frame took to render). The general instability is what results in “micro stutter” and why many of us feel that 120fps feels like 60fps. During this video you will also notice times where the framerate goes completely crazy for around 3 - 6 seconds at a time with hugely varied frame times. You will also notice times where regular spikes appear… Personally the game feels almost unplayable to me and its just really starting to frustrate me now.
Example of the horrendous frame rate stability...
Tast1c
(Bangtastic)
#2
No matter how fluent the game is the spikes are still there. especially for example third part of whitechapel the spikes are real. Graphically it shouldnt be lot demanding, but somehow it is. I still hope for some performance fixes when it hits open beta, otherwise ppl with older rigs will open like 10k new threads^^
But it isnt only up to power to run the game apparently, because pixel got more than a decent pc^^
When I played Blacklight back in beta, the game was as laggy or even laggier than DB, back then I didnt played it further cause it was also unplayable for me. I dont know if they have ever fixed that.
prophett
(prophett)
#5
[QUOTE=Tast1c;517782]
When I played Blacklight back in beta, the game was as laggy or even laggier than DB, back then I didnt played it further cause it was also unplayable for me. I dont know if they have ever fixed that.[/QUOTE]
It improved since the beta but up until a year ago it was never fully resolved.
BomBaKlaK
(BomBaKlaK)
#8
Yep but optimization need the game to be build first, then optimize.
I don’t wait for a super optimize game yet but something a bit beter can be nice, I mean with more stable framerate.
ailmanki
(ailmanki)
#9
[QUOTE=BomBaKlaK;517794]Yep but optimization need the game to be build first, then optimize.
I don’t wait for a super optimize game yet but something a bit better can be nice, I mean with more stable framerate.[/QUOTE]
Yes optimizing to early isn’t good, but they should at least get the game to run fluid, else echo data and feedback will be flawed.
PixelTwitch
(PixelTwitch)
#10
well honestly, if Echo takes average fps data (like I would assume) over a certain amount of time then they will likely not even see these spikes in their data.
JBRAA
#11
In Dirty Bomb, are the spikes also related to or effected by FPS fluxuations?
In practical terms: Are spikes still there if you cap FPS to 30, 60, 90, 120, 144, 150, 200, etc? (I know its not Q3 engine. Im just using those values, as examples, for values to test. 30 example, here you should absolutely be able to hold 30FPS rock solid, and have zero fluxuations in FPS. Then to see how the spikes act.
If then, then spikes calm down, with rock solid FPS, then maybe, try set max FPS to a value which is high enough, but low enough to keep FPS rock rolid, and spikes down to zero.) But its just an idea. I dont actually know if they are related in DB, or in any other game, or if spikes and FPS-fluxuations are effected by each other.
If you do the test, thanks in advance. 
Good job, btw with making this topic, video, and finding the cvar for this graph.
In CS:GO, on all highest, except using 2xAA and 4xAF, I get perhaps 80-200+FPS, mostly around 140-170FPS. But this feels bad for my eyes. If in CS:GO use fps_max “120” it instantly feels so much better, due to no fluxuations.
I dont know if this is related to DB, since I do know fluxuations and spikes are different, but I do also know flux can feel very bad as well.
In COD:BO2 fluxuations 150-200 FPS doesnt feel bad at all for some reason. Maybe because it has zero or very low level of spikes.
COD:BO2 max 120 FPS vs max 200 (150-200 due to my systems performance), is very very similar in what my eyes notice in terms of stutter & smoothness / frametime ms.
If it matters:
I use 4690K @ 4,2, 660 ti, 120hz, win7x64, 32GB RAM.
If it matters part2:
Im very sensitive to hz, ms etc. I tested 144hz vs 120hz, vs 120hz@100hz, and I could notice the difference (144hz feels best of course.)
Another similar topic to FPS-spikes, is network-spikes.
But when looking at your video, it does not look like network-spikes
Network-spikes would make mini-freezes on screen, with lost packets & visual frames, if Im not mistaken.
onYn
(onYn)
#12
Shouldn´t really care about performance if we are not close to a release date (and I hope we aren´t). Otherwise it´s just going to be wasted work time.
PixelTwitch
(PixelTwitch)
#15
yea I was right mate… you dont have Stat Raw activated. You can tell by the curve on the initial spike…
JBRAA
#16
Kinda off topic:
PixelTwich: Request: Wanna test if DB has server side vs client sides problems with hitboxes like CS:GO has with jumping for example? https://www.youtube.com/watch?v=H1lblTOV1mg
PixelTwitch
(PixelTwitch)
#17
JBRAA, Dirty Bomb uses Client Side hit boxes so if you shoot them on your screen you do damage almost 100% of the time (there are a couple of glitches).
However, that said, I am not saying that the client side hitboxes always match up with the animations. I do believe there are some problems there but we are locked out of the commands to do testing.
JBRAA
#18
[QUOTE=PixelTwitch;517855]JBRAA, Dirty Bomb uses Client Side hit boxes so if you shoot them on your screen you do damage almost 100% of the time (there are a couple of glitches).
However, that said, I am not saying that the client side hitboxes always match up with the animations. I do believe there are some problems there but we are locked out of the commands to do testing.[/QUOTE]
Ah ok. Thanks for the info.
Oh how unfortunate. It would be wonderful to see how good or bad animations actually match up to hitboxes. But if I remember correctly DB doesnt use hitboxes? Or atleast not classic ones?
Here is some old stuff from Wolf ET, which was golden info back then.
http://denarii.info/ethitboxes/
(like this http://denarii.info/ethitboxes/strafe_left_nade2.jpg )
http://users.kymp.net/feuer/etcomic/056.jpg
