Can you go into your config, change "FrameTimeDisplay" to 1 and record some taiko gameplay so that we can analyze further? Please also make sure the shown display looks the same when you're recording and when you are not.
I don't have that line in my .cfg. Should I add it?Tom94 wrote: 4j5f55
Can you go into your config, change "FrameTimeDisplay" to 1 and record some taiko gameplay so that we can analyze further? Please also make sure the shown display looks the same when you're recording and when you are not.
Seeing from one of your posts, you've got an Nvidia GTX 970, which has NVENC. This thing can easily record up to 4K videos at 60 Hz using the H.264 codec (MP4) with an extremely low impact (pretty much non-existent) on your game, since it runs in a dedicated chip on your GPU.lolcubes wrote: u93m
I don't have that line in my .cfg. Should I add it?Tom94 wrote: 4j5f55
Can you go into your config, change "FrameTimeDisplay" to 1 and record some taiko gameplay so that we can analyze further? Please also make sure the shown display looks the same when you're recording and when you are not.
Also, mind giving some advice on what recording method to use? Pretty much all recording applications which capture video that I used either don't capture 60 fps correctly or have A/V desync.
My phone has a shitty 30fps camera (I'll try using that for now, the stutters could be visible).
Also there is one more issue and that is that gameplay with vsync is MUCH more smooth than watching a replay under vsync (however this is also present in live). The replay is unwatchable due to insane stuttering, while gameplay is playable with occasional stutter and some input lag.
Shadowplay doesn't work with OpenGL yet and NVENC codec actually has issues with vfps, which leads to A/V desync. So far not a single recording I produced in this way was proper (after re-encoding to cfps), and I was even using a relatively fast SSD to record on.Nabile wrote: 2548t
Seeing from one of your posts, you've got an Nvidia GTX 970, which most certainly has NVENC. This thing can easily record 1080p videos at 60 Hz using the H.264 codec (MP4) with an extremely low impact (pretty much non-existent) on your game, since it runs in a dedicated chip on your GPU.
You can use this encoder with either Open Broadcaster Software or ShadowPlay (comes with GeForce Experience).
Thanks for your ! The delay you are feeling when in windowed mode is normal and caused by windows. It also exists on stable.abraker wrote: 6cx2d
Issues presented in this thread are still existent on b20150521cuttingedge, especially the crashing when alt-tabbing on fullscreen.
Oops, intel HD 4000 graphics mobile. That implies that I am using a laptop. Also these recordings are 30 FPS and were made without the HDMI out to the TV this time.Tom94 wrote: 4j5f55
Can you let me know whether you are using a specific device or whether you just have the intel HD4000 graphics on your desktop machine?
I was playing on unlimited (gameplay). My SwapBuffer still consumes a considerate amount, just not as much. You can actually see how much at the beginning of the song select idling video. Yes this is the latest cutting edge version. I went from stable to cutting edge (20150521.1) just to record these videos. It might look that way because I disabled triangles (too laggy).Tom94 wrote: 4j5f55
Just to make sure: You are not running VSync, right? Does your SwapBuffer still consume that much of your frame times while you are not recording? And finally, are you sure you are on the latest cuttingedge version? Your graph looks a bit like it isn't from the very newest version.
Thanks for all the info! The time spent in SwapBuffer is mostly waiting for all the drawing instructions to finish executing on the GPU. We are looking into how we could improve the performance here.abraker wrote: 6cx2d
I was playing on unlimited (gameplay). My SwapBuffer still consumes a considerate amount, just not as much. You can actually see how much at the beginning of the song select idling video. Yes this is the latest cutting edge version. I went from stable to cutting edge (20150521.1) just to record these videos. It might look that way because I disabled triangles (too laggy).Tom94 wrote: 4j5f55
Just to make sure: You are not running VSync, right? Does your SwapBuffer still consume that much of your frame times while you are not recording? And finally, are you sure you are on the latest cuttingedge version? Your graph looks a bit like it isn't from the very newest version.
bigfeh wrote: 44154m
This apparently is cuttingedge-specific (the same message would show up when sent to a stable client)
When sending a message whose first non-space character is a forward slash (e.g. " /test", minus quotes - note the space before the slash), it will only show up on the client of the sender but not on the client of the recipient.
@edit
sending the message without the space obviously counts it as a command instead, so it doesn't show up in either client. Just figured I'd clear this up
Pics might help explainSender
Recipient![]()
what's the performance like at menus? is the drop just during gameplay? does it improve by turning off snaking sliders?Yomura wrote: 36d16
This is NOT playable at all
OpenGL cuttingedge running on a 2007 machine (HP tx1000) with OpenGL 2.1
https://osu-ppy-sh.jeuxcrack.net/ss/3164377
also is buggy as hell as you can see
Getting ~50-70fps when with DirectX fullscreen is around 200fps
As you can see that's not fullscreen because fullscreen is directly 10fps in the face.
I always have snaking sliders off. At F9 screen and main menu I get around 9~20fps as you can see here:peppy wrote: 6w4l42
what's the performance like at menus? is the drop just during gameplay? does it improve by turning off snaking sliders?Yomura wrote: 36d16
This is NOT playable at all
OpenGL cuttingedge running on a 2007 machine (HP tx1000) with OpenGL 2.1
https://osu-ppy-sh.jeuxcrack.net/ss/3164377
also is buggy as hell as you can see
Getting ~50-70fps when with DirectX fullscreen is around 200fps
As you can see that's not fullscreen because fullscreen is directly 10fps in the face.
You can either keep the shortcut in your desktop or close the update window whenever you start the game for the time being (it doesn't matter if you close the update window since osu! can update itself in-game if an update is available).Renevant wrote: 3e2y38
Im also getting this.VeilStar wrote: 4fbc
A shortcut to osu!.exe gets created on your desktop every time you launch osu!.
Happens since today(?).
I was so confused thinking i accidentally created a shortcut for it. Then it kept happening.
Please test taiko again on the latest release (...25.3)lolcubes wrote: u93m
Okay, as far as FPS goes, it's much more stable now, but there are still issues with Auto or spectator mode in Taiko when under 300ish FPS. It's difficult to explain, but every time a note gets hit a stutter appears, however it's not ed in the frame time window (doesn't happen to me above 300 FPS). Could it be that the replays just work that way? Is it possible to make it smoother?
Tested with ...25.4, it's the same. Tonight I will record a video when I get back from work (I did a test video on my phone, it's difficult to see but when I switch from 120 to 1200 fps, it becomes more clear).peppy wrote: 6w4l42
Please test taiko again on the latest release (...25.3)lolcubes wrote: u93m
Okay, as far as FPS goes, it's much more stable now, but there are still issues with Auto or spectator mode in Taiko when under 300ish FPS. It's difficult to explain, but every time a note gets hit a stutter appears, however it's not ed in the frame time window (doesn't happen to me above 300 FPS). Could it be that the replays just work that way? Is it possible to make it smoother?
Or that's what I thought. Apparently osu! doesn't save slider adjustments made with arrow keys.Pampel wrote: 4m5tc
I'm on windows 7 using default mouse sensitivity.Tom94 wrote: 4j5f55
On which operating system are you? Do you by any chance use a custom mouse sensitivity?
This is expected, yes. Full-screen is getting exclusive access to your display and bying the windows display system.Shizuyi wrote: 6f5u28
My FPS stay the same but i noticed it feels a bit more smooth since the latest patch! i like it.BUt another thing: Is it normal that the play feeling between borderless and Fullscreen is SOO much different? It feels weird. :S
you're seeing the same issue on both cutting edge and stable? only when using opengl?_index wrote: 4p2k47
Experiencing random slight lag, according to Ctrl+F11, drawcall spike once per 5-9 seconds, mes any input (either aim-wise or key-wise). Pretty damn annoying.
Latest cutting edge/stable, opengl fullscreen, AMD FX-8150 w/ GTX 970 on a windows 8.1
Is it on my end or might be fixable?
yes, on bothpeppy wrote: 6w4l42
you're seeing the same issue on both cutting edge and stable? only when using opengl?_index wrote: 4p2k47
Experiencing random slight lag, according to Ctrl+F11, drawcall spike once per 5-9 seconds, mes any input (either aim-wise or key-wise). Pretty damn annoying.
Latest cutting edge/stable, opengl fullscreen, AMD FX-8150 w/ GTX 970 on a windows 8.1
Is it on my end or might be fixable?
Looks better, 1.83-> 1.74ms with the same benchmark as before. Most of the random frame time spikes are gone. worst 0.1% from 4.19 to 2.97 mspeppy wrote: 6w4l42
New build up with some heavy performance tweaking. Please report back if you previously had issues.
directX overall lags way more than opengl, because of how much its depended on a single core performance, rendering 80-100% u1 load and framelaggingpeppy wrote: 6w4l42
same spikes are not present in direct x?_index wrote: 4p2k47
yes, on both
frame limiter doesnt help, however issue seems to be totally random, sometimes appearing for a few hours and sometimes not. u load is usually at 70% one core and 20-30% other cores, while gpu stays about 50-60%
Graphics: [Tom94] Improve behavior of resetting the screen resolution to desktop settings. (fixes /t/259747)Reference post: p/4158960
Yea comfirmed.Fr33z3R wrote: 6w5243
the issue
Do you by any chance have GX mode enabled? Could you describe a bit more exactly what kind of popping you experience? I also have an STX and do not get any such things.Bmxant wrote: 3a2i8
I'm having issues with occasional sound popping. Issues only occurs in Osu!, no other game or sound file. It's random, but very noticeable and annoying.
Using a Xonar STX
Can reproduce the problem using both the Beta and Cutting-edge build
I don't seem to have the issue using the Stable client (so far)
Thanks for the reply, Tom94.Tom94 wrote: 4j5f55
Do you by any chance have GX mode enabled? Could you describe a bit more exactly what kind of popping you experience? I also have an STX and do not get any such things.Bmxant wrote: 3a2i8
I'm having issues with occasional sound popping. Issues only occurs in Osu!, no other game or sound file. It's random, but very noticeable and annoying.
Using a Xonar STX
Can reproduce the problem using both the Beta and Cutting-edge build
I don't seem to have the issue using the Stable client (so far)
Thanks a lot for the detailed information! I hear what you mean and will try to reproduce and look into it when it find the time.Bmxant wrote: 3a2i8
Pop/distort sound is at 0:31
https://www.youtube.com/watch?v=hCV06IM ... e=youtu.be
Application: osu!.exe
Framework Version: v4.0.30319
Description: The process was terminated due to an unhandled exception.
Exception Info: System.AccessViolationException
Stack:
at OpenTK.Graphics.OpenGL.GL.GenBuffers(Int32, Int32 ByRef)
at #=qye0$5YzUohuFy0qK3Za7DiKyzeAlwodjqvDQZfm$HquzOWcW$YiBeK2gt5l4gn.#=qj5RzzukSvCEAK8mrB4nHsw==()
at #=qYXxbL8qIL6wewuGnq7zT2g==.#=qj5RzzukSvCEAK8mrB4nHsw==()
at #=qGDXvskXP5HpR5NXG0JB9iQDg1aJEao7jWVFxAxjte40=.#=qeUlHWXdBKhAVDyWeU2F5Vg==()
at #=qnkb9whIkP9kmVLKYGzx3fA==.#=qKMe5x3OHNZe6N44TOzJJkw==(System.String, Boolean, Boolean, Boolean)
at #=qnkb9whIkP9kmVLKYGzx3fA==.#=qNv9AY$nrGIhAPK9YlteP4w==(System.String[])
at #=qnkb9whIkP9kmVLKYGzx3fA==.#=q0WrRcV01CL0r5Ftd$XtIEw==()
at #=qzuZqONQlNterlD6omxQ0Yf$k9FHWxTwA19sWgYs9vj4=.#=qYSqKQhx$diQxrdfMePEwAQ==()
I looked into this issue and the new VBO code requires a higher version of OpenGl to run. Your graphics card needs to at least OpenGl 1.5 in order to run the current code implementation. We will need to think of an alternative for GMA / ATI cards if we want to keep for those models. Here is a link describing the compatibility issue.Winshley wrote: 5616h
I posted this few days ago, telling me to post the issue here instead.
The game was running fine 3 weeks ago, but went on hiatus for a while before returning again. I went through all the updates, showing osu! splash screen, only to get itself exploded. Below is the error message I got from Event Viewer:
Application: osu!.exe
Framework Version: v4.0.30319
Description: The process was terminated due to an unhandled exception.
Exception Info: System.AccessViolationException
Stack:
at OpenTK.Graphics.OpenGL.GL.GenBuffers(Int32, Int32 ByRef)
at #=qye0$5YzUohuFy0qK3Za7DiKyzeAlwodjqvDQZfm$HquzOWcW$YiBeK2gt5l4gn.#=qj5RzzukSvCEAK8mrB4nHsw==()
at #=qYXxbL8qIL6wewuGnq7zT2g==.#=qj5RzzukSvCEAK8mrB4nHsw==()
at #=qGDXvskXP5HpR5NXG0JB9iQDg1aJEao7jWVFxAxjte40=.#=qeUlHWXdBKhAVDyWeU2F5Vg==()
at #=qnkb9whIkP9kmVLKYGzx3fA==.#=qKMe5x3OHNZe6N44TOzJJkw==(System.String, Boolean, Boolean, Boolean)
at #=qnkb9whIkP9kmVLKYGzx3fA==.#=qNv9AY$nrGIhAPK9YlteP4w==(System.String[])
at #=qnkb9whIkP9kmVLKYGzx3fA==.#=q0WrRcV01CL0r5Ftd$XtIEw==()
at #=qzuZqONQlNterlD6omxQ0Yf$k9FHWxTwA19sWgYs9vj4=.#=qYSqKQhx$diQxrdfMePEwAQ==()
Makes sense.TheVileOne wrote: 4cy22
I looked into this issue and the new VBO code requires a higher version of OpenGl to run. Your graphics card needs to at least OpenGl 1.5 in order to run the current code implementation. We will need to think of an alternative for GMA / ATI cards if we want to keep for those models. Here is a link describing the compatibility issue.
smoogipooo wrote: 5z25a
@Winshley: Please test again![]()
System.ArgumentException: Version string portion was too short or too long.
at System.Version.VersionResult.SetFailure(ParseFailureKind failure, String argument)
at System.Version.TryParseVersion(String version, VersionResult& result)
at System.Version.Parse(String input)
at System.Version..ctor(String version)
at #=q2RGho6OA$Q3AoWh_rJwdls32VE_DUTHpVM8biS5r4Po=.#=qHKbzGg9kaFcq$RuZ$p_8dQ==()
at #=qKj0H3V8xQ79JF8kPW$lQug==.#=qoP7fGILcSyqozyfclHWrIQ==()
at #=qKj0H3V8xQ79JF8kPW$lQug==.#=qnIX3ASoyevvlxD3AyiwggQ==()
at #=qPGkRY7FPh2MJQZJeBrF$erptF5ECo1MIGWucSoBgfUA=.#=q2duPe9AdIZmY9_cLc5rY2g==()
at #=qjU1NY2$E$b30iCIRJpxceA==.#=q4jFVuRSoX3GtOoihu0UHuw==(String #=q1zPTpHvpacVtwEgORwqMIw==, Boolean #=qpSA6cyZ7qdVcwC5yf4x88INfhNqLSul0BwL5gHAkXAc=, Boolean #=qay2a5Y_f6jY8_7K36ruRfUiyieYRHoScqyjtuUdP7FA=, Boolean #=qJmm0ixjT64t4slNnd2vSVg==)