Feedback on 0.46.0

General discussion regarding the OpenMW project.
For technical support, please use the Support subforum.
User avatar
afritz1
Posts: 47
Joined: 05 Sep 2016, 01:18
Contact:

Feedback on 0.46.0

Post by afritz1 » 16 Jun 2020, 05:48

Hello everyone, good work on 0.46! I played it for a few minutes and it feels like a big change from 0.45. Shadows are really cool and I don't remember distant land being a thing with the view distance slider. Both immediately made the game feel more natural.

My biggest issue with 0.46 is that the UI is super blurry at low resolutions now and looks really bad. I reinstalled 0.45 to verify, and there's a significant difference on my 4k screen (Windows 10 by the way). Previously I was playing at 720p or 1080p because the UI was comfortably-sized. Now it seems like there's no good resolution! 1440p and above, the UI is too tiny, and 1080p or below, it's too blurry. So there isn't really a comfortable resolution now :/

Okay, I tested it some and it seems like 0.45 is not DPI-aware but 0.46 is, because when I run them both in windowed mode at 720p, the 0.46 window is way smaller. With my desktop DPI scaling, 2160p is effectively like 1234p or something, which is why the 0.45 UI looks so much bigger. It's possible that you guys didn't change anything related to DPI and this is just Windows being weird with a new installation before I restart the computer. I dunno.

On another note, it seems the Red Mountain wind sound is the vanilla Morrowind sound, not the GOTY sound. Back when I played the original Xbox version and switched from vanilla Morrowind to GOTY, I remember the wind sound being different. Maybe there could be a global value that if the player has Tribunal and Bloodmoon loaded, it should play the GOTY sound instead?

I noticed shadows in 0.46 become very jagged if the camera looks directly at or away from the sun, even at 4096 shadow map size, but I followed a bit of AnyOldName3's shadow work in the past so I understand there are some difficult algorithms and complications involved (the fact that it works at all... well, you know).

Another thing I noticed in 0.46 is that long-duration destruction spells on my character from like a Dagoth ash ghoul or ash zombie made the destruction particle kind of obscure the first-person view which didn't seem like a problem before. Maybe the alpha is thicker than it used to be?

Also, don't quote me on this but I vaguely remember the Xbox version's bow handling being such that the velocity of the arrow was tied to how long you held down the trigger during the attack animation. In 0.45 and 0.46, it seems like it only differentiates between either a quick click to start the arrow-knocking animation or releasing the left-mouse button after it is done animating.

Lastly, the Win64 installer from the GitHub releases tried restarting my computer without asking when it finished, an instant before the part where it shows the Run OpenMW checkbox. That seems wrong...

Anyway yeah, cool release. The things I mentioned above pushed beyond the threshold of 'should I post something to the forums' so that's why I'm saying anything at all :P

User avatar
akortunov
Posts: 785
Joined: 13 Mar 2017, 13:49
Location: Samara, Russian Federation

Re: Feedback on 0.46.0

Post by akortunov » 16 Jun 2020, 07:20


User avatar
Atahualpa
Posts: 1136
Joined: 09 Feb 2016, 20:03

Re: Feedback on 0.46.0

Post by Atahualpa » 16 Jun 2020, 07:24

afritz1 wrote:
16 Jun 2020, 05:48
Lastly, the Win64 installer from the GitHub releases tried restarting my computer without asking when it finished, an instant before the part where it shows the Run OpenMW checkbox. That seems wrong...
Oh damn! That happened on my PC as well but I blamed a parallel Windows update for it. Maybe, for the first time, Windows wasn't the culprit.
Can anyone else confirm this?

User avatar
psi29a
Posts: 4987
Joined: 29 Sep 2011, 10:13
Location: Belgium
Gitlab profile: https://gitlab.com/psi29a/
Contact:

Re: Feedback on 0.46.0

Post by psi29a » 16 Jun 2020, 08:55

Could it be that it was installing the c++ redist package and that triggered windows to say; reboot!

I think that will happen every time that redist is updated.

User avatar
Atahualpa
Posts: 1136
Joined: 09 Feb 2016, 20:03

Re: Feedback on 0.46.0

Post by Atahualpa » 16 Jun 2020, 10:21

That might be the culprit, indeed. A similar case is described here: Upgrading to Microsoft Visual C++ Redistributable for Visual Studio 2017.
Because of the dependency removal on the EOL Microsoft Visual C++ Redistributable, CA UIM 9.0.2 installation deploys the VS 2017 package 1.0 (vs2017_vcredist_x86 1.0 and vs2017_vcredist_x64 1.0). By default, the above-listed probe versions use this VS 2017 package 1.0 (vs2017_vcredist_x86 1.0 and vs2017_vcredist_x64 1.0) as a dependency. This dependency on the 1.0 package causes Windows operating system to restart on the robot where the probe is deployed. In the VS 2017 package version 1.0, there is no "norestart" option specified in the Post Install command. Therefore, the Post Install might trigger operating system reboot as part of the package installation. We recommend you to download the vs2017_vcredist_x86 1.01 and vs2017_vcredist_x64 1.01 package (as required) from the Nimsoft archive to avoid the computer auto-restart for the dependent probes.

User avatar
AnyOldName3
Posts: 2048
Joined: 26 Nov 2015, 03:25

Re: Feedback on 0.46.0

Post by AnyOldName3 » 16 Jun 2020, 13:23

Regarding UI scaling, don't we have a separate setting for that so it's not tied to resolution?
AnyOldName3, Master of Shadows

clocknova
Posts: 52
Joined: 11 Sep 2013, 17:34

Re: Feedback on 0.46.0

Post by clocknova » 16 Jun 2020, 14:08

There is a setting for that, under [GUI]. I like to set it to "scaling factor = 1.2" on my screen at high resolutions.

User avatar
afritz1
Posts: 47
Joined: 05 Sep 2016, 01:18
Contact:

Re: Feedback on 0.46.0

Post by afritz1 » 16 Jun 2020, 17:53

The UI scaling factor works great -- seems way better for me at 2.0. 0.46 is much more playable now :)

I should try using that high DPI fix on OpenTESArena since it is not DPI aware yet and I've been experiencing weirdness on Windows where it sometimes decides to mark an .exe as DPI aware if it has ever changed from exclusive fullscreen to windowed, so I have Release and Debug builds which are apparently DPI aware on my PC due to my experimentations with SDL fullscreen but RelWithDebInfo is not.

The Windows restart issue with the Win64 installer seems to be 100% reproducible since I uninstalled and reinstalled 0.46 last night to check.

It seems it is also possible that Windows 10 will completely hard lock and require a computer restart if 0.46 and Unity are both open. Don't know if that's just AMD drivers or what. Probably just a rare edge case with my computer? Might be worse than that because Windows will usually reset the display driver if it hangs for too long. This was a complete hard lock at the main menu load game window; couldn't Ctrl-Alt-Delete or anything to get out of it.

User avatar
Capostrophic
Posts: 776
Joined: 22 Feb 2016, 20:32

Re: Feedback on 0.46.0

Post by Capostrophic » 16 Jun 2020, 21:04

Another thing I noticed in 0.46 is that long-duration destruction spells on my character from like a Dagoth ash ghoul or ash zombie made the destruction particle kind of obscure the first-person view which didn't seem like a problem before. Maybe the alpha is thicker than it used to be?
That's also the case in most previous releases. It'll be resolved in 0.47.

mjmax
Posts: 14
Joined: 06 May 2012, 22:31

Re: Feedback on 0.46.0

Post by mjmax » 17 Jun 2020, 02:53

Atahualpa wrote:
16 Jun 2020, 07:24
afritz1 wrote:
16 Jun 2020, 05:48
Lastly, the Win64 installer from the GitHub releases tried restarting my computer without asking when it finished, an instant before the part where it shows the Run OpenMW checkbox. That seems wrong...
Oh damn! That happened on my PC as well but I blamed a parallel Windows update for it. Maybe, for the first time, Windows wasn't the culprit.
Can anyone else confirm this?
Same here on the Win64 installer! No warning or anything.

Post Reply