Pages: « 1 2
  Print  
Author Topic: Phase 2/2 - Final testing for the fonts and scrolling issues (RESOLVED)  (Read 19276 times)
Offline (Unknown gender) Darkstar2
Reply #15 Posted on: June 10, 2014, 12:16:49 pm
Member
Joined: Jan 2014
Posts: 1238

View Profile Email
Yeah if you notice I used the delta_time or w/e constant to calculate my scrolling background.

There are many ways to do your own scrolling but majority of people will use traditional methods - When I modified your demo, it looked 1:1 identical yet did not have the gaps anymore.

Quote
Anyway, no I said I don't wtf to your code you sent me in a PM because your rounding was mixed with earlier suggestions we made for vertices, not the projection coordinates. Thus my wtf response.

I did try the proper rounding method before but I used X+ instead of x =, so now it all works.  I just finished testing this with DX9, window and full screen and it worked too.
I removed the +0.25 thing that was there, I checked before and after and fonts looked identical, and replaced it with the rounding fix.  So now OpenGL1, OpenGL3 and DX9 are patched with this fix. 

Quote
Anyway, wonderful, good thing it works, and yeah I'm certain nobody would ever have a need for a camera at a half pixel would they? When the camera is at half-pixels all you ever notice are artifacts, it looks the same as if it were at the rounded coordinate except shittier.

Exactly, I mean those are little details, what matters is the game, and visually the majority of people playing the games will not dissect the pixel placement on the screen.  With all the movement and concentrating on the game those are details :D

Quote
So am I correct in my assumptions or does someone have evidence to the contrary?

Actually your assumption is logic.

We established at 0.01 works for us all,
and you've guided me to discovering that whenever user x/y + engine offsets = an integer, that the artifacts appear again, and yes half a pixel looks uglier than rounded to the nearest, after looking closely did not see .01 adding any visual effects to the fonts, I mean most people don't care or don't have trained eyes, and besides when playing at your monitor's native resolution 1:1 with your game res, it is never an issue.  Most of the blurriness or other spacing/thickness inconsistencies will be seen due to monitor upscaling and game scaling, as  a lot of people use room sizes < than their monitor res anyway. 

« Last Edit: June 22, 2014, 12:54:34 pm by Darkstar2 » Logged
Offline (Male) Goombert
Reply #16 Posted on: June 10, 2014, 12:43:33 pm

Developer
Location: Cappuccino, CA
Joined: Jan 2013
Posts: 2991

View Profile
Alright good I was just going to tell you to patch DX as well. But please leave any FS fixes or anything out of this pull request. Anyway, go ahead and send it and I'll tell Josh or Harri to merge it.
Logged
I think it was Leonardo da Vinci who once said something along the lines of "If you build the robots, they will make games." or something to that effect.

Offline (Unknown gender) Darkstar2
Reply #17 Posted on: June 10, 2014, 01:14:06 pm
Member
Joined: Jan 2014
Posts: 1238

View Profile Email
Alright good I was just going to tell you to patch DX as well. But please leave any FS fixes or anything out of this pull request. Anyway, go ahead and send it and I'll tell Josh or Harri to merge it.

Ok will do this now, and don't worry, the DX9 fixing was just a personal experiment, I'll show you the script I made for it later so you can tell me what you think. I'll integrate that into my own projects using DX9, will wait for the window resizing bug to get fixed first.

want me to patch DX11 too ? lol !
:P
Logged
Offline (Male) Rusky
Reply #18 Posted on: June 10, 2014, 01:28:43 pm

Resident Troll
Joined: Feb 2008
Posts: 954
MSN Messenger - rpjohnst@gmail.com
View Profile WWW Email
Off topic- fixed time step for physics and input handling is essential to get deterministic behavior with varying performance. delta_time is good for getting a higher visual framerate using interpolation when rendering, but applying it to motion is a recipe for frustration on the player's side.
Logged
Offline (Unknown gender) Darkstar2
Reply #19 Posted on: June 10, 2014, 01:59:04 pm
Member
Joined: Jan 2014
Posts: 1238

View Profile Email
Added a pull request
https://github.com/enigma-dev/enigma-dev/pull/750

Fixed OGL1, OGL3 and DX9, rounds first then adds offset to prevent any issues with moving projections or fractional coordinates that may negate the fix and cause the visual anomalies.
Tested all 3 and working.

Logged
Offline (Unknown gender) egofree
Reply #20 Posted on: June 10, 2014, 02:46:04 pm
Contributor
Joined: Jun 2013
Posts: 601

View Profile Email
I've an Alienware X51 with an Geoforce GTX 660 OEM. On Windows 7 64 bits, everything is ok.
Logged
Offline (Unknown gender) Darkstar2
Reply #21 Posted on: June 10, 2014, 03:17:32 pm
Member
Joined: Jan 2014
Posts: 1238

View Profile Email
I've an Alienware X51 with an Geoforce GTX 660 OEM. On Windows 7 64 bits, everything is ok.

Thanks egofree.

A merge will be made soon where the X/Y projection gets rounded first before .01 applied to it, I won't make a new EXE and ask people to test again, because it is not necessary.  There was one test missing which was moving projection using fractional increments, and that was problematic, so rounding fixes everything up.   This rounding thing might have also proactively fixed any potential issues not covered by the test.
At least in theory, but we all know anything is possible beyond belief, with ENIGMA, almost paranormal lol.

@Robert:
https://www.dropbox.com/s/ktliuq5czsp7e8k/DX9Fonts1.png

Who asked me why I removed the 0.25 offset that was present in DX9.  With rounding I don't think this needed, besides I tested and .01 gave identical results to 0.25, no blurriness from not using 0.25 as you can see above the left 1 is 0.25 and right 1 is 0.01
« Last Edit: June 10, 2014, 04:04:40 pm by Darkstar2 » Logged
Offline (Male) edsquare
Reply #22 Posted on: June 10, 2014, 05:43:35 pm

Member
Location: The throne of ringworld
Joined: Apr 2014
Posts: 402

View Profile
Ok on Windows7 32 with ATI [Radeon X1200] no problem with or without fix.

Also ok on linuxmint 15 under wine with the free drivers.
Logged
A child of five would understand this. Send someone to fetch a child of five.
Groucho Marx
Offline (Male) Goombert
Reply #23 Posted on: June 11, 2014, 08:59:39 am

Developer
Location: Cappuccino, CA
Joined: Jan 2013
Posts: 2991

View Profile
0.01,0.01 works in Direct3D9 as well and fixes blurry text for me. The pull request can be merged.
https://github.com/enigma-dev/enigma-dev/pull/750

Now it is important to note that font characters in Direct3D9 appear 1px to the left of where they do in OGL, this is further related to the issue we've been discussing. So this is clearly differences in the sampler of the two API's as well as differences in driver implementations of the two API's. Also D3D9 does this whether you have offsets 0.25 or 0.01 or no offsets.
Logged
I think it was Leonardo da Vinci who once said something along the lines of "If you build the robots, they will make games." or something to that effect.

Offline (Unknown gender) Darkstar2
Reply #24 Posted on: June 12, 2014, 11:19:11 pm
Member
Joined: Jan 2014
Posts: 1238

View Profile Email
Thanks to everyone who took part in this test the results so far are positive.  It has since been merged.  Robert, are you going to make a new portable zip with the recent changes?

I will keep this topic open for a while in case people want to add anything.

Ok topic locked as this issue was resolved with success, no complaints, and everything so far is working as it should.  Again, thanks for your help.
« Last Edit: June 22, 2014, 12:53:57 pm by Darkstar2 » Logged
Pages: « 1 2
  Print