|
Update 05/06/2025: SoulKey's settings
In Korean...if someone can help translate please PM me.
+ Show Spoiler [Speed/Mouse] +
Update 04/06/2025: Bisu's settings
+ Show Spoiler +OK so Bisu is doing something weird. He plays with hardware cursor enabled, yet it is not scaled to the regular size expected. From comparing with different display resolutions and observing his VODs, it appears he manual sets up 960x720 resolution in Windowed mode. The 960x720 resolution would explain the "blurry" look that his Starcraft client has in-game, even though he is streaming at 1080p 8000kpbs. I wonder if this strange setup at home has something to do with his less-than-stellar performance in the ASL studio, where he might not be able to replicate his at-home settings perfectly (c.f. flash's ruler)Anyway: Mouse: : Logitech Mini Optical (implanted Logitech G303 sensor) (DPI 1050) Keyboard: QSENN DT35 Ingame: + Show Spoiler [Speed] + + Show Spoiler [Video] +Unknown, likely Windowed @ 960x720 or 1280x720 pillarbox
Update 04/02/2025: Light's settings
+ Show Spoiler +
Update 04/01/2025: Jaedong's settings
+ Show Spoiler +
Update 10/21/2024: FlaSh's settings
+ Show Spoiler +
Information in this post was compiled from various sources including community members, korean pro-gamers, blizzard forums, etc. There is no conclusive evidence that there is one single "best setting".
Your mileage may vary.
Purpose:
1. Inform the foreign community about these settings
2. Inform the foreign community that these settings are used by some pro-gamers
3. Give players the information they need to experiment with the settings and find the ones which work for them.
I. Battle.net Settings (Optional)
+ Show Spoiler +1. Click Ingame Settings
2. Add "--renderer=" (no quotes) to custom command line arguments
Available options (your mileage may vary):
--renderer=legacy *not recommended
--renderer=dx9
--renderer=dx10 *recommended
--renderer=dx11
--renderer=dx12
--renderer=opengl
3. Enable 32bit mode (Optional)
32-bit mode is only needed if you need DPI scaling for the mouse compatibility fix in Section III (see below)
II. Starcraft Settings
A. Video
+ Show Spoiler +1. Fullscreen mode
2. Disable FPS Cap
3. Set Gamma = 5/7 (helps with DT/burrowed unit visibility) 4. Set SD Filter = 3/4 ( reduces Mouse Lag even if HD gfx enabled)
B. Speed
+ Show Spoiler +1. Enable hardware cursor: (Optional) EXPLANATION: Technically speaking, it is drawing the mouse cursor overlay separately using a separate GPU register on top of the framebuffer. To explain it a little more clearly, it is a method of drawing the in-game screen and mouse cursor movement as separate layers. Disabling this option will cause the game engine to draw the in-game content and the cursor position on a single layer. If the in-game frame is low, the cursor movement will also have low frame rates. If you check this option, the output will be matched to the refresh rate (monitor refresh rate) regardless of the game engine's frame rate. Even if the in-game frame rate is low, the cursor movement is still expressed smoothly without appearing to be interrupted. As a result, the delay between the actual mouse movement and the cursor movement output can be shortened. So which one is better? In fact, Starcraft rarely has situations where the screen is displayed below the refresh rate of your monitor. In reality, it is an option that has no real meaning other than the difference in cursor size. Nevertheless, from a technical standpoint, it's a good idea to turn it on. However, if you feel more comfortable deactivating it, you can deactivate it and use it. 2. Adjust mouse sensitivity to your preference. C.f "Progamer Mouse Settings" thread for further info on mouse sensitivity Various progamer mouse settings: + Show Spoiler +Flash Windows Mouse Sensitivity (in control panel): 6/11 Mouse 800dpi, 500hz Scan Windows: 5/11 DPI: 800 In-Game Sensitivity: 65% Hardware Cursor: Off Mouse Scaling: Off Mouse: FKMini3 Light: (this has since been updated, but the gist is that his sensitivity is insanely high -- he barely moves his mouse in handcam videos) Windows: 10/11 DPI: 450 Mouse Scaling: On In-Game Sensitivity: Off Resolution: 1080p In-Game Sensitivity: Off Source: Mouse: Logitech G Pro Superlight 2 (but with these settings he was using a Logitech G Pro Wired) Soma Windows: 6/11 DPI: 800 In-Game Sensitivity: 65% Hardware Cursor: Off Mouse Scaling: Off FKMini3]
III. Windows Settings
Note: Advanced windows optimizations are beyond the scope of this thread.
A. Mouse
+ Show Spoiler +1. Open Control Panel and Select Mouse Options: --set slider to 6/11 --uncheck "Enhanced Pointer Precision"
B. Performance
+ Show Spoiler +1. Run (⊞ Win + R) the following: %windir%\system32\SystemPropertiesPerformance.exe 2. Adjust for best performance:
C. Enable classic mouse scaling AND Hardware cursor
+ Show Spoiler +We all know and love the Starcraft original interface. Note the size of the mouse cursor -- it is almost as large as a dragoon. This is the "classic" mouse icon scale.
So what is the problem?Enabling hardware cursor is best because it lowers mouse lag. But the cursor shrinks very small. Compare the normal mouse cursor size (shown above) to below: its sooo tiny!!
In addition, Starcraft will tell you "Hardware cursor not available when using mouse scaling":
But never fear! To play with hardware cursor and enable classic "large cursor", we just need to adjust windows settings
FIX: 0. 32-bit mode must be enabled (see Section I above) 1. Open Starcraft x86 directory (C:\Program Files (x86)\StarCraft\ x86), Right-click Starcraft.exe and select Properties 2. Select compatibility tab, click "Change settings for all users" 3. Follow settings as shown below:
4. Open Windows Settings -> select System -> Display Set Display Scaling = 150%
END RESULT:
D. Optional Settings for Advanced Users:+ Show Spoiler +Since we are using a Windows Scale not equal to 100%, applying the following registry tweak is strongly recommended, because it gieves exactly 1-to-1 mouse to pointer response when playing at 150% Scale: https://www.mediafire.com/file/ixrk61ns5cgp4no/Windows_10 8.x_MouseFix.zip/fileDownload, Unzip, and double-click "Windows_10+8.x_MouseFix_ItemsSize=150%_Scale=1-to-1_@6-of-11.reg". When prompted to Merge, choose yes:
To reverse above changes, double-click "Windows_10+8.x_Default.reg" and again choose yes to merge.
IV. References
+ Show Spoiler +
|
thank you!
what do the bnet settings actually do ingame?
|
On June 21 2024 21:45 WGT-Baal wrote:
what do the bnet settings actually do ingame?
--render setting selects the "original" (1.16.1) graphics pipeline (dx9, very fast)
--32-bit mode is required for the mouse compatibility fix in Section III
|
|
|
Will this fix my issues with not building depots?
|
On June 22 2024 00:35 kogeT wrote:
Will this fix my issues with not building depots?
that depends on whether the missed depots are due to lost keyboard/mouse inputs, or due to player error
|
|
|
Yes. I have experimented with all the available renderers at uncapped FPS. DX9 give highest uncapped performance (1800 FPS on RTX3060), but "--renderer=legacy" has the best "game feel", and feels particularly "snappy" when rapid changes in camera location occur, e.g F2-F4 screen location hotkeys, minimap clicks/drags.
I have not investigated too deeply since the inevitable differences between each player's hardware, drivers, and personal "game feel" preference suggests that "mileage may vary". There is no "one best setting" for all possible situations (or if there is such a "best setting", it is unknown)
Feel free to experiment and find a renderer setting that works/feels/plays best for you and your hardware.
|
Using DX9 if you care about input latency just seems very counterproductive, and Blizzard had planned to remove the renderer entirely before they axed the team, so I wouldn't really recommend it. Higher FPS does not imply lower latency. Any statements that things "feel snappier" without any actual data or measurement are basically worthless, humans are incredibly bad at identifying this when asked.
I would highly recommend using the modern (e.g. default) renderer, and using Windowed Fullscreen instead of exclusive Fullscreen (this will have identical latency to exclusive Fullscreen on modern graphics APIs, without all of the downsides that come with exclusive Fullscreen). Note that if you decide to ignore me and use DX9 anyway, you definitely should NOT use Windowed Fullscreen as in DX9 it will force the game output to be copied an extra time and add an extra frame of input latency at minimum.
I also really have to wonder about the effects of forcing DPI scaling on the game, this seems like a pretty bad idea (especially since it is just for making the mouse cursor a little bigger). I'd test out the latency effects with PresentMon but I'd really rather not mess with DPI scaling on my system, so maybe since you're the one suggesting it you should do that yourself.
|
Tec27:
Sincerely, as sincerely as possible, Thank you SO MUCH for sharing your thoughts in this thread. I know your background with OpenBw SB, etc. so i was really hoping you would see this thread and "chime in"
That being said, let us join a debate, late-19th-century-in-Vienna-style: dispassionate and cordial. (dont forget to equip monocle and wig). i will choose the side of pragmatist.
On June 22 2024 02:07 tec27 wrote:
Using DX9 if you care about input latency just seems very counterproductive
why?
On June 22 2024 02:07 tec27 wrote:
and Blizzard had planned to remove the renderer entirely before they axed the team
source?
On June 22 2024 02:07 tec27 wrote:
Higher FPS does not imply lower latency.
Yes you are correct. Please quote in the OP where this was implied? I did not intend to make such implications.
On June 22 2024 02:07 tec27 wrote:
Any statements that things "feel snappier" without any actual data or measurement are basically worthless
Worthess to whom? Let us observe player behavior and list the facts:
Fact #1: (good) players DO use them
Now, lets ask ourselves why why?
I will argue that "game feel" matters for StarcraftBW, more than any other game.
I'd test out the "subjective degree of worthlessness" by emailing Bisu/Flash/Light to ask each of them how much value they assign to their setups, but I'd really rather not, so maybe since you're the one suggesting it you should do that yourself.
On June 22 2024 02:07 tec27 wrote:
I would highly recommend using the modern (e.g. default) renderer, and using Windowed Fullscreen instead of exclusive Fullscreen (this will have identical latency to exclusive Fullscreen on modern graphics APIs, without all of the downsides that come with exclusive Fullscreen).
What downsides? What about plus-sides? For example: if using windowed mode, player cannot adjust Gamma levels ingame
On June 22 2024 02:07 tec27 wrote:
Note that if you decide to ignore me and use DX9 anyway, you definitely should NOT use Windowed Fullscreen as in DX9 it will force the game output to be copied an extra time and add an extra frame of input latency at minimum.
source?
and please show where in the OP that windowed fullscreen was recommended?
On June 22 2024 02:07 tec27 wrote:
I also really have to wonder about the effects of forcing DPI scaling on the game
what effects?
On June 22 2024 02:07 tec27 wrote:
this seems like a pretty bad idea
why?
On June 22 2024 02:07 tec27 wrote:
(especially since it is just for making the mouse cursor a little bigger)
i believe you missed the point about the cursor.
first ill address your point. let us assume "just make the cursor a little bit bigger" WAS the point. your choice of words ("a little bit bigger") really depends on player hardware. For example, if you play at 1440p or above, the default hardware cursor is so small its practically invisible. So it is not "just a LITTLE bit bigger". But again, that was NOT the point in the first place.
Next, I think your choice of words marginalizes the effect the cursor size has on the player and the gameplay. But perhaps I am overstating the effect? To find out, perhaps you could go speak with Bisu/Light/Sacsri and kindly inform them that they have been using the wrong mouse cursor size! And please do let us know their response!
Lastly, I decided to update the OP and make clearer the purpose of this guide. Summary: I am just the messenger. Dont shoot the messenger.
On June 22 2024 02:07 tec27 wrote:
I'd test out the latency effects with PresentMon but I'd really rather not mess with DPI scaling on my system, so maybe since you're the one suggesting it you should do that yourself.
Nowhere is anyone advocating the use of any of these tweaks.
|
On June 22 2024 10:04 tankgirl wrote:Tec27: Sincerely, as sincerely as possible, Thank you SO MUCH for sharing your thoughts in this thread. I know your background with OpenBw SB, etc. so i was really hoping you would see this thread and "chime in" That being said, let us join a debate, late-19th-century-in-Vienna-style: dispassionate and cordial. (dont forget to equip monocle and wig). i will choose the side of pragmatist.
Show nested quote +On June 22 2024 02:07 tec27 wrote:
Using DX9 if you care about input latency just seems very counterproductive
why? Show nested quote +On June 22 2024 02:07 tec27 wrote:
and Blizzard had planned to remove the renderer entirely before they axed the team
source? Yes you are correct. Please quote in the OP where this was implied? I did not intend to make such implications. Show nested quote +On June 22 2024 02:07 tec27 wrote:
Any statements that things "feel snappier" without any actual data or measurement are basically worthless
Worthess to whom? Let us observe player behavior and list the facts: Fact #1: (good) players DO use them Now, lets ask ourselves why why? I will argue that "game feel" matters for StarcraftBW, more than any other game. I'd test out the "subjective degree of worthlessness" by emailing Bisu/Flash/Light to ask each of them how much value they assign to their setups, but I'd really rather not, so maybe since you're the one suggesting it you should do that yourself. Show nested quote +On June 22 2024 02:07 tec27 wrote:
I would highly recommend using the modern (e.g. default) renderer, and using Windowed Fullscreen instead of exclusive Fullscreen (this will have identical latency to exclusive Fullscreen on modern graphics APIs, without all of the downsides that come with exclusive Fullscreen).
What downsides? What about plus-sides? For example: if using windowed mode, player cannot adjust Gamma levels ingame Show nested quote +On June 22 2024 02:07 tec27 wrote:
Note that if you decide to ignore me and use DX9 anyway, you definitely should NOT use Windowed Fullscreen as in DX9 it will force the game output to be copied an extra time and add an extra frame of input latency at minimum.
source? and please show where in the OP that windowed fullscreen was recommended? Show nested quote +On June 22 2024 02:07 tec27 wrote:
I also really have to wonder about the effects of forcing DPI scaling on the game
what effects? why? Show nested quote +On June 22 2024 02:07 tec27 wrote:
(especially since it is just for making the mouse cursor a little bigger)
i believe you missed the point about the cursor. first ill address your point. let us assume "just make the cursor a little bit bigger" WAS the point. your choice of words ("a little bit bigger") really depends on player hardware. For example, if you play at 1440p or above, the default hardware cursor is so small its practically invisible. So it is not "just a LITTLE bit bigger". But again, that was NOT the point in the first place. Next, I think your choice of words marginalizes the effect the cursor size has on the player and the gameplay. But perhaps I am overstating the effect? To find out, perhaps you could go speak with Bisu/Light/Sacsri and kindly inform them that they have been using the wrong mouse cursor size! And please do let us know their response! Lastly, I decided to update the OP and make clearer the purpose of this guide. Summary: I am just the messenger. Dont shoot the messenger. Show nested quote +On June 22 2024 02:07 tec27 wrote:
I'd test out the latency effects with PresentMon but I'd really rather not mess with DPI scaling on my system, so maybe since you're the one suggesting it you should do that yourself.
Nowhere is anyone advocating the use of any of these tweaks.
Didn't know that 19th century monocle debate involved high levels of sarcasm and appeals to authority. Source?
|
On June 22 2024 16:24 Jealous wrote:
Didn't know that 19th century monocle debate involved high levels of sarcasm and appeals to authority. Source?
haha you're right! good one! thanks for your contribution =)
|
On June 22 2024 16:45 tankgirl wrote:Show nested quote +On June 22 2024 16:24 Jealous wrote:
Didn't know that 19th century monocle debate involved high levels of sarcasm and appeals to authority. Source? haha you're right! good one! thanks for your contribution =)
haha you're welcome! thanks! i enjoy contributing =)
Next, I think your choice of words marginalizes the effect my comment has on your thread and the discourse. But perhaps I am overstating the effect? To find out, perhaps you could go speak with TL administrators and kindly inform them that they have been using the wrong posting standards! And please do let us know their response!
|
:Michael Jackson eats popcorn:
|
On June 22 2024 10:04 tankgirl wrote:Tec27: Sincerely, as sincerely as possible, Thank you SO MUCH for sharing your thoughts in this thread. I know your background with OpenBw SB, etc. so i was really hoping you would see this thread and "chime in" That being said, let us join a debate, late-19th-century-in-Vienna-style: dispassionate and cordial. (dont forget to equip monocle and wig). i will choose the side of pragmatist.
Show nested quote +On June 22 2024 02:07 tec27 wrote:
Using DX9 if you care about input latency just seems very counterproductive
why? Show nested quote +On June 22 2024 02:07 tec27 wrote:
and Blizzard had planned to remove the renderer entirely before they axed the team
source? Yes you are correct. Please quote in the OP where this was implied? I did not intend to make such implications. Show nested quote +On June 22 2024 02:07 tec27 wrote:
Any statements that things "feel snappier" without any actual data or measurement are basically worthless
Worthess to whom? Let us observe player behavior and list the facts: Fact #1: (good) players DO use them Now, lets ask ourselves why why? I will argue that "game feel" matters for StarcraftBW, more than any other game. I'd test out the "subjective degree of worthlessness" by emailing Bisu/Flash/Light to ask each of them how much value they assign to their setups, but I'd really rather not, so maybe since you're the one suggesting it you should do that yourself. Show nested quote +On June 22 2024 02:07 tec27 wrote:
I would highly recommend using the modern (e.g. default) renderer, and using Windowed Fullscreen instead of exclusive Fullscreen (this will have identical latency to exclusive Fullscreen on modern graphics APIs, without all of the downsides that come with exclusive Fullscreen).
What downsides? What about plus-sides? For example: if using windowed mode, player cannot adjust Gamma levels ingame Show nested quote +On June 22 2024 02:07 tec27 wrote:
Note that if you decide to ignore me and use DX9 anyway, you definitely should NOT use Windowed Fullscreen as in DX9 it will force the game output to be copied an extra time and add an extra frame of input latency at minimum.
source? and please show where in the OP that windowed fullscreen was recommended? Show nested quote +On June 22 2024 02:07 tec27 wrote:
I also really have to wonder about the effects of forcing DPI scaling on the game
what effects? why? Show nested quote +On June 22 2024 02:07 tec27 wrote:
(especially since it is just for making the mouse cursor a little bigger)
i believe you missed the point about the cursor. first ill address your point. let us assume "just make the cursor a little bit bigger" WAS the point. your choice of words ("a little bit bigger") really depends on player hardware. For example, if you play at 1440p or above, the default hardware cursor is so small its practically invisible. So it is not "just a LITTLE bit bigger". But again, that was NOT the point in the first place. Next, I think your choice of words marginalizes the effect the cursor size has on the player and the gameplay. But perhaps I am overstating the effect? To find out, perhaps you could go speak with Bisu/Light/Sacsri and kindly inform them that they have been using the wrong mouse cursor size! And please do let us know their response! Lastly, I decided to update the OP and make clearer the purpose of this guide. Summary: I am just the messenger. Dont shoot the messenger. Show nested quote +On June 22 2024 02:07 tec27 wrote:
I'd test out the latency effects with PresentMon but I'd really rather not mess with DPI scaling on my system, so maybe since you're the one suggesting it you should do that yourself.
Nowhere is anyone advocating the use of any of these tweaks.
Fair enough, killer logic, no wonder you got no contextual response, I like your enthusiasm & mindset, keep it on, unfortunately I rarely use Remastered so I can't talk from experience, I also don't see any appeal to authority here, the topic title & content aren't saying those are the best settings for everyone, so the fallacy isn't on your side, thanks for sharing your knowledge.
|
This is a very interesting thread indeed because it forces us to think and investigate for ourselves!
I suspect everyone has different settings.
I've noticed the game loads very slowly on my integrated graphics and almost crashes.
Also the cloaked units are harder to see.
I may fiddle around, ask if ppl have taken similar actions and report back.
Please don't kill each other over the thread tho! Foreign community small enough as it is.. kek
|
fiddled with what you said, aside from the gamma (which is visually anyway) everything else made it worse so I undid everything. I had already mouse 50% and hw cursor enabled though so this I would recommend for sure.
The various rendered were all worse.
the SD filter setting also was slightly worse (but more of a feeling, hard to say, i was at the setting just to the left of yours).
I am not on 1080p though and i suppose for mouse, in particular, your particular config of cpu,drivers,mouse drivers and windows setting all come into play.
I d like more input on the FPS cap though (tec?) I havent used it since 2017 when they released a buggy thing that would overheat the gpus, but it got patched quickly. I tried enabling it to 300, and then 250 and removing but couldnt see any difference. More details on what it does and does not do would be neat from either of you, thanks!
This is an interesting thread, thanks for making it.
|
I was wondering what went wrong after I got my new PC and it's that I forgot to increase the gamma correction.
Thanks! Great post 
|
On June 27 2024 21:20 WGT-Baal wrote:
fiddled with what you said, aside from the gamma (which is visually anyway) everything else made it worse so I undid everything. I had already mouse 50% and hw cursor enabled though so this I would recommend for sure.
The various rendered were all worse.
the SD filter setting also was slightly worse (but more of a feeling, hard to say, i was at the setting just to the left of yours).
I am not on 1080p though and i suppose for mouse, in particular, your particular config of cpu,drivers,mouse drivers and windows setting all come into play.
I d like more input on the FPS cap though (tec?) I havent used it since 2017 when they released a buggy thing that would overheat the gpus, but it got patched quickly. I tried enabling it to 300, and then 250 and removing but couldnt see any difference. More details on what it does and does not do would be neat from either of you, thanks!
This is an interesting thread, thanks for making it.
Tbh if you have a 240hz monitor, any FPS beyond 240 won't be felt.
I got the 300 fps cap tho (240hz 1080p monitor here) :D But there isn't any difference whatsoever.
The gamma correction helps against lurkers for sure. The rest of the settings are arbitrary imho.
|
On June 27 2024 23:32 BlueStar wrote:Show nested quote +On June 27 2024 21:20 WGT-Baal wrote:
fiddled with what you said, aside from the gamma (which is visually anyway) everything else made it worse so I undid everything. I had already mouse 50% and hw cursor enabled though so this I would recommend for sure.
The various rendered were all worse.
the SD filter setting also was slightly worse (but more of a feeling, hard to say, i was at the setting just to the left of yours).
I am not on 1080p though and i suppose for mouse, in particular, your particular config of cpu,drivers,mouse drivers and windows setting all come into play.
I d like more input on the FPS cap though (tec?) I havent used it since 2017 when they released a buggy thing that would overheat the gpus, but it got patched quickly. I tried enabling it to 300, and then 250 and removing but couldnt see any difference. More details on what it does and does not do would be neat from either of you, thanks!
This is an interesting thread, thanks for making it. Tbh if you have a 240hz monitor, any FPS beyond 240 won't be felt. I got the 300 fps cap tho (240hz 1080p monitor here) :D But there isn't any difference whatsoever. The gamma correction helps against lurkers for sure. The rest of the settings are arbitrary imho.
gamma correction is also awesome for minimap. helps tell apart things easier.
fullscreen is less screen lagg. Ive talked about it with some pros and they all notice a difference. I also notice it. Hardware cursor is amazingly fluid movement.
|
On June 27 2024 23:32 BlueStar wrote:. The rest of the settings are arbitrary imho.
you are entitled to your opinion. but some settings are NOT arbitrary. e.g SD graphics filter changing mouse latency.
On June 27 2024 16:09 franklyyes wrote: thanks for sharing your knowledge.
you're welcome.
i did not make this thread on a "whim". ive put in several months of work trawling through old stream vods on afreeca, tweaking settings, confirming what works and what does not =)
gl hf =))))
|
Thanks for doing this. I've also been keeping a short list of progamer mouse settings from Afreeca streams, but I've never come across this. For example:
Scan
Windows: 5/11
DPI: 800
In-Game Sensitivity: 65%
Hardware Cursor: Off
Mouse Scaling: Off
Resolution: 1080p
Mouse: FKMini3
Light: (this has since been updated, but the gist is that his sensitivity is insanely high -- he barely moves his mouse in handcam videos)
Windows: 10/11
DPI: 450
Mouse Scaling: On
In-Game Sensitivity: Off
Resolution: 1080p
In-Game Sensitivity: Off
Source: https://www.youtube.com/watch?v=eLEvO7tmUoY
Mouse: Logitech G Pro Superlight 2 (but with these settings he was using a Logitech G Pro Wired)
Soma
Windows: 6/11
DPI: 800
In-Game Sensitivity: 65%
Hardware Cursor: Off
Mouse Scaling: Off
FKMini3
|
Man this is awesome. I never knew about this.
Anyone know what mouse and keyboard Flash is using now btw?
|
On June 22 2024 17:03 Jealous wrote:
Next, I think your choice of words marginalizes the effect my comment has on your thread and the discourse. But perhaps I am overstating the effect? To find out, perhaps you could go speak with TL administrators and kindly inform them that they have been using the wrong posting standards! And please do let us know their response!
Right! To which standards are you referring? Did you mean these standards?
On September 13 2004 22:42 mensrea wrote:
...don't be a backseat moderator...
...it's very important to be informed about a discussion before posting your thoughts...
...some people contribute by being helpful when someone posts a question...Others provide resources or news...
Do you have Starcraft: Broodwar installed? Feel free to test out some of the settings in the guide and report back with any questions, news from your own research findings, or to share alternatives to these settings with any other Starcraft: Broodwar players who might stumble upon this thread =)
On June 29 2024 00:15 LostDevil wrote:
Thanks for doing this. I've also been keeping a short list of progamer mouse settings from Afreeca streams, but I've never come across this. For example:
...
Great! I'm updated the OP with this new info. Thanks =)))
On June 29 2024 01:52 llIH wrote:
Man this is awesome. I never knew about this.
Anyone know what mouse and keyboard Flash is using now btw?
Most recent info I have is from 2021:
Mouse: Zowie Mico
Keyboard: QSENN DT35
Anyway, sorry folks...I looked around for 6-7 hours again this morning to find the Sacsri VOD showing "in detail" the mouse "lag" changing with the SD graphics filter settings...no luck...bleh...ill try again later...not that I think it is absolutely necessary...to me it is quite a noticeable difference...however since this is kindof my own humble little contribution to the community I think it dutiful to provide accurate sources...I will keep updating the thread as time allows. gl hf =)
|
What about HD graphics, smooth unit turning, shadow stacking and pillar box? This varies between the pros?
|
Thanks. I think it made the SD graphics look better (more like the pre-SCR ones) at least for me..
|
My two cents: Scan uses 1000dpi, 55% sensitivity IG
|
Hi guys, I don't know about the renderer setting + display scale 150%(it did not work for me + starcraft screen becomes blurry) + SD Filters and the others.
But I'm here to share my information that I looked up on Korean community.
If you are using hardware cursor, this feature ignores game engine's framerate(FPS) and show your cursor frame as your monitor refresh rate.
If you are not using this feature, your computer is using GPU register and individually draws/shows your cursor on screen/layer.
So people who did not understand my poor explanation, if your computer spec is poor like using 15 year old laptop/desktop or whatever, playing on 60hz monitor with bad graphics card, you will feel the lag
Ex) If game client is lagging(whatever the reason is, like UMS too many units) + having a FPS drop a lot = your cursor will lag as well
15 seconds video explanation: www.fmkorea.com
You can also use google translate and read what he has explained there + recorded video.
On July 01 2024 21:22 Glioburd wrote:
My two cents: Scan uses 1000dpi, 55% sensitivity IG
I was, but I've switched to 800dpi 65% long time ago.
for people who want to know my setting,
Scan
Windows: 5/11
DPI: 800
In-Game Sensitivity: 65%
Hardware Cursor: Off
Mouse Scaling: Off
FKMini3
|
On July 02 2024 18:09 LaStScan wrote:Hi guys, I don't know about the renderer setting + display scale 150%(it did not work for me + starcraft screen becomes blurry) + SD Filters and the others. But I'm here to share my information that I looked up on Korean community. If you are using hardware cursor, this feature ignores game engine's framerate(FPS) and show your cursor frame as your monitor refresh rate. If you are not using this feature, your computer is using GPU register and individually draws/shows your cursor on screen/layer. So people who did not understand my poor explanation, if your computer spec is poor like using 15 year old laptop/desktop or whatever, playing on 60hz monitor with bad graphics card, you will feel the lag Ex) If game client is lagging(whatever the reason is, like UMS too many units) + having a FPS drop a lot = your cursor will lag as well 15 seconds video explanation: www.fmkorea.comYou can also use google translate and read what he has explained there + recorded video. Show nested quote +On July 01 2024 21:22 Glioburd wrote:
My two cents: Scan uses 1000dpi, 55% sensitivity IG I was, but I've switched to 800dpi 65% long time ago. for people who want to know my setting, Scan Windows: 5/11 DPI: 800 In-Game Sensitivity: 65% Hardware Cursor: Off Mouse Scaling: Off FKMini3
that video shows why hardwarw cursor is cool!
|
I can attest tankgirl's post is accurate if you want the best competitive settings. At the same time, there are a lot of pros that use their own settings. Mouse speed not being 6/11, using mouse scaling, using over 60% mouse speed so on. The only thing I can say to that is it is their own personal settings and should not be replicated. While tankgirls settings are a good starting point, and you can tweak those to create your own personal settings
|
I could´nt get the cursor to work as described.
|
This fixed many issues, thanks! I still have so many issues though that make playing BW painful.
Every time I start SC (from the battle.net launcher usually), I need to log in again from within SC. Everything is laggy, usually takes a few minutes. Once I enter my username/pw, it says "Downloading Settings". Takes a couple minutes, until saying "Download failed." Everything is laggy, takes a minute to click OK, pick a server, and get ready to search for ladder games. Usually takes like 3-4 searches - first few searches end in error (error 50:6 iirc).
Once in game, everything works perfectly, with one exception. If I open menu in-game, it usually lags for like 3-5 seconds. If I need to change latency.
Tried running b.net launcher and starcraft.exe as administrator. Tried reinstalling many, many times. Whitelisting everything starcraft in antivirus (and turning off antivirus completely). Nothing has helped, until following these instructions! Now the lag and wait times have been reduced by about half, but still very annoying. Any help with these issues is greatly appreciated
|
|
|
On June 29 2024 18:25 llIH wrote:
What about HD graphics, smooth unit turning, shadow stacking and pillar box? This varies between the pros?
yes, it varies.
+ Show Spoiler +Sacsri:
HD - on
smooth unit turning - off
shadow stacking - off
pillar box - off
Flash:
HD - on
smooth unit turning - on
shadow stacking - on
pillar box -on
On July 04 2024 03:52 Mutaller wrote:
I can attest tankgirl's post is accurate if you want the best competitive settings. At the same time, there are a lot of pros that use their own settings. Mouse speed not being 6/11, using mouse scaling, using over 60% mouse speed so on. The only thing I can say to that is it is their own personal settings and should not be replicated. While tankgirls settings are a good starting point, and you can tweak those to create your own personal settings
100%!!!
On July 04 2024 04:29 Suikakuju wrote:
I could´nt get the cursor to work as described.
Sometimes the settings do not save properly:
1. Set windows scale to 150% in settings
2. Sign out of windows and log back in.
3. Set Starcraft.exe compatibility option DPI mode to "Application" in programfiles/starcraft/x86/starcraft.exe
4. Make sure you run starcraft in 32-bit mode.
5. Exit starcraft.
6. Finally, set compatibility mode DPI scaling to "System" (desired setting)
For me, this only works in fullscreen mode (afaik) and gpu scaling must be turned off in your ATI/AMD/Nvidia graphics settings.
On July 02 2024 18:09 LaStScan wrote:Hi guys, I don't know about the renderer setting + display scale 150%(it did not work for me + starcraft screen becomes blurry) + SD Filters and the others. But I'm here to share my information that I looked up on Korean community. If you are using hardware cursor, this feature ignores game engine's framerate(FPS) and show your cursor frame as your monitor refresh rate. If you are not using this feature, your computer is using GPU register and individually draws/shows your cursor on screen/layer. So people who did not understand my poor explanation, if your computer spec is poor like using 15 year old laptop/desktop or whatever, playing on 60hz monitor with bad graphics card, you will feel the lag Ex) If game client is lagging(whatever the reason is, like UMS too many units) + having a FPS drop a lot = your cursor will lag as well 15 seconds video explanation: www.fmkorea.comYou can also use google translate and read what he has explained there + recorded video.
Thank you for sharing -- updated OP with this great info!!!
On July 06 2024 23:03 fgt4w wrote:
This fixed many issues, thanks! I still have so many issues though that make playing BW painful.
Every time I start SC (from the battle.net launcher usually), I need to log in again from within SC. Everything is laggy, usually takes a few minutes. Once I enter my username/pw, it says "Downloading Settings". Takes a couple minutes, until saying "Download failed." Everything is laggy, takes a minute to click OK, pick a server, and get ready to search for ladder games. Usually takes like 3-4 searches - first few searches end in error (error 50:6 iirc).
Once in game, everything works perfectly, with one exception. If I open menu in-game, it usually lags for like 3-5 seconds. If I need to change latency.
Tried running b.net launcher and starcraft.exe as administrator. Tried reinstalling many, many times. Whitelisting everything starcraft in antivirus (and turning off antivirus completely). Nothing has helped, until following these instructions! Now the lag and wait times have been reduced by about half, but still very annoying. Any help with these issues is greatly appreciated
Can you run dxdiag and copy/paste the first two pages here?
In Windows, select Start and enter dxdiag into the search box on the taskbar. Select dxdiag from the results.
In the tool, select Save All Information... for an easily shared format.
also try these fixes in the video. they seem to have worked for others with similar-sounding problems:
+ Show Spoiler +
|
On July 01 2024 09:54 HaFnium wrote:
Thanks. I think it made the SD graphics look better (more like the pre-SCR ones) at least for me..
In my experience, the SD graphics only look good if you are running at a true 4:3 resolution. It has something todo with the way the scaling is performed.
The graphics look "crisp" at 1280x960 and 1400x1050. BUT if you are running a 16 : 9 monitor resolution with pillarbox turned on, the SD graphics look "blurry" regardless of the SD graphics filtering setting.
The most striking example of this scaling issue is with Zerg which looks AMAZING at a true 4:3 resolution but look like garbage in 16 : 9 with pillarbox.
Zerg @ true 4:3
+ Show Spoiler +
Terran @ true 4:3
+ Show Spoiler +
Protoss @ true 4:3
+ Show Spoiler +
Perhaps the part-time Blizzard Team Intern Janitor Hamster looking after SC:R will do something about it...but i wouldnt hold your breath.
If you are reading this thread and if you want to setup a custom 4 : 3 resolution to replicate the above screenshots, check out Sizer. It is the tool used to setup the game as shown above. I will post an updated version of the guide that covers these settings soon.
|
|
|
Update: renderer presetmon results (windows 10)
Fullscreen
--renderer=legacy has lowest memory footprint, DPC latency, lowest frametimes, cpu/gpu performance, and lowest display latency average 2.5ms
it is 2x-2.5 "faster" than the other renderers (john carmack can flame me but this is approximate truth)
it also looks really ugly...i mean its really really ugly...fast and responsive! but ugly
--renderer=opengl is slightly slower but still 1.2-1.5x faster than default (average 3.9ms) and it still looks pretty good
the default (dx10) renderer takes 3rd place (avg 4.7ms)
dx11 and dx12 are actually slower than the default (5.3-5.7ms)
Windowed (Fullscreen)
--renderer=dx9 has lowest memory footprint, DPC latency, lowest frametimes, cpu/gpu performance, and lowest display latency
as mentioned by Tec27:
if you are using Windowed or windowed(fullscreen) mode, don't use --legacy or --opengl because it will require a gpu buffer copy (bad)
at the end of the day, does it matter? will it make you into FlaSh? probably not, but if you hate the laggy feel then you can improve by playing with Fullscreen Mode only and use --opengl or --legacy renderer.
keyboard warriors who come into this thread talking about "blah blah you cant tell difference hurr durr i have PhD in physiology from NASA and am 9000 ladder points": if you dont dont care, or dont feel difference, dont use. if you do feel difference, do use. world is your oyster. =)
if someone wants the raw data they are here:
https://www.mediafire.com/file/nnnvgxzlbx34onl/PresetMon.zip/file
https://www.virustotal.com/gui/url/f5bfe0d432a8d6a4b763f49611b32101701fd45b315d865333957a1b09e5119d
gl hf gg =)
|
Interesting update thank you.
I have a question for you tankgirl, does any pro play at more than 1080p (say 1440?) and if so do the settings vary significantly?
|
On March 29 2025 20:29 tankgirl wrote:Update: renderer presetmon results (windows 10) Fullscreen --renderer=legacy has lowest memory footprint, DPC latency, lowest frametimes, cpu/gpu performance, and lowest display latency average 2.5ms it is 2x-2.5 "faster" than the other renderers (john carmack can flame me but this is approximate truth) it also looks really ugly...i mean its really really ugly...fast and responsive! but ugly --renderer=opengl is slightly slower but still 1.2-1.5x faster than default (average 3.9ms) and it still looks pretty good the default (dx10) renderer takes 3rd place (avg 4.7ms) dx11 and dx12 are actually slower than the default (5.3-5.7ms) Windowed (Fullscreen) --renderer=dx9 has lowest memory footprint, DPC latency, lowest frametimes, cpu/gpu performance, and lowest display latency as mentioned by Tec27:if you are using Windowed or windowed(fullscreen) mode, don't use --legacy or --opengl because it will require a gpu buffer copy (bad) at the end of the day, does it matter? will it make you into FlaSh? probably not, but if you hate the laggy feel then you can improve by playing with Fullscreen Mode only and use --opengl or --legacy renderer. keyboard warriors who come into this thread talking about "blah blah you cant tell difference hurr durr i have PhD in physiology from NASA and am 9000 ladder points": if you dont dont care, or dont feel difference, dont use. if you do feel difference, do use. world is your oyster. =) if someone wants the raw data they are here: https://www.mediafire.com/file/nnnvgxzlbx34onl/PresetMon.zip/filehttps://www.virustotal.com/gui/url/f5bfe0d432a8d6a4b763f49611b32101701fd45b315d865333957a1b09e5119dgl hf gg =)
On my machine, I see identical latencies between --renderer=legacy exclusive fullscreen and the default renderer with borderless fullscreen. Are you getting higher than the 300 fps limit in the legacy mode? Knowing what I know about the code I don't think that should be possible, but your data shows otherwise (1ms frame times, e.g. ~= 1000 fps). If I disable the frame cap setting while in legacy rendering mode it seems to set a 100 fps cap in menus (and the latencies increase accordingly) and 300 fps ingame, so I'm not sure what is different in your setup. Do you only have a single monitor?
Modern games typically try to avoid rendering so many extra frames and instead process input + render a frame as close to the display refresh time as possible (this is what "frame pacing" is) but Blizzard's updated graphics code doesn't do anything like that, so the way to decrease latency is to just render as many frames as possible. Thus, if you can get past the 300 fps limit, you'll likely see a reduction in latency, but I don't see how that is possible. Would definitely be interested in what is different about your hardware/software setup that makes those reduced frame times happen.
|
On March 29 2025 20:29 tankgirl wrote:Update: renderer presetmon results (windows 10) Fullscreen --renderer=legacy has lowest memory footprint, DPC latency, lowest frametimes, cpu/gpu performance, and lowest display latency average 2.5ms it is 2x-2.5 "faster" than the other renderers (john carmack can flame me but this is approximate truth) it also looks really ugly...i mean its really really ugly...fast and responsive! but ugly --renderer=opengl is slightly slower but still 1.2-1.5x faster than default (average 3.9ms) and it still looks pretty good the default (dx10) renderer takes 3rd place (avg 4.7ms) dx11 and dx12 are actually slower than the default (5.3-5.7ms) Windowed (Fullscreen) --renderer=dx9 has lowest memory footprint, DPC latency, lowest frametimes, cpu/gpu performance, and lowest display latency as mentioned by Tec27:if you are using Windowed or windowed(fullscreen) mode, don't use --legacy or --opengl because it will require a gpu buffer copy (bad) at the end of the day, does it matter? will it make you into FlaSh? probably not, but if you hate the laggy feel then you can improve by playing with Fullscreen Mode only and use --opengl or --legacy renderer. keyboard warriors who come into this thread talking about "blah blah you cant tell difference hurr durr i have PhD in physiology from NASA and am 9000 ladder points": if you dont dont care, or dont feel difference, dont use. if you do feel difference, do use. world is your oyster. =) if someone wants the raw data they are here: https://www.mediafire.com/file/nnnvgxzlbx34onl/PresetMon.zip/filehttps://www.virustotal.com/gui/url/f5bfe0d432a8d6a4b763f49611b32101701fd45b315d865333957a1b09e5119dgl hf gg =)
thank you for the deep dive. going to check what I am on when I get home.
|
|
|
on top of my previous question regarding resolution (1440p vs 1080p), i have another one: what is the benefit, if any, of pillarbox? it just makes it square so everything is centered and feels like the original but in HD? but then you lose the extra lateral view depth so in effect you see less of the map? Why would pro pick this?
|
On April 02 2025 09:00 WGT-Baal wrote:
on top of my previous question regarding resolution (1440p vs 1080p), i have another one: what is the benefit, if any, of pillarbox? it just makes it square so everything is centered and feels like the original but in HD? but then you lose the extra lateral view depth so in effect you see less of the map? Why would pro pick this?
Mini map is closer to the center of the screen with pillarbox, making it easier to look at mini map quickly in theory. on fullscreen its further to the left. there is an option to move minimap to the right a bit with fullscreen but that creates an open space to the left of the minimap, which means the mouse can overshoot the minimap more than on pillarbox with a quick move into minimap, which some pros use for fast screen change.
|
On April 02 2025 09:00 WGT-Baal wrote:
on top of my previous question regarding resolution (1440p vs 1080p), i have another one: what is the benefit, if any, of pillarbox? it just makes it square so everything is centered and feels like the original but in HD? but then you lose the extra lateral view depth so in effect you see less of the map? Why would pro pick this?
As of September 2024 Mini finally switched to Fullscreen 16/9 (not pillarbox / 4:3) aspect ratio. He had long-been a 4:3 user.
The only remaining (competitive) 4:3 players are Bisu, Mind, JyJ, Jaedong, Rain and Flash.
Jaedong experimented with wide mode for several months a few years ago but eventually switched back to pillarbox 4:3.
The most obvious advantage is in TvT where siege tank range plays such a huge role. As seen in this recent Flash vs TY game, it doesn't seem to be a problem for Flash. JyJ also has decent TvT record.
I suspect that for the highly "automated" players especially Bisu/Flash/Jaedong, it mostly boils down to muscle memory?
On March 30 2025 07:22 tec27 wrote:
On my machine, I see identical latencies between --renderer=legacy exclusive fullscreen and the default renderer with borderless fullscreen. Are you getting higher than the 300 fps limit in the legacy mode? Knowing what I know about the code I don't think that should be possible, but your data shows otherwise (1ms frame times, e.g. ~= 1000 fps). If I disable the frame cap setting while in legacy rendering mode it seems to set a 100 fps cap in menus (and the latencies increase accordingly) and 300 fps ingame, so I'm not sure what is different in your setup. Do you only have a single monitor?
Modern games typically try to avoid rendering so many extra frames and instead process input + render a frame as close to the display refresh time as possible (this is what "frame pacing" is) but Blizzard's updated graphics code doesn't do anything like that, so the way to decrease latency is to just render as many frames as possible. Thus, if you can get past the 300 fps limit, you'll likely see a reduction in latency, but I don't see how that is possible. Would definitely be interested in what is different about your hardware/software setup that makes those reduced frame times happen.
I discovered a method to disable the arbitrarily imposed 300FPS limit, which I did for testing. I dont know if I want to share the method publically at this time because I don't know the consequences and/or player advantage for anyone not using it. I will send a PM with instructions on how to disable. Perhaps I should re-do the tests with the standard FPS limit because what you are saying suggests it wont actually make any difference at 100-300 FPS? Please advise. I wanted to test the racecars at the racetrack, not the grocery store, if that makes sense...
On March 30 2025 01:03 WGT-Baal wrote:
Does any pro play at more than 1080p (say 1440?) and if so do the settings vary significantly?
I don't believe so. The settings do vary. The main problem with higher resolutions is the cursor size scales down accordingly, meaning you cant play with hardware cursor enabled.
In fact, when Flash returned to streaming late last year, there were some humorous discussions about "progamer computer illiteracy" on his stream because he had a 4K 2nd-monitor setup and he was driven insane by the tiny size of the hardware cursor on his 4K monitor for a few days. Users in his streamchat had to guide him through how to setup "primary monitor" in the windows settings. Personally, I think its OK to dedicate yourself to being the best at one single thing and be unskilled at others, as seems to be the case with FlaSh.
Also the ASL studio uses a relatively "standard" 144hz 24" 1080p LG Ultragear monitor. Serious player probably attempt to replicate the studio setup at home? Light for example uses a BenQ XL2411P. But I dont know to what extent this is universally the case.
|
On April 02 2025 10:43 tankgirl wrote:Show nested quote +On April 02 2025 09:00 WGT-Baal wrote:
on top of my previous question regarding resolution (1440p vs 1080p), i have another one: what is the benefit, if any, of pillarbox? it just makes it square so everything is centered and feels like the original but in HD? but then you lose the extra lateral view depth so in effect you see less of the map? Why would pro pick this? As of September 2024 Mini finally switched to Fullscreen 16/9 (not pillarbox / 4:3) aspect ratio. He had long-been a 4:3 user. The only remaining (competitive) 4:3 players are Bisu, Mind, JyJ, Jaedong, Rain and Flash. Jaedong experimented with wide mode for several months a few years ago but eventually switched back to pillarbox 4:3. The most obvious advantage is in TvT where siege tank range plays such a huge role. As seen in this recent Flash vs TY game, it doesn't seem to be a problem for Flash. JyJ also has decent TvT record. I suspect that for the highly "automated" players especially Bisu/Flash/Jaedong, it mostly boils down to muscle memory? Show nested quote +On March 30 2025 07:22 tec27 wrote:
On my machine, I see identical latencies between --renderer=legacy exclusive fullscreen and the default renderer with borderless fullscreen. Are you getting higher than the 300 fps limit in the legacy mode? Knowing what I know about the code I don't think that should be possible, but your data shows otherwise (1ms frame times, e.g. ~= 1000 fps). If I disable the frame cap setting while in legacy rendering mode it seems to set a 100 fps cap in menus (and the latencies increase accordingly) and 300 fps ingame, so I'm not sure what is different in your setup. Do you only have a single monitor?
Modern games typically try to avoid rendering so many extra frames and instead process input + render a frame as close to the display refresh time as possible (this is what "frame pacing" is) but Blizzard's updated graphics code doesn't do anything like that, so the way to decrease latency is to just render as many frames as possible. Thus, if you can get past the 300 fps limit, you'll likely see a reduction in latency, but I don't see how that is possible. Would definitely be interested in what is different about your hardware/software setup that makes those reduced frame times happen. I discovered a method to disable the arbitrarily imposed 300FPS limit, which I did for testing. I dont know if I want to share the method publically at this time because I don't know the consequences and/or player advantage for anyone not using it. I will send a PM with instructions on how to disable. Perhaps I should re-do the tests with the standard FPS limit because what you are saying suggests it wont actually make any difference at 100-300 FPS? Please advise. I wanted to test the racecars at the racetrack, not the grocery store, if that makes sense... Show nested quote +On March 30 2025 01:03 WGT-Baal wrote:
Does any pro play at more than 1080p (say 1440?) and if so do the settings vary significantly? I don't believe so. The settings do vary. The main problem with higher resolutions is the cursor size scales down accordingly, meaning you cant play with hardware cursor enabled. In fact, when Flash returned to streaming late last year, there were some humorous discussions about "progamer computer illiteracy" on his stream because he had a 4K 2nd-monitor setup and he was driven insane by the tiny size of the hardware cursor on his 4K monitor for a few days. Users in his streamchat had to guide him through how to setup "primary monitor" in the windows settings. Personally, I think its OK to dedicate yourself to being the best at one single thing and be unskilled at others, as seems to be the case with FlaSh. Also the ASL studio uses a relatively "standard" 144hz 24" 1080p LG Ultragear monitor. Serious player probably attempt to replicate the studio setup at home? Light for example uses a BenQ XL2411P. But I dont know to what extent this is universally the case.
Thank you for the details!
For players not being very computer savvy, it is indeed ironic and I experienced it firsthand in LANs, both with foreigners and Koreans. From everything to lan and network settings to mouse, keyboard language and windows language settings. I know my kespa friends were always somewhat joking that it was half the job to help tje proteams set up correctly
|
On April 02 2025 10:43 tankgirl wrote:Show nested quote +On April 02 2025 09:00 WGT-Baal wrote:
on top of my previous question regarding resolution (1440p vs 1080p), i have another one: what is the benefit, if any, of pillarbox? it just makes it square so everything is centered and feels like the original but in HD? but then you lose the extra lateral view depth so in effect you see less of the map? Why would pro pick this? As of September 2024 Mini finally switched to Fullscreen 16/9 (not pillarbox / 4:3) aspect ratio. He had long-been a 4:3 user. The only remaining (competitive) 4:3 players are Bisu, Mind, JyJ, Jaedong, Rain and Flash. Jaedong experimented with wide mode for several months a few years ago but eventually switched back to pillarbox 4:3. The most obvious advantage is in TvT where siege tank range plays such a huge role. As seen in this recent Flash vs TY game, it doesn't seem to be a problem for Flash. JyJ also has decent TvT record. I suspect that for the highly "automated" players especially Bisu/Flash/Jaedong, it mostly boils down to muscle memory? Show nested quote +On March 30 2025 07:22 tec27 wrote:
On my machine, I see identical latencies between --renderer=legacy exclusive fullscreen and the default renderer with borderless fullscreen. Are you getting higher than the 300 fps limit in the legacy mode? Knowing what I know about the code I don't think that should be possible, but your data shows otherwise (1ms frame times, e.g. ~= 1000 fps). If I disable the frame cap setting while in legacy rendering mode it seems to set a 100 fps cap in menus (and the latencies increase accordingly) and 300 fps ingame, so I'm not sure what is different in your setup. Do you only have a single monitor?
Modern games typically try to avoid rendering so many extra frames and instead process input + render a frame as close to the display refresh time as possible (this is what "frame pacing" is) but Blizzard's updated graphics code doesn't do anything like that, so the way to decrease latency is to just render as many frames as possible. Thus, if you can get past the 300 fps limit, you'll likely see a reduction in latency, but I don't see how that is possible. Would definitely be interested in what is different about your hardware/software setup that makes those reduced frame times happen. I discovered a method to disable the arbitrarily imposed 300FPS limit, which I did for testing. I dont know if I want to share the method publically at this time because I don't know the consequences and/or player advantage for anyone not using it. I will send a PM with instructions on how to disable. Perhaps I should re-do the tests with the standard FPS limit because what you are saying suggests it wont actually make any difference at 100-300 FPS? Please advise. I wanted to test the racecars at the racetrack, not the grocery store, if that makes sense... Show nested quote +On March 30 2025 01:03 WGT-Baal wrote:
Does any pro play at more than 1080p (say 1440?) and if so do the settings vary significantly? I don't believe so. The settings do vary. The main problem with higher resolutions is the cursor size scales down accordingly, meaning you cant play with hardware cursor enabled. In fact, when Flash returned to streaming late last year, there were some humorous discussions about "progamer computer illiteracy" on his stream because he had a 4K 2nd-monitor setup and he was driven insane by the tiny size of the hardware cursor on his 4K monitor for a few days. Users in his streamchat had to guide him through how to setup "primary monitor" in the windows settings. Personally, I think its OK to dedicate yourself to being the best at one single thing and be unskilled at others, as seems to be the case with FlaSh. Also the ASL studio uses a relatively "standard" 144hz 24" 1080p LG Ultragear monitor. Serious player probably attempt to replicate the studio setup at home? Light for example uses a BenQ XL2411P. But I dont know to what extent this is universally the case.
I'm fairly certain if you test and keep the 300 FPS limit you'll see the same thing as me (same latency across the different renderers). Without it, the legacy ones probably *do* have less latency as your machine is *likely* able to render more frames in those modes (although honestly, I'm not entirely sure of that, more recent graphics APIs may be more efficient/better mapped to modern hardware). The ideal is probably to disable the FPS limit across all the renderers if you can. If you do want to DM me the method it might be a setting I could add to SB for people that want it.
As far as the mouse cursor resizing, I'm honestly not totally sure what causes that and this is something I'd like to fix for SB. It's weird because they're using the same API to change between all the different cursors, but one of them is the correct size and all the other ones are small. Likely something with how they're loading the cursors but I need to look into it further.
|
Update 04/06/2025: Bisu's settings
OK so Bisu is doing something weird. He plays with hardware cursor enabled, yet it is not scaled to the regular size expected.
From comparing with different display resolutions and observing his VODs, it appears he manual sets up 960x720 resolution in Windowed mode. The 960x720 resolution would explain the "blurry" look that his Starcraft client has in-game, even though he is streaming at 1080p 8000kpbs.
I wonder if this strange setup at home has something to do with his less-than-stellar performance in the ASL studio, where he might not be able to replicate his at-home settings perfectly (c.f. flash's ruler)
Anyway:
Mouse: : Logitech Mini Optical (implanted Logitech G303 sensor) (DPI 1050)
Keyboard: QSENN DT35
Ingame:
+ Show Spoiler [Speed] +
+ Show Spoiler [Video] +Unknown, likely Windowed @ 960x720 or 1280x720 pillarbox
@tec27 PM'd
|
|
|
|
|
|
|
|
EPT![[image loading]](https://stimg.sooplive.co.kr/PHOTO_BBS/6/13745176/1535259702269.jpg)
![[image loading]](https://i.imgur.com/6ZBUfiX.jpeg)
![[image loading]](https://i.imgur.com/NAo8XFj.jpeg)
![[image loading]](https://i.imgur.com/L6TL8fq.jpeg)
![[image loading]](https://i.imgur.com/SxdQP80.jpeg)
![[image loading]](https://i.imgur.com/N6wj2CW.jpeg)
![[image loading]](https://i.imgur.com/4jvlEw8.png)
![[image loading]](https://i.imgur.com/fHVX4A7.png)
![[image loading]](https://i.imgur.com/YGMi4mi.png)
![[image loading]](https://i.imgur.com/ug7g7Dw.jpeg)
![[image loading]](https://i.imgur.com/y7J61IL.jpeg)
![[image loading]](https://i.imgur.com/nsKQA7q.jpeg)
![[image loading]](https://i.imgur.com/SPa8uu2.jpeg)
![[image loading]](https://i.imgur.com/uJa63eh.jpeg)
![[image loading]](https://i.imgur.com/msaAn2D.jpeg)
![[image loading]](https://i.imgur.com/UqhDOj7.jpeg)
![[image loading]](https://i.imgur.com/7Pz22ok.jpeg)
![[image loading]](https://i.imgur.com/53HnUO5.jpeg)
![[image loading]](https://i.imgur.com/bnlCOSC.jpeg)
![[image loading]](https://i.imgur.com/HjXJLcv.jpeg)
![[image loading]](https://i.imgur.com/j5RvMBH.png)
![[image loading]](https://i.imgur.com/Toodndt.png)
![[image loading]](https://i.imgur.com/GNNisYK.png)
![[image loading]](https://i.imgur.com/4hrbDug.png)
