Jump to content

Welcome to MSFN Forum
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. This message will be removed once you have signed in.
Login to Account Create an Account



Photo

Strange mouse behavior on an LCD monitor

- - - - -

  • Please log in to reply
22 replies to this topic

#1
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

After having waited long enough for LCD technology to improve to an acceptable level of performance, I finally upgraded my Win98SE computer's monitor from a CRT to a Dell U2413 professional level IPS. I compared the latency between both monitors side by side, and the LCD lags a bit as expected. However, there was an unexpected weird behavior with the mouse pointer on the LCD. There's the expected slight lag, but moving it around the screen, it appeared to have a floating feeling, as if a smoothing filter was applied to the mouse movement. No such option exist in the mouse settings. Mouse acceleration is off. If I click and drag a window around the screen, the pointer does not exactly follow the window where one would expect the pointer to be. Same thing happens with scrollbars, or dragging anything around which provide an object of reference.

I though maybe it was the input lag. I use IPS and TN monitors all the time, at work, in school, and at home. I have a cheap TN that is connected to another computer and an LCD TV connected to my brother's computer. LCD TVs are notorious for terrible input lag, yet there was no floating mouse phenomenon. This only happens on my Win98SE computer with this monitor.

To eliminate some factors, production models of the U2413 from the first two years had problems with ghosting and overshoot, which has been resolved in the latest revision, and is unrelated to this problem. The U2413 has a game mode that makes it one of the fastest IPS monitors, making the display lag performance as Class 1 in an in-depth review by TFT Central. Activating game mode doesn't fix it. Tried swapping different mice, optical or laser, high or low DPI, PS/2 or USB, and it's still there.

Searching for a solution, I found something that might explain this behavior. According to this condition, the GPU renders the mouse pointer separately from the rest of the screen. It is usually not apparent in CRTs, but sometimes, the floating mouse behavior appears on an LCD. Some users experiencing this issue reported it to be resolved if the hardware acceleration slider is moved one step to the left, while others needed to completely turn it off. I tried it, and it did not help. Anyone recognize this problem?

My configuration:
PIII 1Ghz
Windows 98SE
NVIDIA GeForce4 Ti 4200 with 56.64 drivers
NUSB 3.6e
KernelEx 4.5.2




How to remove advertisement from MSFN

#2
Drugwash

Drugwash

    MSFN Expert

  • Member
  • PipPipPipPipPipPip
  • 1,448 posts
  • Joined 21-June 06
  • OS:98SE
  • Country: Country Flag

Maybe related or maybe not: what brand exactly is your videocard? I'm asking because I have such card myself, the Ti4200 AGP8X model made by MSI and the latest driver I could use is the MSI-provided 45.32 from their site. Any official NVIDIA driver later than that would completely disable AGP texture (check in DxDiag if yours is enabled).

It's possible that this particular card model and/or particular manufacturers may require specific drivers to function properly. Therefore my advice would be for you to check if there are any manufacturer-specific drivers for your card model at the manufacturer's site and perform a test with those. Or if you have the original drivers CD you could try those too, if only to rule out this possible issue.

 

I'm using my card with the MSI 45.32 drivers on an 667MHz Pentium III (+KernelEx 4.5.2, AutoPatcher98 with updated DirectX 9.0c, RP9.7.2 and many other updates) on a Yakumo 17XF8 TFT monitor and there is no lag whatsoever, the cursor is stuck where it should be when dragging windows around. Granted it's a VGA connection, not DVI.



#3
MrMateczko

MrMateczko

    Junior

  • Member
  • Pip
  • 65 posts
  • Joined 29-June 13
  • OS:none specified
  • Country: Country Flag
Try different graphic drivers, or if you can, try different graphic cards.

About the AGP Texture Acceleration in DxDiag: http://www.intel.com...b/CS-009689.htm and https://support.micr...en-us/kb/810772
It seems it's not needed at all, and is just for the show.

Edited by MrMateczko, 15 August 2015 - 03:26 AM.


#4
Drugwash

Drugwash

    MSFN Expert

  • Member
  • PipPipPipPipPipPip
  • 1,448 posts
  • Joined 21-June 06
  • OS:98SE
  • Country: Country Flag

Well, M$ as always have been trying to minimalize things (and Intel follows suit). At the time, I was trying to update the video driver for my card in order to be able to play one of the 'Tomb raider' series games and that game was unable to detect available resolutions/color depth and other video-related options because of misfit drivers. So the driver does matter a lot.

Anyway, I offered this personal example more on the lines of "if that visible issue happened, who knows what other hidden issues may arise with a misfit driver". Mouse lag included. So certain tests are required in order to narrow down or pinpoint the issue.

 

Nonetheless, thank you for taking the time to search and publish the links above.

 

(fingers pressing buttons when brains are not thereā€¦)


Edited by Drugwash, 15 August 2015 - 04:50 AM.


#5
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

Thanks for the replies.

 

Checked DxDiag and AGP Texture Acceleration is enabled. I have DirectX 9.0c, Dec. 2006 release.

 

I don't have a spare graphics card available, and there's no integrated graphics on the mobo, so no hardware swapping can be done.

 

My card is an ASUS AGP4X using the reference drivers. I'll try out the ASUS drivers.

 



#6
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

Installed and tested the ASUS drivers. No improvement on the mouse problem. The drivers were a little bloated, so I went back to the reference drivers.

 

Might have to let this sit for a little bit and come back to it when I have more time.



#7
Drugwash

Drugwash

    MSFN Expert

  • Member
  • PipPipPipPipPipPip
  • 1,448 posts
  • Joined 21-June 06
  • OS:98SE
  • Country: Country Flag

You may have to edit and install the drivers provided by Dell. And while at that, tell the id!ots at Dell that "installation" (in Source Disk Names section) has two As so they should better revise their inf files in the package.

Honestly, I wouldn't buy Dell - all my experience with such hardware was at least frustrating but mostly aggravating. And limiting a MONITOR (!!!) usage to a narrow range of operating systems is simply moronic.



#8
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

I called Dell tech support. They searched through their knowledge base and haven't found anything related to this. All they could come up with was some kind of compatibility problem between the OS and monitor. Due a lack of data for this monitor and Win98SE, they couldn't pin it on anything else.

 

All that's left to do is to try and see if I could modify the drivers to work with Win98SE, if I can get it work at all. I am using the monitor on Windows 7 without any drivers, and it doesn't have that problem.



#9
Drugwash

Drugwash

    MSFN Expert

  • Member
  • PipPipPipPipPipPip
  • 1,448 posts
  • Joined 21-June 06
  • OS:98SE
  • Country: Country Flag

If you're using the DVI connection it's possible that the EDID is not correctly read or processed by 98SE. I've had a hard time getting a Hanns-G Hi221 monitor work at native resolution (1680x1050) even under XP-SP3 because of this and had to create custom resolutions in NVIDIA's control panel despite having installed the very-hard-to-find driver, so installing a driver may or may not work for you but you'll have to try this just to be sure. Just change signature to "$CHICAGO$", comment out the services sections as they don't exist in 9x and you're all set.

 

I believe there are tools that can modify a monitor's EDID data through reflashing but it's a complicated and dangerous operation so I wouldn't advise to try that unless you can afford to lose the monitor.



#10
jaclaz

jaclaz

    The Finder

  • Developer
  • 15,997 posts
  • Joined 23-July 04
  • OS:none specified
  • Country: Country Flag

EDID related (maybe useful, maybe not) :unsure:

http://www.msfn.org/...-blurry-screen/

 

jaclaz



#11
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

Thanks for the ideas. Every little bit helps. I haven't had any success in getting the system to recognize the modifed inf. It probably wouldn't do much anyway since Windows is already aware that a Dell U2413 is connected, and recognized its native resolution without the drivers.

 

There's a possiblity that it could be somewhat EDID related since I am not connected via VGA, but instead of the OS, it could be the monitor not getting the return signal. The U2413 is completely digital and does not include VGA or DVI-A connections. Win98SE is connected via DVI-D, Win7 via Display Port, and changed via an input selector. I've done some more testing, and found what could be a possible explanation why the problem occurs. If I boot up Win98SE system first, the floating mouse behavior appears on the monitor. If I later boot up the Win7 system and switch input to it, Win7 inherits the problem, and remains that way. However, if I boot Win7 first, there is no floating mouse problem, and if I later boot up and switch to the Win98SE system, it stays problem free there too. Note that power must be completely cut to the monitor at the surge protector in order for this to work (a monitor cold boot if you will), otherwise, the monitor will remember the inherited problem from a previous Win98SE boot and keep it that way.

 

I think the monitor is not receiving the proper communication from Win98SE on DVI, and thus, it operates in some form of compatibility mode. When Win7 sends the proper signal to the monitor from a monitor cold boot, everyting is all fine and dandy if I switch over to Win98SE. However, booting up another system if I only want to use 98 is just silly. While this is not a solution, maybe it can lead to one.



#12
jumper

jumper

    2015 All-American Masters HJ'er

  • Member
  • PipPipPipPip
  • 579 posts
  • Joined 21-January 11
  • OS:98SE
  • Country: Country Flag
Connect the Win7 system via DVI (and the SE via Display Port if possible).
Design feedback requested:
KernelEx 4.5.2015
IHAtool - IpHlpApi tester; call various functions and report results
--status-> framework is solid; 22 api's fully supported; preview release coming soon
Future projects: Kexter - IP40+Ktree+Kexstubs

#13
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

Win98SE is using a GeForce4 Ti 4200, which only has VGA and DVI connections. DVI is the only choice here.



#14
jumper

jumper

    2015 All-American Masters HJ'er

  • Member
  • PipPipPipPip
  • 579 posts
  • Joined 21-January 11
  • OS:98SE
  • Country: Country Flag
(So a Display Port to DVI or VGA adapter would be needed, but....)
The point was to connect the Win7 system to the monitor via DVI to see if it's a problem with the DVI input on the monitor.
Design feedback requested:
KernelEx 4.5.2015
IHAtool - IpHlpApi tester; call various functions and report results
--status-> framework is solid; 22 api's fully supported; preview release coming soon
Future projects: Kexter - IP40+Ktree+Kexstubs

#15
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

Checked Win7 with DVI. No problem there.



#16
jumper

jumper

    2015 All-American Masters HJ'er

  • Member
  • PipPipPipPip
  • 579 posts
  • Joined 21-January 11
  • OS:98SE
  • Country: Country Flag
What video card is in the Win7 system? You might need a similar one in SE.
Design feedback requested:
KernelEx 4.5.2015
IHAtool - IpHlpApi tester; call various functions and report results
--status-> framework is solid; 22 api's fully supported; preview release coming soon
Future projects: Kexter - IP40+Ktree+Kexstubs

#17
MrMateczko

MrMateczko

    Junior

  • Member
  • Pip
  • 65 posts
  • Joined 29-June 13
  • OS:none specified
  • Country: Country Flag
Hmm...I just had an idea, try using VBEMP, or revert back to standard VGA 640x480.

#18
Drugwash

Drugwash

    MSFN Expert

  • Member
  • PipPipPipPipPipPip
  • 1,448 posts
  • Joined 21-June 06
  • OS:98SE
  • Country: Country Flag

To test:

1. Modify original monitor driver and install. Reboot.

2. Create and apply custom resolution(s) such as 1920x1199 (one pixel difference) that can override the EDID. Reboot.

3. Replace the monitor with a more resiliant one. Enjoy. :)


  • jumper likes this

#19
Foxbat

Foxbat

    Member

  • Member
  • PipPip
  • 130 posts
  • Joined 18-January 11
  • OS:none specified
  • Country: Country Flag

What video card is in the Win7 system? You might need a similar one in SE.

MSI GeForce GTX 970. Too recent, not gonna work. My 98 system doesn't have PCIe.

 

Hmm...I just had an idea, try using VBEMP, or revert back to standard VGA 640x480.

VBEMP provides no harware acceleration. I need the system to be fully functional with nothing nerfed, otherwise, it would defeat the purpose of using 98.

 

To test:

1. Modify original monitor driver and install. Reboot.

2. Create and apply custom resolution(s) such as 1920x1199 (one pixel difference) that can override the EDID. Reboot.

3. Replace the monitor with a more resiliant one. Enjoy. :)

1. I couldn't get my modified drivers to work.

2. Haven't had much luck in messing with custom resolutions in the past. I'll give it another go this time around.

3. The next alternative monitor that fits my specs start in the $1000 range, so I bought the $500 Dell monitor. It has to be one single monitor to share with 98 and 7. In order to meet the requirements of my line of work, I had to get rid of my beloved CRT and deal with some pet peeves of an LCD. Works well with 7, but not optimal for 98. It's either one or the other. Can't have both--everything's a compromise.


Edited by Foxbat, Yesterday, 08:30 PM.


#20
jumper

jumper

    2015 All-American Masters HJ'er

  • Member
  • PipPipPipPip
  • 579 posts
  • Joined 21-January 11
  • OS:98SE
  • Country: Country Flag
Both video cards are GeForce, but different chipsets....
Contact Asus and MSI concerning DVI-D compatibility with the two different GeForce chipsets. The GeForce4 might not fully support all features of the latest version of DVI-D. If this is a known issue, their engineers (of either company) might have a (unsupported) software utility for you to try.

As for a workaround, once the monitor has been booted with Win7 and is working correctly, don't power it off! Let it hibernate when the PC's are off.
Design feedback requested:
KernelEx 4.5.2015
IHAtool - IpHlpApi tester; call various functions and report results
--status-> framework is solid; 22 api's fully supported; preview release coming soon
Future projects: Kexter - IP40+Ktree+Kexstubs

#21
Drugwash

Drugwash

    MSFN Expert

  • Member
  • PipPipPipPipPipPip
  • 1,448 posts
  • Joined 21-June 06
  • OS:98SE
  • Country: Country Flag

@ Foxbat:

It's just a hunch but the monitor may use EDID extensions which may not be recognized by the 98 driver.

I have a very raw and incomplete script that reads EDID data from the registry and another one that dumps the registry data to text files. Comparing the registry data between 98 and 7 might shed some light (or not). I could compile and upload these if you want. Done, check my repository, the EDID folder.

 

What exactly did not work with modified drivers: wouldn't install at all or got installed but wouldn't make any difference?

Overriding EDID with custom resolutions  is a bit different in 9x than in NT-based Windows. I couldn't fully understand it myself and didn't pursue the matter further after resolving the issue with Hanns-G.

 

Thing is, if EDID extensions are the problem you may never get to fix it under 9x unless you find a driver that can cope with them.


Edited by Drugwash, Today, 06:16 AM.


#22
jaclaz

jaclaz

    The Finder

  • Developer
  • 15,997 posts
  • Joined 23-July 04
  • OS:none specified
  • Country: Country Flag

Again maybe useful, maybe not:

https://archive.org/...x_EDID_Designer

http://peteprlm.hubp...-corrupted-edid

 

jaclaz



#23
Drugwash

Drugwash

    MSFN Expert

  • Member
  • PipPipPipPipPipPip
  • 1,448 posts
  • Joined 21-June 06
  • OS:98SE
  • Country: Country Flag

It's precisely because of the non-working Phoenix tool that I stepped up to create my CROWE. ;)

Phoenix just freezes in my 98SE when trying to get the data from registry (no matter which KernelEx compatibility I set it to) so it's practically useless on a 9x system.

As for edid-rw, that would eventually be useful for EDID data correction and reflash, but who would go to the lengths of installing and configuring a Linux/Ubuntu machine, then reading EDID data, editing with a hex editor - wouldn't it be better doing it directly in a humanly-readable interface? - and finally writing it back to the monitor on the I2C interface?

 

Anyway, your efforts are deeply appreciated and it's for good reason that you're entitled as The Finder. :)






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users