Damn forced upgrade to Win7!



  • So finally, with much kicking and screaming, I'm being forced to migrate from XP to Win7 (64 bit) by IT because of XP EOL.

    I get my shiny new computer (Dell E6540 for anyone that cares) and it's replete with WTFs, some of which I blame on Windows, some I blame on people cutting costs, and others I don't even...

    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!

    2. How many wireless drivers do I need anyway? Dell DW WLAN with an icon from 1985. Intel PROSet. And Windows native. Seriously folks...

    3. The computer won't wake from sleep. Initially it woke up and then every app threw an error dialog about inability to write to some memory address due to an I/O error. Forms suggest updating the video drivers. Oh, did I mention? This laptop has two video chips - Intel HD and AMD Radeon (WTF?). So I updated the drivers. Try to wake from sleep... black screen, then the system hard reboots itself. Ok, so, try again... delete the AMD Radeon drivers and... now "sleep" isn't even available on the machine. More forum searching... "Sleep may not be available on your system unless you install the correct video drivers." WTF!?! WHY DOES SLEEP HAVE ANY DEPENDENCE ON VIDEO DRIVERS!?!?!

    4. Some of my customer-specific apps don't work on Win7 64 bit. Third party, not mine, I know, I know. (Best part? The app was written in .Net...)

    To top it off, the guy at the desk next to me also got a new computer today, same make/model, but his sleeps just fine. So I probably have bad hardware....



  • @too_many_usernames said:

    This laptop has two video chips
    Par for the course these days, use Intel when you need to save power and the dedicated one when you need graphics power.
    @too_many_usernames said:
    Intel HD and AMD Radeon (WTF?)
    And here's the problem... the AMD implementation of it is nowhere near as smooth as the nVidia one, which Just Works and you can mix and match drivers (I have a newer Intel and nVidia drivers than what Lenovo provides in their Optimus driver package).



  • @Douglasac said:

    Par for the course these days, use Intel when you need to save power and the dedicated one when you need graphics power.

    Still a WTF in my mind... why don't nVidia and AMD have a low power mode so two chips aren't required?

    And yes, one of my other colleagues has a similar laptop with nVidia not AMD and his just works... actually, I ought to check the computer of the guy sitting next to me and see if he's got nVidia or AMD... I also agree that I'd now much rather have nVidia than AMD....



  • @too_many_usernames said:

    2. How many wireless drivers do I need anyway? Dell DW WLAN with an icon from 1985. Intel PROSet. And Windows native. Seriously folks...

    You need one. Any others you should remove. (Pro-tip: keep the one Windows Update installs for you.)

    @too_many_usernames said:

    This laptop has two video chips - Intel HD and AMD Radeon (WTF?).

    Wow, welcome to 4 years ago. You have been in the time-pod for a long time, haven't you?

    @too_many_usernames said:

    WTF!?! WHY DOES SLEEP HAVE ANY DEPENDENCE ON VIDEO DRIVERS!?!?!

    Because video cards have sleep mode? Just a wacky wild guess there.

    @too_many_usernames said:

    4. Some of my customer-specific apps don't work on Win7 64 bit. Third party, not mine, I know, I know. (Best part? The app was written in .Net...)

    In what way do they not work?

    @too_many_usernames said:

    To top it off, the guy at the desk next to me also got a new computer today, same make/model, but his sleeps just fine. So I probably have bad hardware....

    The sleep issue is almost certainly bad hardware. The wifi issue is your IT staff not being able to find their ass with both hands. The rest are issues you're imagining.



  • @too_many_usernames said:

    Still a WTF in my mind... why don't nVidia and AMD have a low power mode so two chips aren't required?
    No idea. On AMD processor based notebooks they have AMD integrated graphics which are more energy efficient... then again considering that AMD processors generate more heat and consume more power you wont notice a difference in battery life really compared to a dedicated GPU setup.



  • @blakeyrat said:

    In what way do they not work?

    Program uses a single USB channel to talk to a USB to CAN converter. There are two client devices on the CAN bus. On WinXP or Win7 32-bit, the software can talk to both clients on the CAN bus. On Win7 64 bit, it can only talk to one client device. I suspect it's not so much a Win64 issue so much as just terrible programming that hit some weird corner case of undocumented API behavior.

    @blakeyrat said:
    Because video cards have sleep mode? Just a wacky wild guess there.

    I guess I'm "living in the old days" where if you remove power from a device, it has no choice but to go to sleep. Stupid "modern" systems that have to be "shut down properly" before they allow being shut off. Also, "back in the day", things would recover to a sensible state upon being powered back up without having all kinds of software state to manage. (To be fair, the computer does have an SSD, so a full shutdown/reboot only takes about 25 seconds. Sleep recovery, assuming it works, would have taken about 5 - most of which is me typing my password.)

    @blakeyrat said:
    The rest are issues you're imagining.

    I'm not imagining the vertical lines; others in my office see them! Of course, they could be patronizing me...and I guess there's no empirical way to know for sure...



  • @too_many_usernames said:

    I'm not imagining the vertical lines; others in my office see them! Of course, they could be patronizing me...and I guess there's no empirical way to know for sure...


    If only there was some way to capture some sort of image of an object so it could be shown to others...

    Or maybe to somehow "zoom" with your eyes and see things as bigger than they really are.



  • @aihtdikh said:

    @too_many_usernames said:

    I'm not imagining the vertical lines; others in my office see them! Of course, they could be patronizing me...and I guess there's no empirical way to know for sure...

    If only there was some way to capture some sort of image of an object so it could be shown to others...
    Or maybe to somehow "zoom" with your eyes and see things as bigger than they really are.
     

    Don't be silly! There is absolutely no such magical device in existence camera.

     



  • @blakeyrat said:

    @too_many_usernames said:
    WTF!?! WHY DOES SLEEP HAVE ANY DEPENDENCE ON VIDEO DRIVERS!?!?!

    Because video cards have sleep mode? Just a wacky wild guess there.

    because the data on the card's ram needs to be written on the disk (and restored) for the programs to be able to wake up without needing to initialise opengl again, a good driver will be able to minimize the data needing to be transfered



  • @too_many_usernames said:

    I guess I'm "living in the old days" where if you remove power from a device, it has no choice but to go to sleep. Stupid "modern" systems that have to be "shut down properly" before they allow being shut off.

    I think the last computer I owned that allowed that, was a ZX Spectrum.



  • @too_many_usernames said:

    Still a WTF in my mind... why don't nVidia and AMD have a low power mode so two chips aren't required?
    It's because Intel now puts a graphic chip directly on their CPUs, so if you want anything more powerful than what Intel offers, you're stuck with 2 cards anyway.
    @blakeyrat said:
    You need one. Any others you should remove. (Pro-tip: keep the one Windows Update installs for you.)
    Be careful with WU drivers (when they're available at all) - I've had it try to install out-of-date drivers, which then caused trouble (sound driver that resulted in IE using 100% CPU, WLAN driver that caused the card to randomly lose connection, etc.).



  • @ender said:

    @blakeyrat said:
    You need one. Any others you should remove. (Pro-tip: keep the one Windows Update installs for you.)
    Be careful with WU drivers (when they're available at all) - I've had it try to install out-of-date drivers, which then caused trouble.
    Windows 7 and 8 have a driver for the USB wireless adapter on my desktop computer that works fine right out of the box. But Windows Update keeps trying to get me install a different driver.  I once made the mistake of using the driver offered by WU and it completely borked my wireless connection.



  • @too_many_usernames said:

    4. Some of my customer-specific apps don't work on Win7 64 bit. Third party, not mine, I know, I know. (Best part? The app was written in .Net...)
    If they're your apps (.EXEs), recompile for "x86" instead of "AnyCPU" and they will magically start working again.



  • @TwelveBaud said:

    @too_many_usernames said:
    4. Some of my customer-specific apps don't work on Win7 64 bit. Third party, not mine, I know, I know. (Best part? The app was written in .Net...)
    If they're your apps (.EXEs), recompile for "x86" instead of "AnyCPU" and they will magically start working again.

    That ticked me off so bad when I first ran into that. AnyCPU doesn't work on 64bit? but x86 does (I know, as does any x86 program for the most part, but still)?

    @too_many_usernames said:

    the computer does have an SSD, so a full shutdown/reboot only takes about 25 seconds

    Yikes - my Asus laptop is 12 seconds running Bodhi Linux (as long as I set BURG to not wait for a choice). It's roughly 6 or 7 seconds from a cold boot. It's not any slower booting to windows 8 either (which i had to install to be able to take a test at school - yech)
    My desktop isn't SSD, and boots to win 7 as fast as your Dell there.



  • @DrakeSmith said:

    That ticked me off so bad when I first ran into that. AnyCPU doesn't work on 64bit? but x86 does (I know, as does any x86 program for the most part, but still)?
    AnyCPU tries to run the program in 64-bit mode, which works fine for managed code - however, if the program relies on some native DLLs, and doesn't ship x64 versions of those DLLs, it will be unable to use them and will fail.


    You don't actually have to recompile the program - it's possible to mark the executable locally to have it run in x86 mode.



  • @DrakeSmith said:

    That ticked me off so bad when I first ran into that. AnyCPU doesn't work on 64bit?

    Dafuq? What are you blubbering about, and how is it so factually wrong?



  • @too_many_usernames said:

    I guess I'm "living in the old days" where if you remove power from a device, it has no choice but to go to sleep.

    If you remove power from a device, it goes to dead and loses all its memory. Unplug your TV instead of turning it off, and next time you turn it back on the volume is set to "deafening" and it's on channel 3. Unless yours is fancier than mine, and has permanent memory.

    @too_many_usernames said:

    Stupid "modern" systems that have to be "shut down properly" before they allow being shut off.

    You'd prefer corrupted files? Actually with journaling it's even better, because instead of corrupted files you get nothing at all, as if your last 5 minutes of computing before cutting the power was transported into the Twilight Zone.

    @too_many_usernames said:

    Also, "back in the day", things would recover to a sensible state upon being powered back up without having all kinds of software state to manage.

    Without sleep mode, it takes a significant amount of time to refill your video card's memory with all the stuff it's supposed to be displaying. If your computer just completely shut off the video card, when it woke from sleep it'd have to spend half a minute with churning HDs getting all that data back where it belongs and in the meantime the user experience is "black screen or perhaps some flickering". Is that really what you want?

    @too_many_usernames said:

    (To be fair, the computer does have an SSD, so a full shutdown/reboot only takes about 25 seconds. Sleep recovery, assuming it works, would have taken about 5 - most of which is me typing my password.)

    25 seconds is actually pretty long for an SSD-installed OS. If your computer also has a spinning drive, make sure they didn't accidentally put the boot sector on it (meaning, BIOS has to spin-up the drive, read like 2k of data, then never use it again.) Or debloat the shitty anti-virus bullshit your crappy IT department put on the thing-- my company uses Kaspersky which adds like 20 seconds to boot time on its own, fucking piece of shit.


  • Discourse touched me in a no-no place

    @Gurth said:

    I think the last computer I owned that allowed that, was a ZX Spectrum.
    We can build systems that are virtually like this now. Just need to make sure that you're using transactional storage mechanisms. (Which most developers don't seem to grok at all, but that's a whole 'nother story.)



  • @blakeyrat said:

    Unplug your TV instead of turning it off, and next time you turn it back on the volume is set to "deafening" and it's on channel 3. Unless yours is fancier than mine, and has permanent memory.
    My cheapass generic TV and our similarly cheap Toshiba at home all remember where they were and the volume level. Our PVR, on the other hand...@blakeyrat said:
    my company uses Kaspersky which adds like 20 seconds to boot time on its own, fucking piece of shit.
    eurgh Kaspersky Endpoint. We've found that we basically have to neuter it to return your computer to something resembling reasonable performance so there's no real point to it. I've been quite happy with Webroot Secureanywhere, teeny-tiny and it works. Also can be had pretty cheap.



  • Kaspersky clings to the boot process like a piece of malware, I still haven't figured out how to make it not autorun at all. But all devs at our company have local admin, so we can just quit it once booted... if you have the patience to wait for the boot.



  • @too_many_usernames said:

    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!
    ...and that's Windows' fault? Get a different monitor...@too_many_usernames said:
    2. How many wireless drivers do I need anyway? Dell DW WLAN with an icon from 1985. Intel PROSet. And Windows native. Seriously folks...
    Windows 7 comes with an extra driver that allows you to set up a hotspot. It shows up as an extra set of drivers, but it uses the same physical network interface.@too_many_usernames said:
    4. Some of my customer-specific apps don't work on Win7 64 bit. Third party, not mine, I know, I know. (Best part? The app was written in .Net...)
    Well... that's pretty much par for the course.@too_many_usernames said:
    To top it off, the guy at the desk next to me also got a new computer today, same make/model, but his sleeps just fine. So I probably have bad hardware....
    Is his display fucked too?


  • ♿ (Parody)

    @blakeyrat said:

    Is that really what you want?

    Look, he didn't say it would be easy. Hardware makers should just figure out how to make good fucking hardware already.



  • @blakeyrat said:

    Kaspersky clings to the boot process like a piece of malware, I still haven't figured out how to make it not autorun at all. But all devs at our company have local admin, so we can just quit it once booted... if you have the patience to wait for the boot.
    I don't understand how they managed to screw it up so badly, even Small Office Security (which is basically rebadged Pure) is a pain, to the point where we've given up on it. The consumer versions are not nearly as leech like on startup resources, so there's clearly something amiss in the Kaspersky offices.



  • @boomzilla said:

    Hardware makers should just figure out how to make good fucking hardware already.
     

    That's a different sort of machinery, not for computers.



  • @blakeyrat said:

    Unless yours is fancier than mine, and has permanent memory.
    I've never seen a TV that wouldn't remember what channel and volume it was set to before it's power was removed. And none of TVs I've had were fancy.



  • @anotherusername said:

    @too_many_usernames said:
    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!
    ...and that's Windows' fault? Get a different monitor...
     

    Wasn't this a laptop?



  • @ender said:

    I've never seen a TV that wouldn't remember what channel and volume it was set to before it's power was removed. And none of TVs I've had were fancy.

    Congratulations?

    Why the fuck do people post this shit.


  • ♿ (Parody)

    @blakeyrat said:

    @ender said:
    I've never seen a TV that wouldn't remember what channel and volume it was set to before it's power was removed. And none of TVs I've had were fancy.

    Congratulations?

    Why the fuck do people post this shit.

    @blakeyrat said:

    Unplug your TV instead of turning it off, and next time you turn it back on the volume is set to "deafening" and it's on channel 3. Unless yours is fancier than mine, and has permanent memory.



  • @ender said:

    @DrakeSmith said:
    That ticked me off so bad when I first ran into that. AnyCPU doesn't work on 64bit? but x86 does (I know, as does any x86 program for the most part, but still)?
    AnyCPU tries to run the program in 64-bit mode, which works fine for managed code - however, if the program relies on some native DLLs, and doesn't ship x64 versions of those DLLs, it will be unable to use them and will fail.


    You don't actually have to recompile the program - it's possible to mark the executable locally to have it run in x86 mode.

    Corflags /32BIT- can usually do it.

    Not that this is a common thing, when folks incorrectly compile for AnyCPU because they only have 32 bit, and likely thought that they needed AnyCPU to work on 64 bit.

    Not that I'm saying you personally didn't know, but some folks in the thread seemed like they could use it.

    Not that I feel like I have enough disclaimers in this post.



  • @Mason Wheeler said:

    @anotherusername said:

    @too_many_usernames said:
    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!
    ...and that's Windows' fault? Get a different monitor...
     

    Wasn't this a laptop?

    Guess it was. My bad. The original post didn't say "laptop" anywhere (and I'm not familiar with Dell's product line enough to know that an E6540 is a laptop). In reading it again, though, it appears that the rant was less focused on Windows than I originally thought. Most of it was actually about the quality of the hardware and the drivers.

    Side note, a 15.6" 1080p HD screen is fairly high-density (141 pixels per inch) and someone whose "eyesight isn't even that good" would have to get really close to even see the individual pixels. I'm a little bit curious what he's talking about. Perhaps he could have someone fax him a screenshot so he can put it on a wooden table and get a good photo for us.



  • @anotherusername said:

    The original post didn't say "laptop" anywhere
     

    Status: FALSE

    Myth: BUSTED

     

    But for some reason, probably the tiny screen diagonal, I magically knew that it was a laptop even though it was not confirmed until later, so yeah. If you glance over numbers you could think it was a desktop case.



  • @dhromed said:

    @anotherusername said:

    The original post didn't say "laptop" anywhere
     

    Status: FALSE

    Myth: BUSTED

     

    But for some reason, probably the tiny screen diagonal, I magically knew that it was a laptop even though it was not confirmed until later.

    For some reason, "Find" is currently broken on my FF 28 so I'll take your word for it. I need more coffee. And the weekend to hurry up and get here.



  • @dhromed said:

    @anotherusername said:
    The original post didn't say "laptop" anywhere
    Status: FALSE

    Myth: BUSTED

     

    But for some reason, probably the tiny screen diagonal, I magically knew that it was a laptop even though it was not confirmed until later, so yeah. If you glance over numbers you could think it was a desktop case.

    The OP (aka the very first post in this thread):

    @too_many_usernames said:

    3. The computer won't wake from sleep. Initially it woke up and then every app threw an error dialog about inability to write to some memory address due to an I/O error. Forms suggest updating the video drivers. Oh, did I mention? <font color="#FF0000" size="5">This laptop has two video chips</font> - Intel HD and AMD Radeon (WTF?). So I updated the drivers.

     

     



  • @El_Heffe said:

    The OP (aka the very first post in this thread):

    @too_many_usernames said:

    3. The computer won't wake from sleep.
    Initially it woke up and then every app threw an error dialog about
    inability to write to some memory address due to an I/O error. Forms
    suggest updating the video drivers. Oh, did I mention? <font color="#FF0000" size="5">This laptop has
    two video chips
    </font> - Intel HD and AMD Radeon (WTF?). So I updated the
    drivers.

     

     

    Gosh, it's so easy to find now. The first time I read it, by the time I got to that part I'd already forgotten that I was trying to figure out whether it was a desktop or a laptop and I completely glossed over the word "laptop".



  • @too_many_usernames said:

    So finally, with much kicking and screaming, I'm being forced to migrate from XP to Win7 (64 bit) by IT because of XP EOL.

    This isn't a bad thing. Win7 is pretty good, once you do some tinkering: ClassicShell, 7+ Taskbar Tweaker for immediate relief.

    MS seems to have slipped into Star Trek Mode (tm): every other Windows release is decent. Win7 is okay, caveat the above. I made the same switch about 3 years ago.

    @too_many_usernames said:

    I get my shiny new computer (Dell E6540 for anyone that cares) and it's replete with WTFs, some of which I blame on Windows, some I blame on people cutting costs, and others I don't even...

    I had an E6xxx at a previous job, and it was okay over all.

    @too_many_usernames said:

    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!

    Your IT is probably paying for Dell's "No questions" support. Hand that laptop in and get a brand new replacement, asap.

    Protip: get the backlit keyboard.

    @too_many_usernames said:

    2. How many wireless drivers do I need anyway? Dell DW WLAN with an icon from 1985. Intel PROSet. And Windows native. Seriously folks...

    The Dell one is worthless. I'm not even sure why they bother to include it... which reminds me. The biggest improvement in the machine I had was when I needed a device wipe and reinstall from my IT's standard windows image. Ask your IT for that too. Gets rid of all the Dell preinstalled crap in one shot. Then hit up Dell's driver site and cherry pick just the ones you need. I recommend video, audio, network(s), and the trackpad driver. I usually install the chipset drivers too.

    @too_many_usernames said:

    4. Some of my customer-specific apps don't work on Win7 64 bit. Third party, not mine, I know, I know. (Best part? The app was written in .Net...)

    MS makes available for free "XP Mode" which is a sort of VM + XP image bundled together. If that doesn't work they also make available for free an XP image for booting under a real VM. One of those will work for your apps, although it's a bit of a PITA. Complain loudly and often.

    @too_many_usernames said:

    This laptop has two video chips - Intel HD and AMD Radeon (WTF?).


    Are you docked? Your dock has a separate video chip in it, which is why the dock requires a slightly different power supply than the laptop itself.

    @too_many_usernames said:

    Program uses a single USB channel to talk to a USB to CAN converter. There are two client devices on the CAN bus. On WinXP or Win7 32-bit, the software can talk to both clients on the CAN bus. On Win7 64 bit, it can only talk to one client device. I suspect it's not so much a Win64 issue so much as just terrible programming that hit some weird corner case of undocumented API behavior.

    If it's not the managed code thing mentioned elsethread, the driver/apps's probably using pointer arimethic to get to the second device, and it's not coded properly for 32- vs 64-bit pointers. This is entirely common.



  • @El_Heffe said:

    The OP (aka the very first post in this thread):
     

    I meant "until somewhat later in the OP"



  • @cdosrun said:

    @ender said:
    @DrakeSmith said:
    That ticked me off so bad when I first ran into that. AnyCPU doesn't work on 64bit? but x86 does (I know, as does any x86 program for the most part, but still)?
    AnyCPU tries to run the program in 64-bit mode, which works fine for managed code - however, if the program relies on some native DLLs, and doesn't ship x64 versions of those DLLs, it will be unable to use them and will fail.


    You don't actually have to recompile the program - it's possible to mark the executable locally to have it run in x86 mode.

    Corflags /32BIT- can usually do it.

    Not that this is a common thing, when folks incorrectly compile for AnyCPU because they only have 32 bit, and likely thought that they needed AnyCPU to work on 64 bit.

    Not that I'm saying you personally didn't know, but some folks in the thread seemed like they could use it.

    Not that I feel like I have enough disclaimers in this post.

    When I first ran into it, it was on a legacy program that was several years old that needed a bug fixed, and I had only been programming for 3 months. I didn't know any of this then. Once I 'fixed' the compilation by changing the target, I researched as to why that did. Now I know.



  • @too_many_usernames said:

    I guess I'm "living in the old days" where if you remove power from a device, it has no choice but to go to sleep.

    What you're describing is not "sleep", it's called "power off".

    Or are you dying every night and magically ressurrected from the dead every morning?



  • @too_many_usernames said:

    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!

    ...

    So I probably have bad hardware....

    I've seen that visible pixel mask effect on a customer's new Dell Vostro laptop. It's not actually the pixel mask - you really can't see that on a 15" 1920px screen; hell, I'm flat out seeing it with a magnifier on this 24" 1920px screen in front of me - it's every second column of pixels staying dark. Get Dell to replace your mobo under warranty.


  • @flabdablet said:

    I've seen that visible pixel mask effect on a customer's new Dell Vostro laptop. It's not actually the pixel mask - you really can't see that on a 15" 1920px screen; hell, I'm flat out seeing it with a magnifier on this 24" 1920px screen in front of me - it's every second column of pixels staying dark. Get Dell to replace your mobo under warranty.

    Ah. Yes.

    Seeing pixel mask at 1080p resolution on a 15" screen, I do not believe.

    Every second column of pixels failing to render, that I believe.



  • And just to prove that you are in fact missing every second column of pixels: use Paint under high zoom to draw a rectangle with 1px line thickness and an even width. If I'm right about what's going on with your machine, you will then only ever be able to see one of the vertical lines when viewing that at 1:1 scale.



  • @flabdablet said:

    @too_many_usernames said:

    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!

    ...

    So I probably have bad hardware....

    I've seen that visible pixel mask effect on a customer's new Dell Vostro laptop. It's not actually the pixel mask - you really can't see that on a 15" 1920px screen; hell, I'm flat out seeing it with a magnifier on this 24" 1920px screen in front of me - it's every second column of pixels staying dark. Get Dell to replace your mobo under warranty.
    Google isn't being helpful. What is it?


  • @flabdablet said:

    And just to prove that you are in fact missing every second column of pixels: use Paint under high zoom to draw a rectangle with 1px line thickness and an even width. If I'm right about what's going on with your machine, you will then only ever be able to see one of the vertical lines when viewing that at 1:1 scale.
    I've actually seen a similar problem on a standalone LCD monitor. Every 2nd row was still lit, but dimly. The way I found best illustrated it was to create a pattern of alternating black and white stripes 1px wide. Moving it horizontally will cause it to pulsate as the white pixels move across the working/not working columns of the display.



  • @flabdablet said:

    It's not actually the pixel mask - you really can't see that on a 15" 1920px screen; hell, I'm flat out seeing it with a magnifier on this 24" 1920px screen in front of me - it's every second column of pixels staying dark.

    That's pretty bad if that's the case - especially since it's on two brand new pieces of hardware. That's either a hideous manufacturing plant, a design defect, or some stupid attempt at power saving by dimming every other line (specific type of design defect).

    I'll have to take a magnifier to my screen on Monday, but I still suspect what I'm seeing is that the new screen doesn't have any black horizontal space between any pixels - at least, that's what it appears to me as I put my face close to that screen as compared to the other screens I had in the office. It was that unbroken vertical coverage that disturbs me - notably, the effect is most pronounced with large swaths of solid colors.



  • @Ben L. said:

    @flabdablet said:
    @too_many_usernames said:

    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!

    ...

    So I probably have bad hardware....

    I've seen that visible pixel mask effect on a customer's new Dell Vostro laptop. It's not actually the pixel mask - you really can't see that on a 15" 1920px screen; hell, I'm flat out seeing it with a magnifier on this 24" 1920px screen in front of me - it's every second column of pixels staying dark. Get Dell to replace your mobo under warranty.
    Google isn't being helpful. What is it?
    Try this one.


  • @too_many_usernames said:

    That's either a hideous manufacturing plant, a design defect, or some stupid attempt at power saving by dimming every other line (specific type of design defect).
    Probably a hairline soldering fault underneath the GPU that didn't manifest until the machine had been bumped around a bit during shipping.



  • @flabdablet said:

    @Ben L. said:
    @flabdablet said:
    @too_many_usernames said:

    1. It's a 15.6" screen, 1920x1080. But someone screwed up badly, because I can see the damn pixel masks on the screen! The screen manufacturers didn't put any horizontal breaks in the pixel mask, so I see Every Single Vertical Line. This isn't an aliasing thing - it's the pixel coloration itself. And my eyesight isn't even that good!

    ...

    So I probably have bad hardware....

    I've seen that visible pixel mask effect on a customer's new Dell Vostro laptop. It's not actually the pixel mask - you really can't see that on a 15" 1920px screen; hell, I'm flat out seeing it with a magnifier on this 24" 1920px screen in front of me - it's every second column of pixels staying dark. Get Dell to replace your mobo under warranty.
    Google isn't being helpful. What is it?
    Try this one.

    SCREEN TEST:

    What color is "The Daily WTF" in this image?



  • @Ben L. said:

    SCREEN TEST
    This reminds me of what I used to have to do at an old workplace where we moved desks frequently, and almost every screen was a cheapo EIZO LCD unit with about 1280 pixels across and a diagonal of around 17". They were pretty much always connected via VGA and not calibrated properly, so I kept in my documents folder a 1px checkerboard repeating pattern to make sure it was aligned properly.

    Another WTF about these monitors is that they beeped when you used the on-screen menu and it sounded exactly like a cheap digital watch. Not wtfy enough for you? It also beeped 3 times when it went into standby. Now consider screens going into standby after 10 minutes of idle. Okay, multiply by 2 because there are 2 per workstation. Now multiply by 100 (large open-plan office). Yeah... I turned off as many of those buggers as I could. Why would this be turned on by default?!

    Extra wtf. Sometimes when they lost signal, they would beep three times, every 30 seconds. Arrgghh!

     



  • @LoremIpsumDolorSitAmet said:

    They were pretty much always connected via VGA and not calibrated properly, so I kept in my documents folder a 1px checkerboard repeating pattern to make sure it was aligned properly.
    When I bought my first LCD, it only had VGA input (not that it'd matter - it was years before DVI outputs became common on graphic cards), so I made this to help me with calibration.



  • @ender said:

    @LoremIpsumDolorSitAmet said:
    They were pretty much always connected via VGA and not calibrated properly, so I kept in my documents folder a 1px checkerboard repeating pattern to make sure it was aligned properly.
    When I bought my first LCD, it only had VGA input (not that it'd matter - it was years before DVI outputs became common on graphic cards), so I made this to help me with calibration.
    Yep! That's exactly what mine was like, only it was just a bitmap approx 1000x1000px. I used to fix up other people's screens too just by hitting 'auto' when I was watching over their shoulder for whatever reason. I feel really sorry for the guys who couldn't tell the difference between a blurry and a properly calibrated screen... do their glasses not work? Or is this something I won't be able to distinguish either when I reach 50? Thankfully that's still some time away.

    Hmm... I was just thinking, health and safety requirements mandate that we fill in a workstation assessment form when we start and whenever we move desks, and it covers all the things you'd expect, like chair height and tilt, screen height and angle, brightness, ambient light, and general 'can you see the screen' types of questions. They could, and should, actually add a basic monitor test screen to the process with something like that TDWTF sample image above.


Log in to reply