wacom's newest tabletPC driver issues!

Discussion in 'Fujitsu' started by stoneseeker, Oct 12, 2012.

Thread Status:
Not open for further replies.
  1. Plognark

    Plognark Scribbler - Standard Member

    Messages:
    176
    Likes Received:
    25
    Trophy Points:
    41
    In the T901 I wound up mitigating the problems with pretty much the same approach we're taking with the T902. I had other issues with the T901 that made me finally give up on it and upgrade to the T902; mainly that goofy scroll wheel circle on the side of the touchpad. Sharing one PS/2 port's bandwidth between three devices was a terrible idea.

    For the brush lag, my suspicion is that it's driver based, but I won't really be able to tell unless I can dig into the source. I've run some proc-mon logs over the last few months, and I can see whenever there's even a tiny bump in CPU activity something goes weird.

    For a while I thought it might have to do with the scratch disk, but with an SSD that seems unlikely.

    It seems like other processes take priority over photoshop, and it gets backed up a bit while they finish up. I've found the same problems in both windows 7 and 8 (64 bit). I'm back on 7 pro now, although eventually I'll get around to doing a dual boot to try and compare the two.

    The most aggravating part is that I don't get these same issues in other art programs. I'm ready to switch to painter as my go-to program after all this aggravation.
     
  2. Tab Rocker

    Tab Rocker Pen Pal - Newbie

    Messages:
    52
    Likes Received:
    6
    Trophy Points:
    15
    I would donate as well. I had a deadline last night and the pen bug was driving me nuts. I was getting it every 10-15 minutes randomly while sketching. Probably should have restarted the computer but didn't want to close everything out.
     
  3. stoneseeker

    stoneseeker Animator and Art Director Senior Member

    Messages:
    3,601
    Likes Received:
    1,072
    Trophy Points:
    231
    I find it happens once every 10 minutes or so average in large Flash files, once every 20 minutes maybe in photoshop and once every 15 or so in Manga Studio 5. So I don't think it's merely photoshop, perhaps larger work just aggravates the problem as more can take priority form the wacom driver. I've even experienced it just surfing the web or opening files and folders in windows, although it's rare .
     
  4. Plognark

    Plognark Scribbler - Standard Member

    Messages:
    176
    Likes Received:
    25
    Trophy Points:
    41
    I ran some more proc-mon tests over lunch, and there is something interesting with the touchscreen. Even when disabled, touching the screen will fire off a barrage of system interrupts. Windows isn't doing anything with them, but the hardware is still sending them. I thought to check this after I debugged the hell out of the T901 scroll wheel and found similar behavior. That was a lost cause, since those settings were stored in bios, well and truly out of my reach.

    This might be different though. If the drivers just aren't handling touch input elegantly, then even when disabled it's going to be adding some serious overhead. I didn't test out pen input at the same time to see what came from that; I'll try that after I get home from work.

    After that I may try some tests wearing a full glove so that there's no skin contact on the screen and see what the results are. I see jan_c has a drafting glove, but the fingers are exposed so it's still going to register events on the capacitive screen.

    I'll let you guys know what I find.
     
  5. SpiralMagus

    SpiralMagus Pen Pal - Newbie

    Messages:
    61
    Likes Received:
    4
    Trophy Points:
    16
    I would happily give cash to anyone who can hack a fix for the glitch too. It's the only complaint I have for the T902.
    Painter is my main art program and the glitch happens with that too.
     
  6. stoneseeker

    stoneseeker Animator and Art Director Senior Member

    Messages:
    3,601
    Likes Received:
    1,072
    Trophy Points:
    231
    I noticed the system interrupts before, but had no clue it was caused by the disabled touchscreen! Is there any "deeper" way of turning off touch layer?

    When you try with a glove, try using a VERY thick glove for the sake of testing, or a rubber glove. I can still register touch through a cotton glove, because I've tried using it with touch enabled, and even with my pinky fully covered, could set off the touchscreen. However, since using a glove to draw (helps my hand glide a bit better while drawing) I have noticed improvements with the glitch, and it happens less often. I had my suspicions that the system interrupts might be associated with the issue...

    I feel like were getting somewhere, this is exciting! If we fix this, it will be a near perfect device.
     
  7. Plognark

    Plognark Scribbler - Standard Member

    Messages:
    176
    Likes Received:
    25
    Trophy Points:
    41
    I think the system interrupts might be remnants from the windows tabletpc input service that we've been disabling. I was messing around with it and had it enabled after a reboot, and I wasn't getting that same huge barrage of them on the process monitor. That being said, the brush runs better with the tablet pc service disabled.

    I will say that I didn't have these issues with my old T5010 running CS4. That was just a core 2 duo, so it had different performance issues, but it was a pen only digitizer without touch input.

    I think that part of it may be that these dual-digitizer drivers aren't designed to work with touch disabled. I mean, I get that that defeats the purpose of having a dual digitizer, but I figure Wacom is probably putting 99% of their coding efforts into making the touch settings work, rather than having them work better when disabled.

    That fact that SpiralMagus gets the same brush lag in painter is depressing, but it's good in a way, because it means it's not something unique to photoshop. I've only messed around with artrage, painter, and sketchbook for short test runs, not enough to tax the system and build up to this error. I'll try to repeat with sketchbook today, since I don't have painter or artrage on the latest OS install yet.
     
  8. stoneseeker

    stoneseeker Animator and Art Director Senior Member

    Messages:
    3,601
    Likes Received:
    1,072
    Trophy Points:
    231
    Well I get the issue when touch is disabled, but not the tabletPC service. I leave that enabled now. I simply go in and disable the wacom touch service since it was eating up memory like crazy. Everything else I leave enabled, but I go and turn off touch in the control panel as well since I don't really use touch. I'm pretty sure with everything enabled (tabletPC pen and touch, Wacom pen and touch) the problem would still present itself, but your saying the system interrupts wouldn't happen at least? So the system interrupts are not the cause of the issue?
     
  9. jan_c

    jan_c Pen Pal - Newbie

    Messages:
    48
    Likes Received:
    4
    Trophy Points:
    16
    Sorry I've been away for a bit; just dropping in to let you know I'm talking to Fujitsu and they're aware of the problem. (Wacom simply tossed out my support request.) We are currently trying firmware and BIOS updates, but no success yet.

    @Plognark Interesting findings! I've also been suspecting that it has something to do with the way interrupts are serviced in Win 7 (again, XP and Ubuntu are just fine). Perhaps of interest, I've been running LatencyMon and it reports that there's quite a bit of latency associated with the USB driver. And that ISD_Tablet.exe gets a hard pagefault from time to time.
    [​IMG]
    [​IMG]
    [​IMG]

    @Stoneseeker Turning off Bluetooth might have helped you because the Bluetooth module shares the same internal USB hub as the Wacom tablet. Again, a hint that something's not quite right with the USB driver.
    [​IMG]

    Also, if anyone needs proof that this happens OUTSIDE of Photoshop/Painter/tablet-aware programs, here it is in MS Paint!
    [​IMG]
     
  10. Plognark

    Plognark Scribbler - Standard Member

    Messages:
    176
    Likes Received:
    25
    Trophy Points:
    41
    Thanks jan_c, that's extremely helpful!

    I wound up disabling my Bluetooth yesterday as well, but I was just trying to shut off as much software as possible to isolate the problem. I may have accidentally improved things more than expected.

    In the proc mon, I noticed that the Bluetooth was firing off events pretty regularly, probably just scanning for active devices in the vicinity. I also disabled the camera service, which was another one firing off events at regular intervals. Do you know if the camera is on the same hub as everything else?

    I'm going to run latency mon tonight and see if I get similar results.

    I think I'll dig up some USB performance monitoring software and see if that's where our bottleneck is. It would be crazy if this wound up being nothing more than a USB issue!
     
Thread Status:
Not open for further replies.

Share This Page