Jump to content

jtyotJOTJIPAEFVJ

Members
  • Content Count

    39
  • Joined

  • Last visited

About jtyotJOTJIPAEFVJ

  • Rank
    Junior Member
  • Birthday 06/08/1992

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. If you read my post again you'll notice I said frame rates are definitely not unimportant. Good post otherwise tho.
  2. Frame rates are definitely not unimportant. SB is almost unplayable now on some maps, with FPS dropping close to 10 when using thermals in woods.
  3. Most of my friends, me included, have 100/100 fiber connections though, so it'd be a bit under 30 seconds per client. Waiting three minutes at scenario start would be more convenient than keeping track of download URLs for all scenarios in my folder. Edit: not to mention LAN parties, where we mostly play. A gigabit lan connection wouldn't take very much time at all to do it.
  4. I don't see any problem in leaving the choice for host to upload maps to clients on server connection. Or did I misunderstand something? It shouldn't take long to share even a few GBs of data on match start with current connection speeds. And if someone has a slow connection, they can download the map beforehand since they know it will take long. I don't see why you would remove a working feature just because some people won't have a fast enough internet to use it. Of course, the ideal solution would be to have a mod tracker similar to Steam Workshop that can tell you the dependencies of a session and download them from a dedicated fileserver before you join, but that might take some programming hours to do.
  5. I tested it by looking out the hatch as the commander, as well as using the gunner's thermal and day sights. The fps is pretty bad when looking out of the hatch, but even worse in low magnification with the gunner's thermal sight. I guess the further render range for thermals could explain that as well.
  6. Sounds like I should've applied to become a tank driver in the military
  7. Maybe it's something that's changed when moving to 4.0. I didn't play 3.0 that much, but I don't recall having this serious issues there. I tried some testing with different viewpoints and zoom levels in the Meeting engagement (winter) vanilla scenario, and there it seems that FPS decreases noticeably when you're looking at dense woods. I've noticed the same on many other maps as well, but tested it more thoroughly just in there. It does seem like the tree models are all identical, so the whole forest should be just a single draw call, but I can't say much more than that without knowing anything about how the engine works. I also noticed that thermals cause another big performance hit, probably due to the depth of field effect. So to reiterate my findings: Thermals vs day sight: FPS noticeably lower on thermals, gets worse with low magnification and when looking at dense woods Wide vs narrow FOV: Wide FOV far worse, when looking at forest out of turret hatch with binocs or with high zoom on gun sight, FPS is smooth enough (30 or so), with wide angle, eg. view from turret hatch with no binocs, FPS drops down to 15. Looking at dense woods vs open ground: another noticeable FPS hit happens when you move your camera so that it's filled with dense woods. With wide angle day sight, FPS is around 30 when looking at open ground, and drops to the 15 mentioned above when looking at woods. With wide angle thermals and looking at woods, the FPS drops as low as 5 in some places, which makes aiming quite impossible. We had to stop using thermals completely in some scenarios with dense forests due to this. This does sound like the graphics engine could use some work, a 1080 should be able to handle SB graphics fine. Other scenarios that had performance issues were Red Decision Point Attack (Leo2A4) winter and the Kouvostoliitto scenario downloaded from here. All of them have dense forests, and all of them had the same issues.
  8. Looks like my current drivers are dated to July, might just be a driver issue in that case. It's hard to say either way though, since the issue is so rare.
  9. Me and my friend recently started seeing weird corruption artifacts in SB, after updating to the latest patch. The issue is quite rare (two or three times over maybe ten hours played, each time lasting no more than a minute), but still highly annoying when it does happen. The artifacts seem to be stable on your screen, they stay the same over time until you move your camera, at which point they change into new random patterns, or eventually disappear. It's likely related to having a GTX 1080, as we were the only ones experiencing these issues in our group, and we're the only ones with 1080's. Resolution is probably not the cause, as I have a 4k screen, while my friend has a 1080p one and we both witnessed similar artifacts. We're also both running the game on Windows 10. It kind of looks like something is left uninitialized on the screen, at least it looks similar to bugs I've experienced when I've ended up with screen buffers that read uninitialized garbage data. It also looks like it happens before the post-processing pass, as you can see the artifacts causing bloom on the screenshot. Hopefully this helps you track down the issue. Screenshot of the artifacts that I grabbed when playing is attached below.
  10. I was at a LAN party where we played friends, and we noticed that forests give you a massive performance hit across many different hardware setups. Even with a GTX 1080, a wide FOV, eg. looking out of your tank as commander, in a forested area would bring down the FPS to around 10 or so, whereas using a high magnification optic worked fine. This points toward tree count on the screen being the main reason for performance loss, as opposed to overdraw from the high amount of overlapping objects with alpha blending. The draw call counts sound insane, so maybe you could try consolidating different objects into the same draw call, or try using some occlusion culling technology to lower the amount.
  11. The M1 and M1A1 have similar behaviour on their FCS though. Only the M1A2 can engage other targets without dumping lead.
  12. Don't know much about your numbered questions, but the M1A1 has a remote controlled machine gun for the commander that you can use as a worse version of a periscope. You can't control the gun through there though, as far as I know. The M1A2 has a very nice thermal-equipped periscope for the commander, so you might want to try that out.
  13. Some more T-72B1 window shooting adventures. This time I used my joystick, which makes things a lot easier to control, as you can't really tell where the neutral position is with mouse control. Turns out it's still not that great for shooting at distant targets.
  14. GPS gone? no problem, tank has window! Also, looks like the Iraqi mechanised troops have helmets rated at 700mm RHA:
×
×
  • Create New...