Mythruna

Hey. I've been reading the forums for a short while and I've got to start by saying your game is moving well. I noticed a bug when I installed the new 301.24 nvidia driver. The fire now loses its glow and becomes separate particles depending on distance. The closer, the more particles. The farther, the more normal. The picture might help.

(http://i.imgur.com/o7Esf.jpg)

I know that you have worked on this fire for a long time, and you probably don't have the time now. Just wanted to make you aware of the bug (maybe it is due to a new rendering of the fire because it is the new gen nvidia drivers.)

And they looked ok before?

There is one known issue where certain cards+drivers only render point sprites a max of a certain size... and I use a particle-based point-sprite effect for my fire.  It's possible that something with the driver changed this to decrease the max... which would really suck.

They also look like they are not clipped properly... which also really sucks.

Hopefully this is a bug and hopefully they fix it.  I need to come up with an alternative to using point sprites anyway because some cards will always have this size limitation but it would suck to have to do the alternate solution on all cards because it uses more memory + processing to do it a different way.

i updated my video card drivers to 301.24 to and i cant seem to replicate your bug i tryed with many difernt flames and took many photos, the only one which was sort of similar is the group of flame where its positioned right in the corner but i think its difernt.
so here are the pics :P

(http://img214.imageshack.us/img214/5309/mythruna1339918197739.png)
(http://img21.imageshack.us/img21/4916/mythruna1339918197738.png)
(http://img715.imageshack.us/img715/2851/mythruna1339918197737.png)
(http://img833.imageshack.us/img833/5747/mythruna1339918197736.png)
(http://img42.imageshack.us/img42/7135/mythruna1339918197735.png)

FB, you have a good card with decent sprite support it seems.  Maybe the issue is with his specific card + latest driver... if it worked before, I mean.

Oh yea :P it just seemed like you where blaming the driver update hehehehe

Seeing your pictures with the new driver suggests that pspeed's explanation is correct. I know it is the driver because this happened after updating the driver with no other changed (from 294.something) What is interesting is the bug occurs dependent on distance. With my initial picture, you can see the flame perfect, about 6-8 blocks away.

Anyway, I will experiment on this on my own with different drivers and some of the beta drivers to see where exactly this problem occurs. If it helps I have an nvidia geforce 555m with 96 shaders, ok for laptop but not in the league of desktop gpus. Thanks for narrowing it down a bit.

Yeah, this is because the "maximum point-sprite size" is in pixels on screen.  So if the max-size for the card is 64x64 then it will never be bigger than 64 pixels wide on screen no matter how close you get.

I thought the problem where you can see seperate points of fire was normal. I never knew that! I also have an Nvidia driver :p. I can't update my drivers since they stopped making updates for this video card since 2009. (And my laptop was made in 2008... Darn you driver updaters)

Ok, sorry for being pretty slow with this. I did some experimenting and noticed that when I switch to the intel HD3000, integrated, and leave nvidia unpowered, the problem persists. However, it looks better as they don't completely move toward the separated state as before but one could still make out the separation. So the problem does not seem hardware specific (in my case).

pspeed, your explanation sounds very accurate, also taking account the distance difference. Maybe a setting was changed with installation of new driver in my case? I'm working at reinstalling clean versions of java and ms visual c++ since that has become unstable.

It's narrowing but I don't know where the problem is. It is not major so for now we can leave it and hopefully we will stumble upon the solution soon without much working.

I'm curious though, how many of you have this problem, as Moonkey does and I have started recently.

I know you are talking about the card but is there maybe a setting somewhere for this which I can read a value. The card itself is fine unless the driver changed some card limits (why they would decrease I don't know).

Uploaded client log. I noticed a big difference from before when fire was good. Seems like something doubled or tripled on itself causing lots of warning errors.
http://freetexthost.com/stbnngm5th (http://freetexthost.com/stbnngm5th)

Those warnings are all caused by the menus... not the game itself.  No worries.

It's weird that it would run in your intel card at all.

It's not _just_ the driver as some nVidia cards have had this limitation for a long time... and nearly all ATI cards seem to.  It's just that the GTX cards seem to let you use any point sprite size you want.

It makes me sad that people don't all  have my graphics card or better. :)

yep, after searching for nvidia past drivers (which happened to be on a separate website) the problem was fixed. That means it was the driver/card combo, although I wonder how a relatively new 555m with a new driver seems to have such a bug. Reverted for now. Moonkey had the same problem so I guess it is a random nvidia problem?

My intel card, the integrated hd3000, runs about 30-40 fps, so it is usable but a little annoying to run. Not many graphics options for laptops unless you want to shell out money, I'm lucky I got an nvidia 555m at about 800 dollars.

I first saw the issue with people using ATI cards and trying to run my waterfall prototype.  The point sprite issue was much bigger in that test and I will definitely have to do water falls a different way because of it.

I don't know what factors go into capping the size of point sprites... but it's annoying. :)

*cough* I have a GTX, but it's a notebook video card.