New Rendering not looking good
by Mark Dynna · in Torque Game Engine Advanced · 07/06/2006 (7:16 pm) · 28 replies
Ok, I got and built the new v3.5 code of TSE. Now, maybe it's geared for the new Atlas2 terrain (which I haven't been able to get working) but using my old CHU and TQT files, it doesn't look as nice now.
Before:

After:

Maybe there are some new settings that I'm missing?
Before:

After:

Maybe there are some new settings that I'm missing?
#22
07/11/2006 (4:01 pm)
Ok when did that page get update? I could swear that other function wasn't there like 10 minutes ago! I'm leaving from work right now and I'll give this stuff a shot when I get home Ben.
#23

I tried the importOldAtlasTQT (named incorrectly on the wiki) but apparently my image doesn't have a mip map ("Ben is lazy"?).
07/11/2006 (8:15 pm)
Ok, I don't know what's going on, but I don't get the memory error anymore. What I do get are a bunch of "strings" going off into infinity!
I tried the importOldAtlasTQT (named incorrectly on the wiki) but apparently my image doesn't have a mip map ("Ben is lazy"?).
#24
That looks like it might be a scaling issue. Hard to say. Might be a missing shader if they're all mostly in the same place.
07/12/2006 (4:06 pm)
Can you please give me the exact error message?That looks like it might be a scaling issue. Hard to say. Might be a missing shader if they're all mostly in the same place.
#25
The image above is using a CHU and TQT combination that worked perfectly fine under the MS3.0 release. That is the result when using the MS3.5 code. I tried re-generating them and got the same result.
07/12/2006 (8:56 pm)
Apologies. The exact error message is: "Ben is lazy. No mip generation, and your image doesn't have 'em." (I wasn't trying to insult you above, I was quoting the error message.)The image above is using a CHU and TQT combination that worked perfectly fine under the MS3.0 release. That is the result when using the MS3.5 code. I tried re-generating them and got the same result.
#26
07/13/2006 (2:40 am)
@Mark : Your last screen is nice. Are you trying to make trees with Atlas? You know... I think Atlas isn't the best solution to make trees... LOL... Sory for the unconstructive purpose.:)
#27
That assert isn't from Atlas - it's from the texture manager.
(BTW - if anyone has some code to generate mips of DXT compressed images, I'd love to integrate it and get rid of that assert. :)
07/13/2006 (3:33 am)
Hmm- that's a bit odd, as it implies the DDS files you have in your TQT don't have mips! But the TQT file always has mips stored out. You've checked the TQT loads properly under a normal AtlasInstance? If it's the same DDS file it'll cause the same crash under Atlas1. Otherwise, corruption of some sort is happening at load time.That assert isn't from Atlas - it's from the texture manager.
(BTW - if anyone has some code to generate mips of DXT compressed images, I'd love to integrate it and get rid of that assert. :)
#28
Trying to convert the TQT to a .atlas file gives me the Assert I mentioned above. If I go back to the original JPG to create the .atlas file I get the blue texture result I show in the other thread.
07/13/2006 (7:50 am)
If I use the TQT file with an AtlasInstance (1) with the MS3.5 code, what I get is the result above. If I switch back to MS3.0 code everything works fine and I can re-create the TQT fine as well.Trying to convert the TQT to a .atlas file gives me the Assert I mentioned above. If I go back to the original JPG to create the .atlas file I get the blue texture result I show in the other thread.
Associate Kyle Carter
You only need to use atlasConvertTextureTOC if you want to change the compression format used by your .atlas file.