RSS SL

1. forum meetups in sl

   Your outfits are so amazing and how you find the time to keep changing your hair, I will never know. Well done as ever.

2. forum meetups in sl

   Zal can be so direct, but makes me laugh so much!

3. isolate and remove only the internal layer ?

   Thanks Aquila for your valuable advice , as clear as always. I’m the one reading too quickly because I’m in a rush.. All right. I thought I had covered everything, but do I need to mark the seams on both sides (since this part is still doubled)? In any case, I completely forgot to add the vertical seam(s).. Yeah, but my goal was to turn this shirt into a vintage dress, so I need to simulate a thicker fabric that doesn’t deform excessively. I’m aiming for a slightly late-60s A-line silhouette. I still have a lot to learn and refine, but here’s the idea: To compensate for the extreme deformation, I’ve been mixing different texture scales in that area to “cheat” the stretching, but it’s messy and not very clean. I’m wondering whether it would make more sense to simply subdivide the quads locally in that zone instead.. Also, GPT suggest scaling up the UVs only in that specific area , don't know if that’s a reasonable approach. I’m not using Blender 4 or newer. I downgraded because the dev kit was made with avastar in Blender 2.7. I update the rig in blender 3.6, and since everything works, I’ve stayed there. I should probably try updating the rig again, because what you’re showing is exactly what I need! And that's really good to know, thank you!

4. What are you listening to?

   🌈

5. Firestorm 7.1.11 worst version yet in busy locations

   I assume the Firestorm crew would have looked at a patch implementing it. But as a pure feature suggestion it is one of many things one could do to speed things up. The suggestion has a few issues: It's not gonna work for old OpenGL < 4.2, SL still supports 3.0 minimum. Apple OpenGL has only 4.1. BPTC Texture Compression - OpenGL Wiki BC7 has a minimum OpenGL feature level of 4.2. So no luck without killing a lot of old boxes for SL. Apple OpenGL stopped at 4.1 for example. Old Intel iGPUs are also pretty bad. So this is technically unfeasable, unless you either move to Vulkan/Metal first or kill off a significant part of the slow machines you intend to improve things for. Linden Labs streams textures via JPEG2000, so highly compressed already. Their bandwidth bill would not change, probably increase as BC7 compression rate is probably worse than JPEG2000. Pushing texture compression to another thread is kind of silly, if the GPU is much much more efficient doing it. But even if you do it, its counterproductive on machines with low CPU counts (e.g. budget/old machines, e.g. i3 or so). It would be fairly trivial to switch the current cache from "handle JPEG2000/bitmap" to BC7, sure. But if you cannot load it (due to no support in OpenGL < 4.2), its not worth the trouble. So, the benefits in short term performance gains look a bit doubtful. A switch to Vulkan and some improvements to cache eviction sound like a better engineering tradeoff to me, honestly.