Replies: 2 comments 1 reply
-
Hm,
That being the case, I didn't really want to upscale an image silently in a way that's going to reduce device support. At the time I added KTX2/Basis support, this library didn't have any other way to resize images... it does now, so maybe this could be addressed with a simple warning about larger images, advising that the user resize them first if needed. I'm assuming
|
Beta Was this translation helpful? Give feedback.
-
Hey Don, really appreciate the explanation! Yep, I was looking for it to be resized to Thank you and have a great day :) |
Beta Was this translation helpful? Give feedback.
-
Description
Hi, I would just like to clarify the
--power-of-two
flag in the texture compression cli commands. It is stated in the docs that it:I was just wondering why is there a 2048x2048px restriction for this closest power-of-two resizing.
To Reproduce
gltf-transform etc1s ./MosquitoInAmber/MosquitoInAmber.gltf ./MosquitoInAmber-resize/MosquitoInAmber.gltf --power-of-two -v
MosquitoInAmber0.jpg
is being resized by--resize 4096x2048
in the debug messages*model taken from: https://sketchfab.com/3d-models/mosquito-in-amber-53a61d58c09b4d2ab30e269aa3e22078
*manually resized
MosquitoInAmber0.jpg
to 4100x4000pxQuestion
In the source code it is written that
Why is the first line needed? Wouldn't it be better to return the closest power-of-two all regardless of input dimension?
Thank you for your time! :)
Beta Was this translation helpful? Give feedback.
All reactions