Runtime texture compression using compute shader

[pedropla]

@joshuacwilde

@TOES
I ended up making a plugin out of the cpp code in QuickETC (not ETC2 branch). It works just barely fast enough for my needs. However, a few days ago I stumbled upon this writeup here : https://zhuanlan.zhihu.com/p/327045410 , which was quite the revelation to find.

I will probably look into this soon and switch over to using some solution similar to theirs, as it was my initial goal to have it done on the GPU rather than CPU, for a few reasons.

Very interesting site you mention. Did you manage to get the ETCCompress.hlsl mentioned on that site? I was unable to download the zip file mentioned. I’d be very interested in testing this out.

[c0d3_m0nk3y]

@pedropla

Very interesting site you mention. Did you manage to get the ETCCompress.hlsl mentioned on that site? I was unable to download the zip file mentioned. I’d be very interested in testing this out.

I’m aware that this is an old thread but here are the files if anybody is interested.

9018082–1243588–cs.rar (7.26 KB)

[unity_CDVKxN7snkcptA]

@Aras

I’m playing with compute shader compression myself right now (though with BC7 format, not ETC2), and so far it’s in the “not really that much faster than doing it on the CPU, but a world of pain in shader compilers, precision mismatches, GPU driver issues” etc.

Runtime compression to ETC/ETC2 is actually there in latest Unity versions (i.e. via Texture2D.Compress), but that’s only starting with 2021.1. Underneath, that runtime compression is the same compressor that bgolus linked to above (“etcpak”, just not the QuickETC2 branch yet).

So we come to May 2023 today… does the QuickETC2 branch feature is in Unity Editor? (Or I need to do the research about transfer that etcpack thing into C#…)

[Prodigga]

We could definitely use some runtime non blocking texture compression love from Unity

[callen]

I’m aware that this is an old thread but here are the files if anybody is interested.

Just wanted to thank you for this, it’s giving me a 100x speedup versus doing compress on the CPU side! I did see their computer shader is assuming no alpha channel (hard codes the alpha block to all 1’s), but despite me having no shader coding skills I was able to make a duplicate of the ReadBlockRGB function which gets a block of source alpha values, and now its perfect. Also if anyone else uses this, what I ended up doing to completely avoid CPU-side slowdowns is to pre-allocate a bunch of textures at the various sizes I need on the first frame of Start, and Apply(true,true) all of them so they are then basically just pointers to a block of compressed GPU memory. No noticable performance hit, and now I can do my texture-compositing weirdness in sub-millisecond times with ease. Would have been impossible if I needed to Compress+Apply on CPU!