A lot of those look like photo sources textures and everything on 1 smoothing group. But I'm not sure exactly what you're looking for crits on; the entire scene or just the thing you pointed out?
I know its standard to use powers of 2, but what is the reasoning? Besides applications that refuse to work with non powers of 2. Something to do with binary and pixel efficacy?
I know its standard to use powers of 2, but what is the reasoning? Besides applications that refuse to work with non powers of 2. Something to do with binary and pixel efficacy?
Has to do with memory limitations, storage, and usage, engines are optimized to with textures that use powers of 2, I'm sure there's a more techy person out there with a better explanation. I know a 2nd reasoning is mipmapping, which uses smaller versions of the texture for sampling (1x1, 2x2, 4x4, 8x8, up to the texture size). There are fixes and work arounds for stuff like that, but generally everything is based around using power of 2 textures.
compression algorithms like power of 2 a lot! most used format .dds only supports power of 2!
mipmap generation works better with it too!... there are many reasons why u should use it and moreless non why u should not.. this page maybe worth a shot http://tech-artists.org/wiki/DirectDraw_Surface
Memory allocation for the GPU. Use powers of 2 like everyone else
Feel free to have non-square textures though, just ensure that the width/height dimensions are each a power of 2. For example 128x64 is fine.
It's also worth noting that if all these objects are intended to be onscreen at once, you will get better performance if you can put them all into 1 single large texture. This avoids the GPU having to switch textures as it's rendering objects.
That being said, these are so low resolution it really shouldn't matter. Just food for thought
It's also worth noting that if all these objects are intended to be onscreen at once, you will get better performance if you can put them all into 1 single large texture. This avoids the GPU having to switch textures as it's rendering objects.)
A lot of engines create texture atlases or have ways of doing that with minimal extra work. As an artist, for personal work, I wouldn't worry too much about having multiple textures on one sheet. Save that for the production environment.
Unless you have 2 assets that go together that uv well together as well.
the one who asked me to do the models told me to use 100*100 textures, I didn't argue with him and thought he knew what he want, he is doing a 3D flash game with away3d.
the number are polygons numbers.
Looks good... guitar should be bigger maybe, looks smaller than the violin.
watch your scaling. the skateboard wheels are too small, and no way a skateboard would ever pass inside the rim of a basketball bucket. the guitar should be bbigger than the violin, or the violin smaler than the guitar, either way watch their proportions.
and, as everyone pointed out, try to work with power of 2 texs. they may even not be equal (you can work with 64x128, 512x2048, doesnt matter). BUT size and height must be power of 2.
Some of these objects are working ok with the photo referenced textures however some things (like the instruments) would look FAR better hand painted and provide alot more accurate details making them readable as what they are meant to be. At the moment some of these things are just a pixelly mess.
Replies
I always thought it was powers of 4. ex] 128 x 128 or 256x256 and up...?
I know its standard to use powers of 2, but what is the reasoning? Besides applications that refuse to work with non powers of 2. Something to do with binary and pixel efficacy?
Has to do with memory limitations, storage, and usage, engines are optimized to with textures that use powers of 2, I'm sure there's a more techy person out there with a better explanation. I know a 2nd reasoning is mipmapping, which uses smaller versions of the texture for sampling (1x1, 2x2, 4x4, 8x8, up to the texture size). There are fixes and work arounds for stuff like that, but generally everything is based around using power of 2 textures.
mipmap generation works better with it too!... there are many reasons why u should use it and moreless non why u should not.. this page maybe worth a shot http://tech-artists.org/wiki/DirectDraw_Surface
cheers next
Feel free to have non-square textures though, just ensure that the width/height dimensions are each a power of 2. For example 128x64 is fine.
It's also worth noting that if all these objects are intended to be onscreen at once, you will get better performance if you can put them all into 1 single large texture. This avoids the GPU having to switch textures as it's rendering objects.
That being said, these are so low resolution it really shouldn't matter. Just food for thought
A lot of engines create texture atlases or have ways of doing that with minimal extra work. As an artist, for personal work, I wouldn't worry too much about having multiple textures on one sheet. Save that for the production environment.
Unless you have 2 assets that go together that uv well together as well.
and then you read the OP's post
OP's post?
Original Poster's post. In this case the person who started this thread.
Anyways.
Please continue.
the number are polygons numbers.
thanks.
and, as everyone pointed out, try to work with power of 2 texs. they may even not be equal (you can work with 64x128, 512x2048, doesnt matter). BUT size and height must be power of 2.
Apparently so dude.
Some weird polycounts there for the amount of detail though
thanks for the crits mates.