commit before updating gdx

This commit is contained in:
minjaesong
2023-10-05 14:57:03 +09:00
parent d81adc449b
commit 1088b1d29b
3 changed files with 9 additions and 5 deletions

View File

@@ -35,17 +35,19 @@ public class Float16FrameBuffer extends FrameBuffer {
public Float16FrameBuffer (int width, int height, boolean hasDepth) {
/*if (!App.gl40capable || App.operationSystem.equals("OSX")) { // disable float framebuffer for Apple M chips
FrameBufferBuilder bufferBuilder = new FrameBufferBuilder(width, height);
bufferBuilder.addColorTextureAttachment(GL20.GL_RGBA, GL20.GL_RGBA, GL20.GL_UNSIGNED_SHORT); // but 16bpp int works perfectly?!
bufferBuilder.addColorTextureAttachment(GL20.GL_RGBA, GL20.GL_RGBA, GL20.GL_UNSIGNED_SHORT); // but 16bpp creates slight banding
if (hasDepth) bufferBuilder.addBasicDepthRenderBuffer();
this.bufferBuilder = bufferBuilder;
}
else {
FloatFrameBufferBuilder bufferBuilder = new FloatFrameBufferBuilder(width, height);
bufferBuilder.addFloatAttachment(GL30.GL_RGBA16F, GL30.GL_RGBA, GL30.GL_FLOAT, false);
bufferBuilder.addFloatAttachment(GL30.GL_RGBA16F, GL30.GL_RGBA, GL30.GL_FLOAT, false); // float16 will not create a banding
if (hasDepth) bufferBuilder.addBasicDepthRenderBuffer();
this.bufferBuilder = bufferBuilder;
}*/
// FIXME temporarily disabling Float16 -- has LWJGL bug that prevents setOpenGLEmulation on Windows (or is this Nvidia's issue?)
FrameBufferBuilder bufferBuilder = new FrameBufferBuilder(width, height);
bufferBuilder.addColorTextureAttachment(GL20.GL_RGBA, GL20.GL_RGBA, GL20.GL_UNSIGNED_SHORT); // but 16bpp int works perfectly?!
if (hasDepth) bufferBuilder.addBasicDepthRenderBuffer();