问题
I have some scientific image data that's coming out of a detector device in a 16 bit range which then gets rendered in an image. In order to display this data, I'm using OpenGL, because it should support ushorts as part of the library. I've managed to get this data into textures rendering on an OpenGL 1.4 platform, a limitation that is a requirement of this project.
Unfortunately, the resulting textures look like they're being reduced to 8 bits, rather than 16 bits. I test this by generating a gradient image and displaying it; while the image itself has each pixel different from its neighbors, the displayed texture is showing stripe patterns where all pixels next to one another are showing up as equal values.
I've tried doing this with GlDrawPixels, and the resulting image actually looks like it's really rendering all 16 bits.
How can I force these textures to display properly?
To give more background, the LUT (LookUp Table) is being determined by the following code:
String str = "!!ARBfp1.0\n" +
"ATTRIB tex = fragment.texcoord[0];\n" +
"PARAM cbias = program.local[0];\n" +
"PARAM cscale = program.local[1];\n" +
"OUTPUT cout = result.color;\n" +
"TEMP tmp;\n" +
"TXP tmp, tex, texture[0], 2D;\n" +
"SUB tmp, tmp, cbias;\n" +
"MUL cout, tmp, cscale;\n" +
"END";
Gl.glEnable(Gl.GL_FRAGMENT_PROGRAM_ARB);
Gl.glGenProgramsARB(1, out mFragProg);
Gl.glBindProgramARB(Gl.GL_FRAGMENT_PROGRAM_ARB, mFragProg);
System.Text.Encoding ascii = System.Text.Encoding.ASCII;
Byte[] encodedBytes = ascii.GetBytes(str);
Gl.glProgramStringARB(Gl.GL_FRAGMENT_PROGRAM_ARB, Gl.GL_PROGRAM_FORMAT_ASCII_ARB,
count, encodedBytes);
GetGLError("Shader");
Gl.glDisable(Gl.GL_FRAGMENT_PROGRAM_ARB);
Where cbias and cScale are between 0 and 1.
Thanks!
EDIT: To answer some of the other questions, the line with glTexImage:
Gl.glBindTexture(Gl.GL_TEXTURE_2D, inTexData.TexName);
Gl.glTexImage2D(Gl.GL_TEXTURE_2D, 0, Gl.GL_LUMINANCE, inTexData.TexWidth, inTexData.TexHeight,
0, Gl.GL_LUMINANCE, Gl.GL_UNSIGNED_SHORT, theTexBuffer);
Gl.glTexParameteri(Gl.GL_TEXTURE_2D, Gl.GL_TEXTURE_MIN_FILTER, Gl.GL_LINEAR); // Linear Filtering
Gl.glTexParameteri(Gl.GL_TEXTURE_2D, Gl.GL_TEXTURE_MAG_FILTER, Gl.GL_LINEAR); // Linear Filtering
theTexBuffer = null;
GC.Collect();
GC.WaitForPendingFinalizers();
The pixel format is set when the context is initialized:
Gdi.PIXELFORMATDESCRIPTOR pfd = new Gdi.PIXELFORMATDESCRIPTOR();// The pixel format descriptor
pfd.nSize = (short)Marshal.SizeOf(pfd); // Size of the pixel format descriptor
pfd.nVersion = 1; // Version number (always 1)
pfd.dwFlags = Gdi.PFD_DRAW_TO_WINDOW | // Format must support windowed mode
Gdi.PFD_SUPPORT_OPENGL | // Format must support OpenGL
Gdi.PFD_DOUBLEBUFFER; // Must support double buffering
pfd.iPixelType = (byte)Gdi.PFD_TYPE_RGBA; // Request an RGBA format
pfd.cColorBits = (byte)colorBits; // Select our color depth
pfd.cRedBits = 0; // Individual color bits ignored
pfd.cRedShift = 0;
pfd.cGreenBits = 0;
pfd.cGreenShift = 0;
pfd.cBlueBits = 0;
pfd.cBlueShift = 0;
pfd.cAlphaBits = 0; // No alpha buffer
pfd.cAlphaShift = 0; // Alpha shift bit ignored
pfd.cAccumBits = 0; // Accumulation buffer
pfd.cAccumRedBits = 0; // Individual accumulation bits ignored
pfd.cAccumGreenBits = 0;
pfd.cAccumBlueBits = 0;
pfd.cAccumAlphaBits = 0;
pfd.cDepthBits = 16; // Z-buffer (depth buffer)
pfd.cStencilBits = 0; // No stencil buffer
pfd.cAuxBuffers = 0; // No auxiliary buffer
pfd.iLayerType = (byte)Gdi.PFD_MAIN_PLANE; // Main drawing layer
pfd.bReserved = 0; // Reserved
pfd.dwLayerMask = 0; // Layer masks ignored
pfd.dwVisibleMask = 0;
pfd.dwDamageMask = 0;
pixelFormat = Gdi.ChoosePixelFormat(mDC, ref pfd); // Attempt to find an appropriate pixel format
if (!Gdi.SetPixelFormat(mDC, pixelFormat, ref pfd))
{ // Are we not able to set the pixel format?
BigMessageBox.ShowMessage("Can not set the chosen PixelFormat. Chosen PixelFormat was " + pixelFormat + ".");
Environment.Exit(-1);
}
回答1:
If you create a texture the 'type' parameter of glTexImage is only the data type your texture data is in before it is converted by OpenGL into its own format. To create a texture with 16 bit per channel you need something like GL_LUMINANCE16 as format (internal format remains GL_LUMINANCE). If there's no GL_LUMINANCE16 for OpenGL 1.4 check if GL_EXT_texture is available and try it with GL_LUMINANCE16_EXT.
One of these should work. However if it doesn't you can encode your 16 bit values as two 8 bit pairs with GL_LUMINANCE_ALPHA and decode it again inside a shader.
回答2:
I've never worked in depths higher (deeper) than 8bit per channel, but here's what I'd try first:
Turn off filtering on the texture and see how it affects the output.
Set texturing glHints to best quality.
回答3:
You could consider using a single channel floating point texture through one of the GL_ARB_texture_float, GL_ATI_texture_float or GL_NV_float_buffer extensions if the hardware supports it, I can't recall if GL 1.4 has floating point textures or not though.
来源：https://stackoverflow.com/questions/508448/opengl-16-bit-display-via-tao-c