Loading a bitmap in OpenGL
Author
I want to load a bitmap image from a resource file in Visual Studio. This is the win32 function to do it: HBITMAP LoadBitmap(HINSTANCE hInstance, LPCTSTR lpBitmapName); But when I have this bitmap handler how can I use it in OpenGL with this function: glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, 256, 256, 0, GL_RGB, GL_UNSIGNED_BYTE, bmp[0]->data) ?
(The problem is I don´t want to load the image from file.)
Edited by - Gandalf on 6/18/00 12:44:27 PM
Author
Thanks!
How many bytes should bee copied if the bmp is 256 x 256, 24 bits?
Gandalf the White
How many bytes should bee copied if the bmp is 256 x 256, 24 bits?
Gandalf the White
Author
Here is little code
I don´t get a error but the bitmap look very ugly!
Gandalf the White
Edited by - Gandalf on June 19, 2000 4:23:18 AM
HBITMAP bitmap = LoadBitmap(GetModuleHandle(NULL),MAKEINTRESOURC(IDB_BITMAP3));unsigned char *data;GetBitmapBits(bitmap, 256*256*3, &data);glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, 256, 256, 0, GL_RGB, GL_UNSIGNED_BYTE, &data); I don´t get a error but the bitmap look very ugly!
Gandalf the White
Edited by - Gandalf on June 19, 2000 4:23:18 AM
How ugly? Is it just that the colors are all screwed up? or is it you cannot recognise your original image?
- make sure your image is 256*256, 24 bits per pixel
- if I remember, I had once to reverse the component order (bitmap is BGR and OpenGL expects RGB)
- and yes, data_size = bmp_width * bmp_height * bytes_per_pixel
maybe you should use LoadImage & GetDIBits instead.
ps: in your code, I assume you allocate enough space for the data!
- make sure your image is 256*256, 24 bits per pixel
- if I remember, I had once to reverse the component order (bitmap is BGR and OpenGL expects RGB)
- and yes, data_size = bmp_width * bmp_height * bytes_per_pixel
maybe you should use LoadImage & GetDIBits instead.
ps: in your code, I assume you allocate enough space for the data!
I don''t know how in hell you aren''t getting an access violation there. You need to allocate memory.
in c++ do this:
-BacksideSnap-
in c++ do this:
int size = 256*256*3data = new unsigned char[size] -BacksideSnap-
Gandalf, are you using 8 bits or 24 bits BMP''s?
if it''s 8 bit, it looks ugly because 8 bit images doesn''t have the colors stored as 24 bit images. In 8 bit, there''s a palette of colors, which is simply an array of RGB values, and each pixel of the image is stored as one byte, that represents the index of the palette.
You have to do a conversion, Here''s some pseudo-code:
- Create an array of 256 entries of the PALETTEENTRY structure (Pal)
- Allocate Width*Height bytes of memory (In)
- Allocate Width*Height*3 bytes of memory (Out)
- Use LoadBitmap to get an HBITMAP
- Use GetDIBColorTable to fill Pal
- Use GetBitmapBits to fill In
now, for (i=0; i < width*height-1;i++) do
Out[i*3+0] = Pal[In].peRed;
Out[i*3+1] = Pal[In].peGreen;<br> Out[i*3+2] = Pal[In].peBlue;<br><br>and use Out in the glTexImage2D:<br>glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, width, height, 0, GL_RGB, GL_UNSIGNED_BYTE, &Out);<br><br>and that''s it! I don''t remember well, but I recall an extension to use 8 bit textures… but I don''t know how to use it.<br><br>For 24 bit images, the only problem is that bitmaps are stored in BGR, and not in RGB (don''t ask me why…), so you have to exchange the bytes, as Jehan said.<br><br>Hope that helps, </i> <br><br>Nicodemus.<br><br>—-<br><I>"When everything goes well, something will go wrong."</I> - Murphy<br>
if it''s 8 bit, it looks ugly because 8 bit images doesn''t have the colors stored as 24 bit images. In 8 bit, there''s a palette of colors, which is simply an array of RGB values, and each pixel of the image is stored as one byte, that represents the index of the palette.
You have to do a conversion, Here''s some pseudo-code:
- Create an array of 256 entries of the PALETTEENTRY structure (Pal)
- Allocate Width*Height bytes of memory (In)
- Allocate Width*Height*3 bytes of memory (Out)
- Use LoadBitmap to get an HBITMAP
- Use GetDIBColorTable to fill Pal
- Use GetBitmapBits to fill In
now, for (i=0; i < width*height-1;i++) do
Out[i*3+0] = Pal[In].peRed;
Out[i*3+1] = Pal[In].peGreen;<br> Out[i*3+2] = Pal[In].peBlue;<br><br>and use Out in the glTexImage2D:<br>glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, width, height, 0, GL_RGB, GL_UNSIGNED_BYTE, &Out);<br><br>and that''s it! I don''t remember well, but I recall an extension to use 8 bit textures… but I don''t know how to use it.<br><br>For 24 bit images, the only problem is that bitmaps are stored in BGR, and not in RGB (don''t ask me why…), so you have to exchange the bytes, as Jehan said.<br><br>Hope that helps, </i> <br><br>Nicodemus.<br><br>—-<br><I>"When everything goes well, something will go wrong."</I> - Murphy<br>
Author
I´m using 24 bits bitmaps. The texture not just looking ugly, it must be something wrong with the color bytes. But I can recognise my original image. Switch from RGB to BGR? Something like this, or what?
My texture still looking bad.
Gandalf the White
Edited by - Gandalf on June 20, 2000 3:46:36 AM
unsigned char temp;for(int i=0; i<(256*256*3); i++)if(i%2==0 && i>=2) { temp = data ; data<i> = data[i-2]; data[i-2] = temp;} My texture still looking bad.
Gandalf the White
Edited by - Gandalf on June 20, 2000 3:46:36 AM
Hi Gandalf,
There is an extension called GL_BGR_EXT / GL_BGRA_EXT , which allows to load the channels in reverse order. The ext is supported by Nvidia ICDs since RivaTNT, maybe even earlier. But I dont know whether it is supported by other ipmlentations. But you can also use RAW or TGA files (onvert them with Photoshop for example), where the first one of them is very easy to use and allways in the right order.
???
Why do you write that in the for slope
temp=data; instead of temp=data ;<br><br>shouldn't it look that way<br><br>int a=3;<br>for(int i=0;i<(256*256*3);i++)<br>{<br> if(a==3)<br> {<br> temp = data <i> ;<br> data = data[i+2];<br> data[i+2] = temp;<br> a=0; <br> }<br> a++; <br>}<br>Hope that helps !<br><br>How do you make these source code frames ?<br><br>Edited by - TheMummy on June 20, 2000 5:18:33 AM </i> <br><br>Edited by - TheMummy on June 20, 2000 5:24:32 AM<br><br>Had to edit it because the forum always creats out of <i> <br><br>Edited by - TheMummy on June 20, 2000 5:26:12 AM </i>
There is an extension called GL_BGR_EXT / GL_BGRA_EXT , which allows to load the channels in reverse order. The ext is supported by Nvidia ICDs since RivaTNT, maybe even earlier. But I dont know whether it is supported by other ipmlentations. But you can also use RAW or TGA files (onvert them with Photoshop for example), where the first one of them is very easy to use and allways in the right order.
???
Why do you write that in the for slope
temp=data; instead of temp=data ;<br><br>shouldn't it look that way<br><br>int a=3;<br>for(int i=0;i<(256*256*3);i++)<br>{<br> if(a==3)<br> {<br> temp = data <i> ;<br> data = data[i+2];<br> data[i+2] = temp;<br> a=0; <br> }<br> a++; <br>}<br>Hope that helps !<br><br>How do you make these source code frames ?<br><br>Edited by - TheMummy on June 20, 2000 5:18:33 AM </i> <br><br>Edited by - TheMummy on June 20, 2000 5:24:32 AM<br><br>Had to edit it because the forum always creats out of <i> <br><br>Edited by - TheMummy on June 20, 2000 5:26:12 AM </i>
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
