Texture prob
Author
This is a really wierd one im making a tile based game right, now the problem is it only likes multiples of 4 images i try loading in a 17 x 16 image and the computer goes super slow ive been through the code over and over but i cant see whats going on.
The file is in TGA format surely opengl can take in none multiple of 4 images ?
The loader works fine textured quad comes on the screen with the texture except the frame rate drops to 2fps if i resize the image to 16 x 16 it works fine 100+ fps.
Any ideas what can cause this sorta problem ?
My best guess is the tga loader i had a similar problem with a bitmap loader with it having extra bytes at the end of each image row because of it being none multiples of 4.
thx alot
fishy
In OpenGL, texture must have power of 2 dimensions(1,2,4,8,16,32...). The weird thing is that 17x16 works, it shouldn't(even if it works slowly). Can you post the code of the loader?
Author
So theres no way i can get it to load in these images except for resizing it first ?
See i had an idea of having 16x16 tiles and a column of pixels before the tile then the colour says what that tile is i.e. red for solid, green for passable that sorta thing.
Now what im doing at the moment is loading in the whole image the when i go to draw i move the texture coordinates 1 pixel to the right so i miss the 1st line that says if its a wall or not but that obviously doesnt work :(. This actually works on my computer but every other computer ive tried it on goes 2fps or loads a white block.
The reason i did it like this was because the level scrolls but the player stays still and im keeping all the wall data in an int array so i can quickly go ok the player is now at 10,0 is he in a wall or not. Its a bit more complex than that really. I also tile each column along so i can use column 0 over and over so i get a nice flat level instead of having a giant image file.
Sure its not the best tho half ripped from opengl game programming book,
[Edited by - fishleg003 on November 5, 2005 1:32:11 PM]
See i had an idea of having 16x16 tiles and a column of pixels before the tile then the colour says what that tile is i.e. red for solid, green for passable that sorta thing.
Now what im doing at the moment is loading in the whole image the when i go to draw i move the texture coordinates 1 pixel to the right so i miss the 1st line that says if its a wall or not but that obviously doesnt work :(. This actually works on my computer but every other computer ive tried it on goes 2fps or loads a white block.
The reason i did it like this was because the level scrolls but the player stays still and im keeping all the wall data in an int array so i can quickly go ok the player is now at 10,0 is he in a wall or not. Its a bit more complex than that really. I also tile each column along so i can use column 0 over and over so i get a nice flat level instead of having a giant image file.
Sure its not the best tho half ripped from opengl game programming book,
typedef struct tag_TARGAFILEHEADER{unsigned char imageIDLength;unsigned char colorMapType;unsigned char imageTypeCode;short int colorMapOrigin; short int colourMapLength;short int colourMapEntrySize;short int imageXOrigin;short int imageYOrigin;short int imageWidth;short int imageHeight;unsigned char bitCount;unsigned char imageDescriptor;}TARGAFILEHEADER;typedef struct tag_TGAFILE{ unsigned char imageTypeCode; short int imageWidth; short int imageHeight; unsigned char bitCount; unsigned char *imageData;}TGAFILE;int LoadTGAFile(char *filename, TGAFILE *tgaFile){ FILE *fileptr; unsigned char ucharBad; short int sintBad; long imageSize; int colorMode; long imageIdx; unsigned char colorSwap; unsigned char colorSwap2; unsigned char colorSwap3; fileptr = fopen(filename, "rb"); if(!fileptr) return 0; //read first 2 bites that are not needed fread(&ucharBad, sizeof(unsigned char), 1, fileptr); fread(&ucharBad, sizeof(unsigned char), 1, fileptr); //read image type fread(&tgaFile->imageTypeCode, sizeof(unsigned char), 1, fileptr); //read 13 bytes of data we dont need fread(&sintBad, sizeof(short int), 1, fileptr); fread(&sintBad, sizeof(short int), 1, fileptr); fread(&ucharBad, sizeof(unsigned char), 1, fileptr); fread(&sintBad, sizeof(short int), 1, fileptr); fread(&sintBad, sizeof(short int), 1, fileptr); //read image size fread(&tgaFile->imageWidth, sizeof(short int), 1, fileptr); fread(&tgaFile->imageHeight, sizeof(short int), 1, fileptr); //read bit depth fread(&tgaFile->bitCount, sizeof(unsigned char), 1, fileptr); //read 1 byte of data not needed fread(&ucharBad, sizeof(unsigned char), 1, fileptr); //colourmode -> 3 = BGR, 4 = BGRA colorMode = tgaFile->bitCount/8; imageSize = (tgaFile->imageWidth * tgaFile->imageHeight) * colorMode; //allocate memory for image tgaFile->imageData = (unsigned char*)malloc(sizeof(unsigned char) * imageSize); //read in image data fread(tgaFile->imageData, sizeof(unsigned char), imageSize, fileptr); //change BGR to RGB for(imageIdx = 0; imageIdx < imageSize; imageIdx += colorMode) { colorSwap = tgaFile->imageData[imageIdx];//red - red colorSwap2 = tgaFile->imageData[imageIdx+2]; //blue - green colorSwap3 = tgaFile->imageData[imageIdx+1]; //green - blue tgaFile->imageData[imageIdx] = colorSwap2; tgaFile->imageData[imageIdx+1] = colorSwap3; tgaFile->imageData[imageIdx+2] = colorSwap; } //close file fclose(fileptr); return 1;}[Edited by - fishleg003 on November 5, 2005 1:32:11 PM]
OpenGL textures have to be power of two to be used, however there are numerous extensions which relax this and allow non PO2 textures (some extensions with various restrictions, like no tiling).
However recently nVidia's drivers have started supporting ARB non-power-of-two (which doesn't have any restrictions). The snag is that they 'support' this even on hardware that doesn't do it, so you get a silent software fallback which is obviously somewhat slow.
For sprites etc. one way around this is to include multiple sprites in a single texture and mess with the texture coords to select the right sub chunks within it.
However recently nVidia's drivers have started supporting ARB non-power-of-two (which doesn't have any restrictions). The snag is that they 'support' this even on hardware that doesn't do it, so you get a silent software fallback which is obviously somewhat slow.
For sprites etc. one way around this is to include multiple sprites in a single texture and mess with the texture coords to select the right sub chunks within it.
Author
Im not sure what you mean by,
"one way around this is to include multiple sprites in a single texture and mess with the texture coords to select the right sub chunks within it."
I thought that is what i am doing but maybe im wrong.
Im stuck no idea how i can get around this any more ideas ?
thx alot for help.
"one way around this is to include multiple sprites in a single texture and mess with the texture coords to select the right sub chunks within it."
I thought that is what i am doing but maybe im wrong.
Im stuck no idea how i can get around this any more ideas ?
thx alot for help.
Quote:Original post by fishleg003
Im not sure what you mean by,
"one way around this is to include multiple sprites in a single texture and mess with the texture coords to select the right sub chunks within it."
I thought that is what i am doing but maybe im wrong.
He means you should make the large texture containing all the sprites a power-of-two, leave in white spaces if you can't get it in the exact dimensions. Then play around changing the texture coordinates, and use the appropriate ones in your game! [smile]
Your idea of having the column of key colors is very good, but another approach would be to have an index file before loading in the sprites, which tells you that - image1.tga - is a solid object etc.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
