View Full Version : tga loading problems

12-12-2002, 05:10 PM
my tga loader isn't loading tga's! grrr... if you could take a look at my loader and spot any errors that i might have missed that would help me out alot. Thanx!

void tga: http://www.opengl.org/discussion_boards/ubb/biggrin.gifefine(){
glTexImage2D(GL_TEXTURE_2D,0,m_type,m_Width,m_Heig ht,0,GL_RGB,GL_UNSIGNED_BYTE,m_data);
void tga::GenerateID(){

void tga::Bind(){

void tga::Filter(){

bool tga::Load(char *filename){
fstream file;
long x = 0;
unsigned char rgbSwitch;

file.open(filename,ios::in | ios::binary);
return false;
file.read((char*)&m_type,sizeof(unsigned char));
if((m_type != 2) | | (m_type != 3)){
return false;
file.read((char*)&m_Width,sizeof(short int));
file.read((char*)&m_Height,sizeof(short int));
file.read((char*)&m_bites,sizeof(unsigned char));

m_channels = m_bites / 8;
m_size = m_Width * m_Height;
m_size *= m_channels;
m_data = new unsigned char[m_size];

file.read((char*)&m_data,sizeof(unsigned char));

while(x < m_size){
rgbSwitch = m_data[x];
m_data[x] = m_data[x + 2];
m_data[x + 2] = rgbSwitch;
x += m_channels;
return true;

Old GLman
12-12-2002, 05:39 PM
file.read((char*)&m_data,sizeof(unsigned char));

Your only reading in one byte of data.

12-12-2002, 05:44 PM
i changed the line to:
file.read((char*)&m_data,sizeof(unsigned char)*m_size);
and it still doesnt work!

12-12-2002, 06:16 PM
i think i found were the error could be happening:

if((m_type != 2) | | (m_type != 3)){
return false;

i dont think im putting in the right value for m_type. i want to skip the first 2 bytes of data of the header and get the type. i try to do this using the file.seekg(2,ios::beg); but i think that im using it wrong and putting wrong data into m_type. how can i skip data in a file/fix this problem?

Old GLman
12-12-2002, 06:20 PM
You want to check the third byte in the file for either 2 or 3. 2 being black and white, 3 being true color, both being uncompressed. Check the spec for the other identifiers. www.wotsit.org (http://www.wotsit.org)


You should be doing a logical AND (&&) rather than a logical OR.

Old GLman

[This message has been edited by Old GLman (edited 12-12-2002).]

12-15-2002, 09:10 AM
Yeah, try wotsi. http://www.opengl.org/discussion_boards/ubb/smile.gif


You are right about your diagnose of the m_type values not being the appropriate ones...
The 1, 2, 3, 4 are simply the number of channels (bytes) you are reading per pixel. But there is no such thing in the header of a TGA file. Instead, you read the number of bits per pixel (bpp). So, you would have 8, 16, 24, 32.
Well, where do these 1, 2, 3, 4 come from?
You divide the bpp, which is the third unsigned char in the header as you are reading, and divide it by 8. 8 is the number of bits in a byte. http://www.opengl.org/discussion_boards/ubb/biggrin.gif

You will also have more problems if the tga is compressed or palletted or so and so and more sos.

8 -> grayscale
16 -> Why use it? Palletted maybe?
24 -> RGB
32 -> RGBA

If you want you can use one I wrote.
Download the normal map generator from: http://cheo.resnet.wayne.edu/miguel/
You can use and abuse it.

Good luck.

[This message has been edited by mancha (edited 12-15-2002).]