Jump to content

  • Log In with Google      Sign In   
  • Create Account

BMP file problem


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
13 replies to this topic

#1 Shawn619   Members   -  Reputation: 342

Like
0Likes
Like

Posted 30 December 2012 - 02:09 PM

My opengl program loads BMP's like such:

 

//Makes the image into a texture, and returns the id of the texture
GLuint loadTexture(Image* image) {
GLuint textureId;
glGenTextures(1, &textureId);
 
//Make room for our texture
glBindTexture(GL_TEXTURE_2D, textureId);
 
//Tell OpenGL which texture to edit
 
 
//Map the image to the texture
glTexImage2D(GL_TEXTURE_2D,
 
//Always GL_TEXTURE_2D
0,
 
//0 for now
GL_RGB,
 
//Format OpenGL uses for image
image->width, image->height,
 
//Width and height
0,
 
//The border of the image
GL_RGB,
 
//GL_RGB, because pixels are stored in RGB format
GL_UNSIGNED_BYTE,
 
//GL_UNSIGNED_BYTE, because pixels are stored
 
 
//as unsigned numbers
image->pixels);
 
//The actual pixel data
 
 
return textureId; //Returns the id of the texture
}
 

 

 

with assistance from the imageloader.cpp&.h files:

 

//imageloader.cpp
#include
 
 
<assert.h>
#include
 
 
<fstream>
#include
 
 
"imageloader.h"
using
 
 
namespace std;
Image::Image(
 
char* ps, int w, int h) : pixels(ps), width(w), height(h) {
 
}
Image::~Image() {
 
 
delete[] pixels;
}
namespace
 
 
{
 
 
//Converts a four-character array to an integer, using little-endian form
 
 
int toInt(const char* bytes) {
 
 
return (int)(((unsigned char)bytes[3] << 24) |
((
 
unsigned char)bytes[2] << 16) |
((
 
unsigned char)bytes[1] << 8) |
(
 
unsigned char)bytes[0]);
}
 
 
 
//Converts a two-character array to a short, using little-endian form
 
 
short toShort(const char* bytes) {
 
 
return (short)(((unsigned char)bytes[1] << 8) |
(
 
unsigned char)bytes[0]);
}
 
 
 
//Reads the next four bytes as an integer, using little-endian form
 
 
int readInt(ifstream &input) {
 
 
char buffer[4];
input.read(buffer, 4);
 
 
return toInt(buffer);
}
 
 
 
//Reads the next two bytes as a short, using little-endian form
 
 
short readShort(ifstream &input) {
 
 
char buffer[2];
input.read(buffer, 2);
 
 
return toShort(buffer);
}
 
 
 
//Just like auto_ptr, but for arrays
 
 
template<class T>
 
 
class auto_array {
 
 
private:
T*
 
array;
 
 
mutable bool isReleased;
 
 
public:
 
 
explicit auto_array(T* array_ = NULL) :
 
 
array(array_), isReleased(false) {
}
 
auto_array(
 
const auto_array<T> &aarray) {
 
 
array = aarray.array;
isReleased = aarray.isReleased;
aarray.isReleased =
 
true;
}
 
~auto_array() {
 
 
if (!isReleased && array != NULL) {
 
 
delete[] array;
}
}
 
T* get()
 
const {
 
 
return array;
}
 
T &
 
operator*() const {
 
 
return *array;
}
 
 
 
void operator=(const auto_array<T> &aarray) {
 
 
if (!isReleased && array != NULL) {
 
 
delete[] array;
}
 
 
array = aarray.array;
isReleased = aarray.isReleased;
aarray.isReleased =
 
true;
}
 
T*
 
operator->() const {
 
 
return array;
}
 
T* release() {
isReleased =
 
true;
 
 
return array;
}
 
 
 
void reset(T* array_ = NULL) {
 
 
if (!isReleased && array != NULL) {
 
 
delete[] array;
}
 
 
array = array_;
}
 
T*
 
operator+(int i) {
 
 
return array + i;
}
 
T &
 
operator[](int i) {
 
 
return array[i];
}
};
}
Image* loadBMP(
 
const char* filename) {
ifstream input;
input.open(filename, ifstream::binary);
assert(!input.fail() || !
 
"Could not find file");
 
 
char buffer[2];
input.read(buffer, 2);
assert(buffer[0] ==
 
'B' && buffer[1] == 'M' || !"Not a bitmap file");
input.ignore(8);
 
 
int dataOffset = readInt(input);
 
 
 
//Read the header
 
 
int headerSize = readInt(input);
 
 
int width;
 
 
int height;
 
 
switch(headerSize) {
 
 
case 40:
 
 
//V3
width = readInt(input);
height = readInt(input);
input.ignore(2);
assert(readShort(input) == 24 || !
 
"Image is not 24 bits per pixel");
assert(readShort(input) == 0 || !
 
"Image is compressed");
 
 
break;
 
 
case 12:
 
 
//OS/2 V1
width = readShort(input);
height = readShort(input);
input.ignore(2);
assert(readShort(input) == 24 || !
 
"Image is not 24 bits per pixel");
 
 
break;
 
 
case 64:
 
 
//OS/2 V2
assert(!
 
"Can't load OS/2 V2 bitmaps");
 
 
break;
 
 
case 108:
 
 
//Windows V4
assert(!
 
"Can't load Windows V4 bitmaps");
 
 
break;
 
 
case 124:
 
 
//Windows V5
assert(!
 
"Can't load Windows V5 bitmaps");
 
 
break;
 
 
default:
assert(!
 
"Unknown bitmap format");
}
 
 
 
//Read the data
 
 
int bytesPerRow = ((width * 3 + 3) / 4) * 4 - (width * 3 % 4);
 
 
int size = bytesPerRow * height;
auto_array<
 
char> pixels(new char[size]);
input.seekg(dataOffset, ios_base::beg);
input.read(pixels.get(), size);
 
 
 
//Get the data into the right format
auto_array<
 
char> pixels2(new char[width * height * 3]);
 
 
for(int y = 0; y < height; y++) {
 
 
for(int x = 0; x < width; x++) {
 
 
for(int c = 0; c < 3; c++) {
pixels2[3 * (width * y + x) + c] =
pixels[bytesPerRow * y + 3 * x + (2 - c)];
}
}
}
 
input.close();
 
 
return new Image(pixels2.release(), width, height);
}
 
 
 
 
 
 
 
 
 

 

 

//imageloader.h
#ifndef
 
 
IMAGE_LOADER_H_INCLUDED
#define
 
 
IMAGE_LOADER_H_INCLUDED
//Represents an image
class
 
 
Image {
 
 
public:
Image(
 
char* ps, int w, int h);
~Image();
 
 
 
/* An array of the form (R1, G1, B1, R2, G2, B2, ...) indicating the
* color of each pixel in image. Color components range from 0 to 255.
* The array starts the bottom-left pixel, then moves right to the end
* of the row, then moves up to the next column, and so on. This is the
* format in which OpenGL likes images.
*/
 
 
char* pixels;
 
 
int width;
 
 
int height;
};
//Reads a bitmap image from file.
Image* loadBMP(
 
const char* filename);
#endif

 

 

I get the error: "Can't load windows V4 bitmaps" when i load this image:

j74u8o.jpg

 

But it works PERFECTLY when i load this image->

wvsxvq.jpg

 

 

Why can't i load the 1st image? Both images are 256x256, 24bit and BMP files.



Sponsor:

#2 Brother Bob   Moderators   -  Reputation: 8197

Like
1Likes
Like

Posted 30 December 2012 - 03:07 PM

You have the error message and you have the code, so see the code where the error message comes from, and then look back in the code to see why the code gets to that point.



#3 radioteeth   Prime Members   -  Reputation: 1054

Like
2Likes
Like

Posted 30 December 2012 - 03:39 PM

there are multiple .bmp formats, and your first image is a funky microsoft one, whereas the second one is a standard 24 bpp bitmap... this is what you can expect from extracting random images from random places.. open the first one in an image editing program (that is NOT Windows Paint) and re-save it as a standard 24-bit bitmap (that means no RLE or anything else funky)... your image loading code relies on the bitmap being stored as essentially raw pixel data.



#4 Shawn619   Members   -  Reputation: 342

Like
0Likes
Like

Posted 30 December 2012 - 07:43 PM

@Brother Bob, i can see the error is because "headersize==108", but i know nothing about the structure of .BMP files.

 

@radioteeth, i tried to save the first image through Gimp 2.0->export->windows BMP->24 bitmap, but with no luck.



#5 radioteeth   Prime Members   -  Reputation: 1054

Like
0Likes
Like

Posted 30 December 2012 - 11:41 PM

did you check 'do not write colorspace information' in the compatibility options dropdown while exporting?  it falls under the "anything else funky" category I mentioned above.

 

edit: to clarify, writing colorspace information is funky, so make sure it's checked this time.


Edited by radioteeth, 30 December 2012 - 11:43 PM.


#6 Shawn619   Members   -  Reputation: 342

Like
0Likes
Like

Posted 31 December 2012 - 12:26 AM

Initially i had it unchecked(by default) and got the error. Now i tried to export with it checked, but it gave me the same error, unfortunately.



#7 GeneralQuery   Crossbones+   -  Reputation: 1263

Like
1Likes
Like

Posted 31 December 2012 - 04:42 AM

Use an image loading library (such as DevIL) to avoid these esoteric problems of hand rolling your own.

#8 Shawn619   Members   -  Reputation: 342

Like
0Likes
Like

Posted 31 December 2012 - 02:27 PM

The problem with other image loading libraries is they hard to setup and often never have a clear tutorial on how to use them. I tried to setup DevIL a few days ago and multiple problems occured.



#9 GeneralQuery   Crossbones+   -  Reputation: 1263

Like
0Likes
Like

Posted 31 December 2012 - 03:07 PM

What's a more productive use of your time: learning to use a library that can effortlessly load and save images in (for all intents and purposes) any format you desire or trying to fix tedious problems in a custom BMP loader? DevIL can even do the heavy lifting for OpenGL texture creation for you.

 

http://openil.sourceforge.net/tuts/tut_1/index.htm



#10 Shawn619   Members   -  Reputation: 342

Like
0Likes
Like

Posted 31 December 2012 - 04:29 PM

All i need is to know how  to get the images in the .BMP loader's specification.

 

Im sorry, but i can't follow your linked tutorial because it is 11 years old and deprecated.



#11 GeneralQuery   Crossbones+   -  Reputation: 1263

Like
0Likes
Like

Posted 31 December 2012 - 04:41 PM

All i need is to know how  to get the images in the .BMP loader's specification.

 

Im sorry, but i can't follow your linked tutorial because it is 11 years old and deprecated.

The steps are still valid in MSVS.



#12 radioteeth   Prime Members   -  Reputation: 1054

Like
1Likes
Like

Posted 31 December 2012 - 06:26 PM

I made two bitmaps in gimp, one with colorspace information, one without, opened them up in a hex editor, found the byte your code does the switch/case statement on, with colorspace info included the value is 108, without it the value is 40, which is what your code relies on in order to operate. Therefore, either your GIMP version is broken (ignoring the checkbox?) or you're missing something..

 

On a personal note, this is why I study all file formats and write my own loading code from scratch. I avoid jpeg format because there are seemingly endless variants. Your code looks like it should work just fine with the bitmaps output by GIMP.

 

If you're getting the SAME windows v4 error, then you are obviously not replacing the actual bitmap file that your program is trying to load.



#13 Shawn619   Members   -  Reputation: 342

Like
0Likes
Like

Posted 31 December 2012 - 10:14 PM

I made two bitmaps in gimp, one with colorspace information, one without, opened them up in a hex editor, found the byte your code does the switch/case statement on, with colorspace info included the value is 108, without it the value is 40, which is what your code relies on in order to operate. Therefore, either your GIMP version is broken (ignoring the checkbox?) or you're missing something..

 

On a personal note, this is why I study all file formats and write my own loading code from scratch. I avoid jpeg format because there are seemingly endless variants. Your code looks like it should work just fine with the bitmaps output by GIMP.

 

If you're getting the SAME windows v4 error, then you are obviously not replacing the actual bitmap file that your program is trying to load.

Hey Radio, you're completely right and i appreciate you going the extra step for me. It loads perfectly now. Apparently exporting over the same file does not write over it. What i  should have done was use GIMP's "overwrite" function, which didn't exist years ago when i used GIMP. Thanks Radio :)



#14 radioteeth   Prime Members   -  Reputation: 1054

Like
0Likes
Like

Posted 02 January 2013 - 11:50 PM

you are welcome, glad I could help ;D






Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS