Writing a CDC to a jpeg file

[robbiegregg]

Hi all
I've been trying all day to sort this out so I was hoping someone could help!!

Basically, I am trying to write the contents of a CDC to a JPEG file and I am trying to use libjpeg.lib. Below is a simplified snippet of the code I am using at present. What I am intending the code to do is to create a picture with a square and a circle and then write this to a JPEG file. It manages to create a JPEG file with the right size however it is completely black so obviously I am doing something wrong! Any help would be greatly appreciated.

Thanks very much for reading,

Robbie

(ps. I am using Visual C++ v6 and am trying to do this within a MFC dialog based project)

void CNucPlot::SaveFile()
{
	CDC myDC;
	HBITMAP myBMP;
	myDC.CreateCompatibleDC(NULL);

	//  Set pictures size;
	int Width = 500;
	int Height = 500;

	// Draw rectangle and an ellipse to myDC ..
	CBrush newColour(RGB(10,10,10));
	myDC.SelectObject(newColour);
	CPen lColourPen (PS_SOLID, 1,RGB(50,100,150));
	myDC.SelectObject(&lColourPen); // 100

	myDC.SetBkColor(RGB(255,255,255));
	myDC.Rectangle(10,20,100,90);
	CRect Rect;
	Rect.left   = 10;
	Rect.top    = 20;
	Rect.right  = 100;
	Rect.bottom = 90;
	myDC.Ellipse(Rect);

	myBMP = CreateCompatibleBitmap(myDC,Width,Height);
	SelectObject(myDC,myBMP);

	struct jpeg_compress_struct cinfo;
	struct jpeg_error_mgr jerr;
	FILE * outfile;
	JSAMPLE* scanline;
	COLORREF pixel;

	cinfo.err = jpeg_std_error(&jerr);
	jpeg_create_compress(&cinfo);

	outfile = fopen("robbie.jpg", "wb");
	if(outfile == NULL) return;

	jpeg_stdio_dest(&cinfo, outfile);
	cinfo.image_width = Width;
	cinfo.image_height = Height;
	cinfo.input_components = 3;
	cinfo.in_color_space = JCS_RGB;	
	jpeg_set_defaults(&cinfo);

	int quality = 90;
	if(quality < 0) quality = 0;
	if(quality > 100) quality = 100;

	jpeg_set_quality(&cinfo, quality, FALSE);
	jpeg_start_compress(&cinfo, TRUE);
	scanline = new JSAMPLE[Width*3];

	for(int posy = 0; posy < Height; posy++) 
	{
		for(int posx = 0; posx < Width; posx++)
		{
			pixel = GetPixel(myDC, posx, posy);
			scanline[posx*3+0] = GetRValue(pixel);
			scanline[posx*3+1] = GetGValue(pixel);
			scanline[posx*3+2] = GetBValue(pixel);

		}
		jpeg_write_scanlines(&cinfo, &scanline, 1);
	}
	jpeg_finish_compress(&cinfo);
	jpeg_destroy_compress(&cinfo);

	delete scanline;
	fclose(outfile);
}

[D_Drmmr]

Please put [code][/code] tags around your code and indent it properly to make it more readable.

Try painting the entire DC white before you paint something else on it.

[robbiegregg]

Hi, I've edited my original post with the tags. I've also added the line myDC.SetBkColor(RGB(255,255,255)); but the JPEG image created is still black!

Any other ideas anyone?

Thanks

[VladimirF]

You have a lot of reading to do about GDI…
First, ignore the JPEG stuff for now.
A newly created DC has a 1x1 monochrome bitmap selected into it.
All your drawings have no effect on that DC,
CreateCompatibleBitmap() creates bitmap *compatible* to that DC (also monochrome), so colored output to that makes a little sense anyway.
The steps should be:

• create DC compatible to a screen;
• create bitmap with a required size, also compatible to a screen;
• select that bitmap into DC;
• NOW draw whatever you want on that DC;
• and don’t forget to clean up after yourself.

I would like to suggest a great GDI debugging tool that let you examine content of your GDI objects (memory DC, bitmap, etc.) while you debug your code, line-by-line.

[robbiegregg]

Thanks VladimirF

I've taken your advice and re-wrote the code as shown below:

void CNucPlot::SaveFile()
{
	CPaintDC dc(this); // device context for painting
	//  Set pictures size;
	int Width = 700;
	int Height = 700;

  //  get the area to draw in.
	CRect rcClient;
	rcClient.top = 0;
	rcClient.left = 0;
	rcClient.bottom = Height;
	rcClient.right = Width;

  //  create a memory dc for double buffering.
	CMemDC myDC(&dc, &rcClient);

	myDC.SetBkColor(RGB(255,255,255)); 
	int nFontHeight = Height / 80.0f;
	CBrush newColour(RGB(255,199,10));
	myDC.SelectObject(newColour);
	CPen lColourPen (PS_SOLID, 5,RGB(50,100,150));
	myDC.SelectObject(&lColourPen); // 100

	myDC.Rectangle(10,10,Width-10,Height-10);
	myDC.Ellipse(10,10,Width-10,Height-10);

	// JPEG Saving ..
	struct jpeg_compress_struct cinfo;
	struct jpeg_error_mgr jerr;
	FILE * outfile;
	JSAMPLE* scanline;
	COLORREF pixel;

	cinfo.err = jpeg_std_error(&jerr);
	jpeg_create_compress(&cinfo);

	outfile = fopen("robbie.jpg", "wb");
	if(outfile == NULL) return;

	jpeg_stdio_dest(&cinfo, outfile);
	cinfo.image_width = Width;
	cinfo.image_height = Height;
	cinfo.input_components = 3;
	cinfo.in_color_space = JCS_RGB;	
	jpeg_set_defaults(&cinfo);

	int quality = 90;
	if(quality < 0) quality = 0;
	if(quality > 100) quality = 100;

	jpeg_set_quality(&cinfo, quality, FALSE);
	jpeg_start_compress(&cinfo, TRUE);
	scanline = new JSAMPLE[Width*3];
//	MessageBox("Finished setting up JPEG - now go through each JPEG line\n");

	for(int posy = 0; posy < Height; posy++) 
	{
		for(int posx = 0; posx < Width; posx++)
		{
			pixel = GetPixel(myDC, posx, posy);
			scanline[posx*3+0] = GetRValue(pixel);
			scanline[posx*3+1] = GetGValue(pixel);
			scanline[posx*3+2] = GetBValue(pixel);

		}
		jpeg_write_scanlines(&cinfo, &scanline, 1);
	}
	jpeg_finish_compress(&cinfo);
	jpeg_destroy_compress(&cinfo);

	delete scanline;
	fclose(outfile);
}

Now it writes correctly to a JPEG file. My problem with this code is that the picture displays on screen as well. Does anyone know how I can modify the code so that it only generates the JPEG file?

Also, when I try to generate a really big JPEG (say 2000 x 2000), it takes along time to complete. It might just be that it is a CPU intensive task to create a big JPEG. However does anyone know a better way of reading the CDC and dumping it to a JPEG file?

Thanks

Robbie

[hoxsiew]

From my experience, GetPixel() is probably the bottleneck in your code. It is best to grab a pointer to the buffer directly and iterate through it directly than calling GetPixel for every pixel. Use GetDIBits() to grab a pointer to the raw buffer, but keep in mind that the buffer is a bottom-up bitmap so you have to iterate backwards.

[Karpatuz]

what about this

#include //For Image
using namespace Gdiplus; //For Image

void BitmapUtils::SaveBitmapToFile(HDC hdc,CRect rect)
{
HDC hMemDC = CreateCompatibleDC(hdc);
HBITMAP hBitmap = CreateCompatibleBitmap(hdc,rect.right, rect.Height() + rect.bottom);
if (hBitmap)
{
HBITMAP hOld = (HBITMAP) SelectObject(hMemDC, hBitmap);
BitBlt(hMemDC, 0, 0, rect.right, rect.Height() + rect.bottom, hdc, 0, 0, SRCCOPY);
SelectObject(hMemDC, hOld);
DeleteDC(hMemDC);
::ReleaseDC(NULL, hdc);
}
CImage im;
im.Attach(hBitmap);
im.Save( "c:\\image.jpg", ImageFormatJPEG);
}