I am using the FreeImage library to store and manipulate bitmap data. Part of my code requires me to take a screenshot of a window (in Windows), and store it as a FBITMAP* (FreeImage's bitmap data structure). The method I have worked out of doing this involves two steps: capture the image, then convert it to a FBITMAP*.
To capture the image, I do something like this:
HWND window; // Assume this is a valid handle to my window
int width; // width of window client area
int height; // height of window client area
HDC windowDC = GetDC(window);
HDC captureDC = CreateCompatibleDC(windowDC);
HBITMAP screenshot = CreateCompatibleBitmap(windowDC, width, height);
SelectObject(captureDC, screenshot);
BitBlt(captureDC, 0, 0, width, height,
captureDC, 0, 0, SRCCOPY|CAPTUREBLT);
ReleaseDC(window, windowDC);
DeleteDC(captureDC);
FreeImage provides a function which returns a raw pointer to the pixel data:
BYTE* FreeImage_GetBits(FBITMAP*)
The FAQ explains that a HBITMAP (WinAPI handle to a bitmap) can be converted to a FBITMAP* using GetDIBits, which takes a source HBITMAP and a destination raw pointer as arguments, and copies the pixel data from one to the other.
The problem with this approach is that I have copied the data twice - once in the BitBlt from the window DC to the HBITMAP selected in the memory DC, and then again from the HBITMAP to the FreeImage memory buffer. I wish to remove this inefficiency and copy the data directly to my raw pointer in the BitBlt operation. For this to work, I need a memory DC which has a HBITMAP selected into it, and where that HBITMAP points to my memory buffer instead of to memory that Windows allocated for it.
How can I achieve this?
Related
I have an HBITMAP handle to a device-dependent bitmap, resulting from this code:
// copy screen to bitmap
HDC hScreen = GetDC(NULL);//don't use hwnd, it doesn't work for non native windows
HDC hDC = CreateCompatibleDC(hScreen);
HBITMAP hBitmap = CreateCompatibleBitmap(hScreen, w, h);
HGDIOBJ old_obj = SelectObject(hDC, hBitmap);
BOOL bRet = BitBlt(hDC, 0, 0, w, h, hScreen, x1, y1, SRCCOPY);
I want to access (read-only) the bits of the bitmap, read some values, do some calculations, then discard it.
According to this answer, I shouldn't use GetDIBits() or GetBitmapBits() since they copy the data, but instead I should use:
BITMAP bitmap;
GetObject(hBitmap, sizeof(bitmap), (LPVOID)&bitmap);
But according to GetObject's documentation:
If hgdiobj is a handle to a bitmap created by calling CreateDIBSection, and the specified buffer is large enough, the GetObject function returns a DIBSECTION structure. In addition, the bmBits member of the BITMAP structure contained within the DIBSECTION will contain a pointer to the bitmap's bit values.
If hgdiobj is a handle to a bitmap created by any other means, GetObject returns only the width, height, and color format information of the bitmap. You can obtain the bitmap's bit values by calling the GetDIBits or GetBitmapBits function.
Which is pointed to in a comment on the above answer:
According to this https://msdn.microsoft.com/en-us/library/dd144904(v=vs.85).aspx if bitmap is not created by CreateDIBSection bitmap.bmBit pointer will be null.
So, is there anyway to access the bits of the bitmap (RGB values), without copying them?
If it helps, the device on which the bitmap is dependent is always a screen, so I guess it's always a 24bit or 32bit bitmap.
In this article, regarding how to save an HBITMAP to a file, there's this section:
// Allocate memory for the BITMAPINFO structure. (This structure
// contains a BITMAPINFOHEADER structure and an array of RGBQUAD
// data structures.)
if (cClrBits < 24) pbmi = (PBITMAPINFO)LocalAlloc(LPTR, sizeof(BITMAPINFOHEADER) + sizeof(RGBQUAD) * (1 << cClrBits));
// There is no RGBQUAD array for these formats: 24-bit-per-pixel or 32-bit-per-pixel
else pbmi = (PBITMAPINFO)LocalAlloc(LPTR, sizeof(BITMAPINFOHEADER));
So, I don't really understand what's happening here, and which is right?
Is it right to create two bitmaps for one DC?
My situation is I have
dc = CreateCompatibleDC(NULL);
bitmap = CreateDIBSection(dc, pbmi1, DIB_RGB_COLORS, (void **) &bmpBits, NULL, 0);
extra_bitmap = CreateDIBSection(dc, pbmi2, DIB_RGB_COLORS, (void **) &extraBmpBits, NULL, 0);
The difference is: bitmap is width*height , extra_bitmap is width/2*height*2
I need to show both images on the screen according to requirements from users.
I tried these code above, it can run (based only on my limited test). But I am not sure if it is the right way to do it. Any suggestions and/or comments are welcome. Thanks.
By the way, if I can have two bitmaps associated with one DC, what is the right way to release them? Can I just call DeleteDC(dc)? Thanks.
Bitmaps are only associated with a DC when they're selected into them via SelectObject. The DC you pass to the CreateDIBSection is only used to define the layout/palette of the newly created bitmap, and only in some circumstances, but it doesn't irrevocably associate that bitmap with the DC or anything like that.
To draw a bitmap to a window, it needs to be selected into a DC (the "source DC"), and this is where you can only have one at a time.
For example:
HDC hdcTarget = GetDC(hWnd); // or hdcTarget = BeginPaint(...)
HDC hdcSource = CreateCompatibleDC(hdcTarget); // create a "source DC")
HGDIOBJ hOldBmp = SelectObject(dc, bitmap); // select bitmap 1 in
BitBlt(hdcTarget, x, y, w, h, hdcSource, 0, 0, SRCCOPY); // draw bitmap 1
SelectObject(hdcSource, extra_bitmap); // select bitmap 2 in
BitBlt(hdcTarget, x1, y1, w1, h1, hdcSource, 0, 0, SRCCOPY); // draw bitmap 2
SelectObject(hdcSource, hOldBmp); // restore previous bitmap
DeleteDC(hdcSource); // no longer needed
ReleaseDC(hWnd, hdcTarget); // or EndPaint(...)
After search around for a few hours, I get
http://www.ucancode.net/Visual_C_MFC_Samples/CreateDIBSection-BITMAPINFOHEADER-CreateCompatibleDC-BITMAPINFO.htm
In this example:
// Create HDCs to hold our surfaces.
hdcSrc1 = CreateCompatibleDC(hdcDest);
if (!hdcSrc1) goto HANDLEERROR;
hdcSrc2 = CreateCompatibleDC(hdcDest);
if (!hdcSrc2) goto HANDLEERROR;
For delete:
DeleteDC(hdcSrc1);
DeleteDC(hdcSrc2);
DeleteObject(hbmSrc1); // hbmSrc1 and hbmSrc2 are created using hbm = CreateDIBSection(hDC, &bmi, DIB_RGB_COLORS,(void **)&pBits, NULL, 0);
DeleteObject(hbmSrc2);
Hope this can be helpful for others. thanks.
You may create as many in-memory bitmaps as you like.
You may create as many device-dependent DCs as you like.
A bitmap may be selected into only one DC at a time. If you try to select the same bitmap into multiple DCs, then you will get unpredictable behavior.
A DC will never have more that one bitmap selected into it at a time, because the act of selecting one in selects the previous one out.
Don't delete a memory DC while your bitmap is still selected into it. You should select the original bitmap (which is a monochrome 1x1 bitmap) into the DC to get your bitmap out before you delete the DC.
Don't delete an in-memory bitmap while it's still selected into a DC.
I need to create the HBITMAP.
Here lies the problem. I have the bmp file's content in memory.
I know how to create a HBITMAP if the bitmap was as a resource.
But since it is in memory, I don't know how to do it.
I do it like this (if in resource) : Link
hDC = BeginPaint(hWnd, &Ps);
// Load the bitmap from the resource
bmpExercising = LoadBitmap(hInst, MAKEINTRESOURCE(IDB_EXERCISING));
// Create a memory device compatible with the above DC variable
MemDCExercising = CreateCompatibleDC(hDC);
// Select the new bitmap
SelectObject(MemDCExercising, bmpExercising);
// Copy the bits from the memory DC into the current dc
BitBlt(hDC, 10, 10, 450, 400, MemDCExercising, 0, 0, SRCCOPY);
// Restore the old bitmap
DeleteDC(MemDCExercising);
DeleteObject(bmpExercising);
EndPaint(hWnd, &Ps);
Please guide me on how to do it if it were a memory resource.
Somehow change char img[10000] into a resource ?
Here, img is the memory contraining the contents of the bitmap.
First, let's remove a little innocent gotcha:
hDC = BeginPaint(hWnd, &Ps);
// Load the bitmap from the resource
bmpExercising = LoadBitmap(hInst, MAKEINTRESOURCE(IDB_EXERCISING));
// Create a memory device compatible with the above DC variable
MemDCExercising = CreateCompatibleDC(hDC);
// Select the new bitmap
HOBJECT oldbmp = SelectObject(MemDCExercising, bmpExercising); //<<<<save it for later ...
// Copy the bits from the memory DC into the current dc
BitBlt(hDC, 10, 10, 450, 400, MemDCExercising, 0, 0, SRCCOPY);
// Restore the old bitmap
SelectObject(MemDCExercising, oldbmp); //<<<... DeleteDC will leak memory if it holds a resource
DeleteDC(MemDCExercising);
DeleteObject(bmpExercising);
EndPaint(hWnd, &Ps);
Now, an HBITMAP is (conceptually speaking) a pointer to an internal structure that holds a "pointer" (in fact is much more a stream) towards the GDI memory space you cannot access.
A "memory bitmap" is not represented into your program as a memory buffer that belogs to your program, but as ... an HBITMAP obtained with CreateCompatibleBitmap, where the HDC parameter id the DC the bitmap has to be compatible with. (usially a screen, window or painting DC).
You can cerate an initialized bitmap passing a buffer containing initial data, or get the data a bitmap holds with CreateBitmap or GetBitmapBits.
In any case, these are your local copy of the bitmap data, not the "live bitmap" GDI draws in.
Note also that the internal structure of those data depends on the format (how many bits per pixel on how many planes and with or without a palette) the bitmap needs to have, and that, to avoid performance penalty in the Blit process, it must coincide with the format used by your screen setup.
This does not necessarily have to be the same as the one the bitmap has when saved into a "bmp" file.
MFC doc/view architecture, GDI drawing/printing. I have a DIB backbuffer I need to display and print.
After the long and painful road I came to the conclusion than I need to use DIB created with CreateDIBSection (rather than DDB created with CreateCompatibleBitmap), and I have to blit it onto printer dc with StretchDIBits (rather than StretchBlt).
But i'm not able to get this thing to work.
Here is what I do:
In my initialization routine, I setup the backbuffer, like this:
// Prepare device context:
CClientDC aDC(this);
OnPrepareDC(&aDC);
// Setup proper backbuffer:
_pMemDc = new CDC;
_pMemDc->CreateCompatibleDC(&aDC);
memset(&_bitmapInfo, 0, sizeof(BITMAPINFO));
_bitmapInfo.bmiHeader.biSize = sizeof(BITMAPINFOHEADER);
_bitmapInfo.bmiHeader.biWidth = _sizeBackBuffer.cx;
_bitmapInfo.bmiHeader.biHeight = -_sizeBackBuffer.cy; // top-down
_bitmapInfo.bmiHeader.biPlanes = 1;
_bitmapInfo.bmiHeader.biBitCount = 24; // Maybe 32?
_bitmapInfo.bmiHeader.biCompression = BI_RGB;
HANDLE hMemBitmap = CreateDIBSection(aDC.GetSafeHdc(), &_bitmapInfo, DIB_RGB_COLORS, (void**)&_pBitmapRawBits, 0, 0);
_hOldSelBitmap = (HBITMAP)_pMemDc->SelectObject(hMemBitmap);
Variables with underscores are (private) member variables, declared like this:
CDC *_pMemDc; // Backbuffer memory dc
HBITMAP _hOldSelBitmap;
BITMAPINFO _bitmapInfo; // Backbuffer DIB (header-only)
unsigned char *_pBitmapRawBits; // Pointer to pixel data of DIB
SIZE _sizeBackBuffer; // Size of backbuffer, i.e. size of that DIB
Now below is what I do in my OnDraw override:
Firstly I get the area to be drawn like this (simplified code):
CRect rectClipBoxPlayground;
if (pDC->IsPrinting())
{
rectClipBoxPlayground = _printingParams.pPrintInfo->m_rectDraw;
}
else
{
pDC->GetClipBox(&rectClipBoxPlayground);
}
Then I calculate the corresponding rect coordinates in my backbuffer, which is usually (much) larger than the DC. Details of this calculation are irrelevant here, I just say that
CRect rectClipBoxBackBuffer;
represents the corresponding rect of backbuffer (in pixel coordinates of backbuffer).
Then I draw onto my backbuffer, using the _pMemDc memory dc.
And finally comes the part where I have troubles: blitting my DIB onto target dc (screen or printer). Here is what I do:
// Copy back buffer to screen/printer dc:
pDC->SetStretchBltMode(HALFTONE);
SetBrushOrgEx(pDC->GetSafeHdc(), 0, 0, 0);
// BELOW COMMENTED CODE OF StretchBlt WORKS(!) INSTEAD OF StretchDIBits.
//
//BOOL bSuccess = pDC->StretchBlt(rectClipBoxPlayground.left, rectClipBoxPlayground.top,
// rectClipBoxPlayground.Width(), rectClipBoxPlayground.Height(),
// _pMemDc, rectClipBoxBackBuffer.left, rectClipBoxBackBuffer.top,
// rectClipBoxBackBuffer.Width(), rectClipBoxBackBuffer.Height(), SRCCOPY);
HBITMAP hMemBitmap = (HBITMAP)_pMemDc->SelectObject(_hOldSelBitmap);
DWORD dwLines = StretchDIBits(pDC->GetSafeHdc(),
rectClipBoxPlayground.left, rectClipBoxPlayground.top, rectClipBoxPlayground.Width(), rectClipBoxPlayground.Height(),
rectClipBoxBackBuffer.left, rectClipBoxBackBuffer.top, rectClipBoxBackBuffer.Width(), rectClipBoxBackBuffer.Height(),
_pBitmapRawBits, &_bitmapInfo, DIB_RGB_COLORS, SRCCOPY);
_pMemDc->SelectObject(hMemBitmap);
The problem is, the commented-out code of StretchBlt works flawlessly (except printing on some printers), but I can't use it because some printers have troubles with it. So I have to use StretchDIBits. Note that I firstly unselect the DIB from its memory dc temporarily, so that it is not associated with any dc. Then I use StretchDIBits but things just don't work! Output is messed up, like I give incorrect coordinates, areas are drawn off-set from where they should be drawn, and sometimes totally black. So I must be missing something (maybe something very trivial). Help! I tried changing signs of rectClipBoxBackBuffer.top and bitmapInfo.bmiHeader.biHeight, the results change, but nothing works like it should.
What am I doing wrong??
Dont't select the DIB into/out DC if you are going to use StretchDIBits.
DC can only contain DDB bitmap, if you supply a DIB, DC will convert it.
Microsoft's documentation about StretchDIBits is totally wrong. I found out that direction of Y axis has to be changed. Now following code works:
// Copy back buffer to screen dc:
pDC->SetStretchBltMode(HALFTONE);
SetBrushOrgEx(pDC->GetSafeHdc(), 0, 0, 0);
HBITMAP hMemBitmap = (HBITMAP)_pMemDc->SelectObject(_hOldSelBitmap);
DWORD dwLines = StretchDIBits(pDC->GetSafeHdc(),
rectClipBoxPlayground.left, rectClipBoxPlayground.top, rectClipBoxPlayground.Width(), rectClipBoxPlayground.Height(),
rectClipBoxBackBuffer.left, _sizeBackBuffer.cy - rectClipBoxBackBuffer.top - rectClipBoxBackBuffer.Height(), rectClipBoxBackBuffer.Width(), rectClipBoxBackBuffer.Height(),
_pBitmapRawBits, &_bitmapInfo, DIB_RGB_COLORS, SRCCOPY);
_pMemDc->SelectObject(hMemBitmap);
P.S.: Now it works for screen drawing and print preview, but fails for actual printing, which was my original problem discussed here: How to print DIB backbuffer on printer - GDI, MFC
How can I create a device context compatible bitmap and then associating the obtained handle to a BITMAP struct?
If I write:
...
HBITMAP hbitmap = CreateCompatibleBitmap(hdc, width, height); // these three arguments are initialized somewhere else
hbitmap = CreateBitmapIndirect(bitmap); // argument already initialized and properly filled
...
A HBITMAP handle compatible with hdc is created, and then a new HBITMAP (filled with bitmap data) is initialized, though without keeping its compatibility. Is there a function which allows not to create a HBITMAP from a BITMAP, but rather fills an initialized HBITMAP with an already existing BITMAP source?
CopyImage function
Creates a new image (icon, cursor, or bitmap) and copies the attributes of the specified image to the new one. If necessary, the function stretches the bits to fit the desired size of the new image.
HANDLE WINAPI CopyImage(
HANDLE hImage,
UINT uType,
int cxDesired,
int cyDesired,
UINT fuFlags
);
hImage A handle to the image to be copied.
uType The type of image to be copied. This parameter can be one of the following values.
IMAGE_BITMAP 0 Copies a bitmap.
IMAGE_ICON 1 Copies an icon.
IMAGE_CURSOR 2 Copies a cursor.
cxDesired The desired width, in pixels, of the image. If this is zero, then the returned image will have the same width as the original hImage.
cyDesired The desired height, in pixels, of the image. If this is zero, then the returned image will have the same height as the original hImage.
fuFlags
CreateBitmapIndirect takes BITMAP on its input. And you can get it via GetObject from HBITMAP:
BITMAP Bitmap;
INT nResult = GetObject((HGDIOBJ) hBitmap, sizeof Bitmap, &Bitmap);
CreateBitmapIndirect will be able create a bitmap from this struct. Or you can use CreateCompatibleBitmap to create compatible bitmap providing width/height from obtained Bitmap.