Creating a QImage from QBytearray

Question

I'm trying to construct QImages from data saved as 16-bit integers in a binary file. I'm able to load the data in fine, but my program fails when I use the QImage::loadFromData(QBytearray ba) function (returning false) as follows:

QBytearray frame;
QImage pic = QImage(256, 256, QImage::Format_RGB888);

for(int i = 0; i < height; i++) {
    for(int j = 0; j < width; j++) {
        // Access value of pixel at each location
        datum = store[i][j];

        for(int c = 0; c < 3; c++) {
            // Calculate colour at given pixel
            col = (255.0f * ((float)datum - (float)min) / ((float)(max - min)));
            // Assign colour value to the pixel
            frame[c+3*j+3*i*width] = ((unsigned char)col);
        }
    }
}

pic.loadFromData(frame);

I repurposed this from Java code I had previously written which worked perfectly as intended (from the exact same data):

BufferedImage image = = new BufferedImage(256, 256, BufferedImage.TYPE_3BYTE_BGR);

byte[] data = image.getRaster().getDataBuffer();

for (j=0; j<height; j++) {
    for (i=0; i<width; i++) {
        //Find value of the pixels at the location
        datum=data[j][i];
        for (c=0; c<3; c++) {
            //Calculate the colour at the given pixel
            col=(255.0f*((float)datum-(float)min)/((float)(max-min)));
            //Assign the colour value to the pixel
            data[c+3*i+3*j*width] = (byte)col;
        } 
    }
}

Can anybody help me to see where I'm getting this wrong? I've been stumped for days and am all out of ideas.

Have you tried making store a hard-coded set of predetermined values? If so, what was the result? If not, could you do that and give us the result? What is frame when you finish building it? — MirroredFate
@MirroredFate I can't really do that considering that the image is 256x256, so hard coding the values for that image is a bit time consuming. And what do you mean exactly by what frame is when I've finished building it? Size? — Strongo
One thing that has baffled me though is that dumping the Bytearray from a separate image loaded in directly from file (which works fine) gives me a QByteArray 2562 elements long - whereas the one I'm constructing is 196609 elements long. I feel this (huge) size discrepancy is at least part of the problem, but I don't know what the difference in the data formatting might be between the two. I mean, the 196609 byte array worked fine in the Java implementation, so I don't really know why the C++ bytearray for a same-sized image would be so much smaller. — Strongo
Also, using the dumped QByteArray in the QImage::loadFromData(QByteArray ba) function works exactly as advertised. It seems as if the size discrepancy is the major issue, but even if I drop the colour setting (c for loop) I still have an array 65536 elements long. — Strongo
This is mainly guesses You have a 2d array store, and a QByteArray frame. You are moving unsigned chars (1 byte) from the store to frame, but you need 3 bytes total for your RGB value. This means frame is essentially an array of bytes equal to 256*256*3, or 196609 bytes. I don't see where you are putting in header info to tell which kind of image it is, or any sort of meta characters to give the size... — MirroredFate

MirroredFate MirroredFate · Accepted Answer · 2014-03-13T00:23:25

Ok, assuming that you are in fact trying to set RGB values for individual pixels, after reading the QImage details, I see you can do it using the following:

value = qRgb(189, 149, 39); // 0xffbd9527
image.setPixel(1, 1, value);

So, something like:

QImage pic = QImage(256, 256, QImage::Format_RGB888);
QRgb value;
int r,b,g;

for(int i = 0; i < height; i++) {
    for(int j = 0; j < width; j++) {
        // Access value of pixel at each location
        datum = store[i][j];
        //I get really confused what is going on here... you don't seem to be actually using `c` for the calculation?
        for(int c = 0; c < 3; c++) { //let's just pretend you set the ints r,b,g in here somewhere
            // Calculate colour at given pixel
            col = (255.0f * ((float)datum - (float)min) / ((float)(max - min)));
        }
        // Assign colour value to the pixel
        value = qRgb(r, g, b);
        pic.setPixel(i, j, value);
    }
}

Creating a QImage from QBytearray

1 Answers