How does Python PIL and PILLOW compress pictures?

Question

I wanted to write a program with Python 3.3 in which a frame is added to a picture. I use the PIL package from Python to do that. But the pictures I get are less than a third as big as the original and they loose quite some focus/sharpness/gloss. Where does Python compress/change the picture and how (if possible) can I subdue it?

I could think of some passages, which might be the one compressing.

1. I load the image with PIL:

   img = Image.open(element)

2. I create a new picture, which is black:

   newImg = Image.new("RGB",(imgWidthNew,imgHightNew),(0,0,0,0))

3. I create a pixel map out of my picture and change some:

   pixels = newImg.load()

4. I paste my picture into the middle of the black one:

   newImg.paste (Image.open(element), (halfFrameWidth, halfFrameHight, imgWidth+halfFrameWidth, imgHight+halfFrameHight))

5. The new image is beeing saved:

   newImg.save(path,dpi=[300,300])

I load mostly .jpg pictures, but due to some problems I extract .bmp files.

 path = path[:path.rfind(".")] + fileEnding

But this shouldnt be the problem, because I have the same issues without changing the data type. My operating system is Windows7 64BIT and I use Pillow 2.3.0 with Python 3.3.

Thank you for any help.

Solution

When saving as JPEG, you can set the quality parameter to control the amount of compression:

newImg.save(path, 'JPEG', dpi=[300,300], quality=90)

Or, save in a lossless format, such as PNG:

newImg.save(path, 'PNG', dpi=[300,300])