Printing/resolution/file size

[martcol]

Can I have some help with understanding printing size/resolution. I understand pixels per inch and dots per inch. What I don't understand is, if I have a file that's say, 8MB how (should) I change it's size to print it at 300 dpi? I send some prints off to an online service (6*4s for £0.15 each) and save them in JPG. I do this because of uploading but I don’t really understand what I am doing. I just use the quality setting in PhotoShop until it tells me at the bottom of that Save As screen that it is less than 1MB. I can't work out where all those lovely pixels go if I change the size of an image for printing? But equally, I can't believe that I need to send huge files to my own ink jet printer for an A4 page.

When I think of digital images I imagine writing on a balloon and blowing it up or visa versa. What you write always looks stronger when the balloon is deflated but then, lacks detail. But if you blow the balloon up too much the image is weak. I'm rambling now and this hasn't led to a question but it might help clever people understand how I think.

Thanks

Martin

[john_houghton]

Compressing an image file using jpeg format has no effect whatsoever on the number of pixels in the image, but the file becomes smaller on disc with the colours of the pixels being less accurately described. Thus, a group of pixels with similar, but slightly different, rgb values may end up with the same rgb value in the compressed file. This can result in a blocky appearance.

You can change the print size just by altering the ppi resolution setting of the image using Image->Image Size (ensuring that the resample option is not checked). You can change the basic image file size (i.e. the number of pixels) by using Image->Image Size and checking the resample box. Just alter the pixel dimensions (also check the "Constrain image proportions" box).

You need to end up with an image resolution of between 200 and 300ppi. In conjunction with the print dimensions, that will determine the basic file size. A 6" x 4" print at 300ppi will be 1800 x 1200 pixels. This is 2.16MP, with a file size of 2.16 x 3 MB (3 bytes per pixel). The file size on disc can be reduced by applying jpeg compression, but not too much or else the image quality will suffer. A compression factor in the range 10 - 20 is usually ok.

John

[martcol]

Thanks John

So, compressing a file for email or posting to the WWW is different to preparing a file for printing to my inkjet? But in some way, preparing a file to print and changing physical size and ppi will reduce it's overal size. So, in that case, where do those lovely pixels go?

If I take two different shots with my camera, they will be different sizes but use the same number of MegaP's. What makes one bigger than the other?

Martin

[john_houghton]

The reason why pictures of different scenes get filed on the memory card with different file sizes is because the effect of compression varies according to to image content. I don't know exactly how jpeg compression works, but a simple example of compression will serve to give some idea of the sort of thing that goes on.

Think of the compressed jpeg file as containing a description of the image rather than the image itself. Suppose you take a picture of a sheet of red paper, i.e. all pixels = same shade of red. A description of this picture might be "Size=1800x1200. All pixels =red". This is very short and to the point, but it enables the full image to be reconstructed. If the image contained a slightly more complicated pattern, with the top half light blue and the bottom half green, the description could be "Size=1800x1200, rows 1-600=light blue, rows 601-1200=green". This description is now twice as long as a result of the content of the image becoming a bit more complex. If the image contains 4 rows of 4 rectangles, each of a different colour, the description will necessarily become correspondingly longer.

So it is generally true that the effect of compression varies with the content of the image: an image with a lot of detail will not compress to as small a file as a bland image with very little detail.

Some forms of compression set out to enable the original image to be reconstructed without any loss of quality. This is called lossless compression, and with sufficiently complex images, the compressed file can be actually bigger than the original! OTOH, jpeg compression is lossy: to enable reasonable degrees of compression to be reliably obtained, the image structure will be simplified by approximating pixel values. The more the data is compressed, the more approximate the reconstructed image will become.

Compression does not throw away pixels, only pixel rgb values, but this can happen when you resize an image by resampling. Actually, you then thow away ALL the pixels and substitute a different set, with the new pixel rgb values being mathematically calculated as a function of the original pixel values in the neighbourhood of the corresponding point in the original image.

John