Photoshop – The Crop Tool

This is one of the first operations to do when preparing an image.  After all, you don’t want to work on parts that may later be discarded!  However, before cropping, it is sensible to correct the image both for convergence (Free Transform, Ctrl T) and for sloping horizons/verticals (the Measure Tool) if necessary.

The first thing to think about is what print shape and size you want to finish with.  If the print is to be specially mounted, say for a competition, then the shape and size will be determined by artistic merit, although the specification of your printer and the paper available will influence this.

On the other hand, if you are going to use a proprietary mount, then the image needs to be printed at a very precise size and aspect ratio, say 8” by 6” or 10” by 8”.

However, before we look at using the tool, let’s look at,

The Theory 

An image taken from a digital camera doesn’t have a pixel resolution, just a matrix of pixels, the number depending on the specification of the detector in the camera.  My Olympus produces an image 2048 pixels by 1536 pixels, giving a total of 3.3 million.  Setting the resolution, pixels per inch, only becomes necessary when you want to view the image either on a monitor or as a print.  For viewing on a monitor, a resolution of 72 ppi is used, as this is approximately the capability of a screen and for printing, the standard is generally accepted as 300 ppi as your eye is unable to see pixels of this size.

However, it isn’t too difficult to calculate that at this resolution, 2048 x 1536 pixels will only give an image of  6.8” by 5.1” (divide the pixels by the resolution).  If you are willing to accept that 200 ppi is satisfactory, then you can produce a print size of  10.2” by 7.7”, ie near enough a 10 by 8.  In practice, I find that 200 pixels per inch produces a very sharp print and in some test prints prepared by Graham Reeves last season, most of us were unable to find fault with prints done at 150 ppi.

This is in part due to the way that printers reproduce the pixels. On the right, Screenshot 1, you can see a greatly enlarged portion of an image showing the individual pixels.  Each pixel can be one of millions of different colours or shades and the printer has to be able to handle these using only yellow, cyan, magenta and black ink.  It also effectively has white if this is your paper colour.  Some printers have light cyan and light magenta to avoid showing too much paper in pale tones.

It perhaps might be a neat trick if the printer could mix up the correct blend of colours and then squirt it on to the paper, but in practice it has separate jets for each colour, and it reproduces each pixel by laying down a patch of tiny dots of the necessary colour(s) so that our eyes see only the blended result.  You can see a similar effect if you look closely at a newspaper picture which consists of black dots on a white background where the size and closeness of the dots is used to produce various shades of grey.

Here’s the maths again.  If your printer has a resolution capability of 2400 dots per inch, then to reproduce pixels at 300 ppi, it can jet out a matrix of 8 x 8 = 64 dots for each pixel. To reproduce pixels at 200 ppi then the number of possible dots of ink increases to 12 x 12 = 144 for each pixel giving a rather better representation.  For the same image size, each pixel at 200 ppi is bigger than those at 300 ppi and if you keep on increasing the actual pixel size, then eventually you will be able to see individual pixels.  Usually the first sign of this is jagged edges in the image.

You need to understand this theory because Photoshop is able to size your images ready for printing, and if you insist on a specific resolution it will give you what you ask for either by dropping some pixels out or by putting additional pixels into the image.  So if I insist on sizing my Olympus prints to 12” x 10” at 300 ppi, then the 2048 pixels across the image will have to be increased to 3600 (12x300) pixels, a massive increase, and to achieve this, Photoshop has to insert pixels at regular intervals across the image.  In order to decide what colour each should be, it looks at nearby pixels and calculates a match, but you can imagine how difficult this can be.  It has to do a similar operation on each row of pixels.

So generally, it is preferable to leave the resolution setting on the options bar empty, and not to set an image size that is going to give a resolution of much less than 150 to 200 ppi.  As far as the printer resolution is concerned, then use the best photo setting, probably something like 2400 dots per inch or better.

If you are cropping heavily, then you may only be printing perhaps 50% of the available pixels, and this will give severe problems if you want to print a large image.  You may find that the resolution is below 100 ppi and you then have the options of cropping less heavily, settling for a smaller image or fixing the resolution and allowing Photoshop to add the additional pixels.  Certainly your choice.

Using the Crop Tool

Whether you are performing a "free crop" or a "fixed size crop", it’s easier to leave the resolution setting to be done on the Image Size panel.

Load your image.

Select the Crop Tool ( key C)

Next clear all of the settings on the Crop options bar by clicking the Clear button, see Screenshot 2

Check that your Image Size panel (click Image on menu bar and then Image Size) has the box for Resample Image unticked.  Then click [OK].   See Screenshot 3

Draw The Crop Marquee

If you are doing a free crop, leave all of the dimension boxes empty.  It’s easier to set these along with the resolution, on the Image Size panel.

If you want a precise size, set the Width and Height making sure the units are correct.  You can do this by typing, after the dimension figure, in for inches or cm for centimetres.

On your image, hold down the left mouse button and drag the crop marquee from top left to bottom right. For a free crop, you can draw any shape or size you like – for a fixed size crop, the ratio of width to height remains fixed and you can only set the size.  You can release the button when the crop looks approximately correct as you can make further adjustments by dragging the handles on the sides or corners.  You can also reposition the crop by placing the cursor inside the cropped area, holding down the left mouse button and dragging to a new position.

When you are happy with the crop, either double-click inside the cropped area, or click the accept icon at the right of the options bar.  Alternatively you can click the Abort icon (also top right) and start again. 

Following cropping, look at your Image Size panel again. 

 If you have done a free crop, it will show how many pixels you have selected and what size the print would be at 72 ppi.  Screenshot 4

At this point, set either the width or the height to whatever you like.  The other dimension will change as will the resolution.  In Screenshot 5, I have set the width to 25.4 cm (10”).  The resolution has changed to 185 ppi and I would try a print at that.

 If you have cropped to fixed dimensions, the panel will be similar to Screenshot 5 and you will see precisely how many pixels you have selected, but the width and height will be the dimensions you set in the crop options bar and the resolution will have changed to match

 If you want to change the resolution then you have two options.

  1. Without resampling the image, when you enter a new resolution, the pixels will remain unchanged and the image size will change.
  2. If you click both Resample Image and Constrain Proportions, and then change the resolution, the pixel numbers will change and the size remain the same.

In Screenshot 6, both boxes are ticked, I have increased the resolution to 200 ppi, and Photoshop has added about 150 extra pixels across the width – not too many so this print would probably be good as well.  However, I would not try 300 ppi which would need the insertion of a further 1000 pixels per line, and I would consider this to be excessive.

Note also, that Bicubic is recommended as the best calculation method for pixel additions.