ICDL Image Editing (ICDL) common mistakes
ICDL Image Editing · Understand the traps and how to avoid them.
Changing the image's resolution value (e.g. from 72 to 300 ppi) in the Image Size dialog and expecting the picture to gain detail or the file size to grow, while the pixel dimensions stay the same.
The ppi field looks like a quality slider, and the dialog's resample checkbox is easy to leave in its default state, so candidates never see that they are only relabelling the print size.
What to do instead
Before answering, decide whether the task changes pixel dimensions (resample on: 1200x800 to 2400x1600) or only the print size (resample off: same pixels, smaller printed area). Read the pixel figures, not the ppi figure, to judge quality and file size.
Estimating file size or choosing a colour mode without working from bit depth - for example assuming a grayscale copy of a photo will be one third the size of the RGB original, or that a 24-bit and an 8-bit image of the same pixel count are the same size.
Candidates memorise 'grayscale is smaller' as a slogan instead of the relationship pixels x bits per pixel, and the exam phrases questions numerically rather than visually.
What to do instead
Use the formula width x height x bit depth / 8 for uncompressed size, and note that grayscale uses 8 bits per pixel against 24 for RGB, so the ratio is roughly 1:3 but only for uncompressed files. Check the format's compression before predicting the on-disk size.
Scanning a photograph at a low setting such as 72 or 100 ppi, then enlarging it later in the editor to reach print size.
The scan dialog defaults to screen resolution and candidates equate 'looks fine on screen' with 'good enough', not realising the detail was never captured.
What to do instead
Scan at target print ppi multiplied by any enlargement factor (e.g. 300 ppi for a 1:1 print, 600 ppi if the scan will be printed at double size), and check the resulting pixel dimensions against the required output size before leaving the scanner dialog.
Acquiring an image (scan, camera import, screenshot, stock download) and dropping it straight into a layout without checking its pixel dimensions against the intended print size or the document's colour mode.
Acquired files open at whatever size they happen to be, and the candidate judges suitability by how the image looks on screen rather than by effective resolution and mode.
What to do instead
Before use, compute effective ppi as pixel width divided by the required printed width in inches and confirm it reaches the target (commonly 300 ppi for print, 72-96 ppi for screen). Convert the mode to match the destination document at the point of import, not after editing.
Stretching or squashing a photograph with the resize command, constrain-proportions switched off, to make it fit a required frame - producing a visibly distorted subject.
Resize and crop both 'make the picture the right shape' in the candidate's mind, and turning off the link icon feels like the quick way to hit exact dimensions.
What to do instead
Use crop to change the shape of an image without distortion, and reserve resize (with proportions locked) for uniform scaling. If a specific aspect ratio is required, crop to that ratio first, then scale.
Applying a brightness, colour or filter adjustment while a selection is still active, so only part of the image changes, and then not noticing because the marching ants were hidden or the change was subtle.
Selections persist invisibly once the marquee is hidden or the tool is switched, and adjustment dialogs do not warn that they will be masked.
What to do instead
Get into the habit of deselecting (Select > Deselect) before any image-wide adjustment, and glance at the document border or the selection indicator in the options bar to confirm nothing is active.
Building a composite by pasting or placing images of mismatched pixel dimensions and colour modes directly into the working document, then wondering why one element is the wrong physical size or has shifted colour.
Paste carries the source's own resolution and mode, and the candidate assumes the destination document will normalise them.
What to do instead
Before combining, resample each element to the composite's resolution with proportions locked, convert its mode to match the document, and place it on its own named layer so it can be repositioned and re-scaled non-destructively.
Flattening a layered composition or saving it straight out as JPEG, so the layers and any transparency are lost and there is no editable master file left.
The Save dialog offers JPEG by default and the candidate treats the final export as the working file.
What to do instead
Save the working file first in a layered format the editor supports (e.g. PSD or TIFF), and export a flattened copy separately for delivery. Only flatten the master when the task explicitly says the layered version is no longer needed.
Choosing the output format by habit rather than by content - saving a photographic image as GIF, a logo with transparency as JPEG, or line art as a low-quality JPEG - and then accepting the artefacts or lost transparency.
Candidates associate 'small file' with 'good' and do not connect format features (colour limit, transparency support, lossy or lossless) to what the image actually contains.
What to do instead
Match format to content: JPEG for continuous-tone photographs with no transparency, PNG for images needing transparency or crisp edges and text, GIF only for simple images with few flat colours. Save a JPEG at maximum quality once rather than re-saving it repeatedly.
Exporting a print-ready image at screen settings - 72 ppi and RGB - or exporting a screen image at 300 ppi CMYK and producing an unnecessarily huge file.
The export dialog remembers the last settings and candidates do not review resolution and colour mode against the stated destination.
What to do instead
Before exporting, state the destination: for print set the physical size with 300 ppi and convert to CMYK if the printer requires it; for screen set pixel dimensions directly, keep RGB, and check the resulting file size against any limit given in the task.