Polly wrote:Can anyone please help with my query about JPEG compression?
I was leaving it to the experts, but if I write some nonsense here you'll get an answer as they all queue up to correct me!!
A raw image has each pixel separately coded for and is huge. It represents the light that reaches the sensor. A compressed image condenses the information in many ways, including effecively 'describing' the image rather than representing it. For example, instead of recording 30 consecutive pixels with the same colour it could indicate simply that 30 identical pixels are coming up. But to describe wholly accurately is still heavy, so approximations and shortcuts are made, with the result that information is lost. There are algorithms that minimise the effect of loss of information by recognising lines, borders &c. and slightly modifying the result to reflect them, accentuating (sharpening) borders for example. The greater the compression, the more 'after work' is needed to keep the picture looking good.
All this introduces artefacts of one sort and another. A common artefact is a kind of heat haze over edges. I imagine this is preferable to the blurred edge that would result from compression without the post-compression adjustments.
Different compression algorithms will result in different effects too, I imagine.
That should get someone to explain it to us properly, Polly!!
Guy