Shoutbox

best way i can think of is to use blur, sharpen and then hand tools to touch up... the computer cannot make educated guesses about the whole picture, only follows certain rules. you however, can work out what the picture is meant to look like and improve the image.

doing it like in the films is impossible... the only way i can think of that being realistic is if the cctv camera is capable of high res images but will be set at low res to monitor movement... when a high res image is needed it has to change to take a better image. just like in a digital camera with different quality settings.

however making something out of nothing is impossible! or else i would be rich

quote:

---

Originally posted by gif83
however making something out of nothing is impossible!

---

This is true, but you haven't got nothing. If you have a pixel, you generally have the average colour of that portion of screen.  By comparing it to the ones around it, or by making a mock comparison by smudging them into one another and comparing it to the unsmudged version (the basis of unsharp masking) you can at least retrieve lost shape from poor focus/definition.  Generally what loses out is the precise colour of these shapes... if you apply a nonsharp mask with extreme values, you will get as perfect a shape as you could ask for, but crazy colour contrast.

It really does enhance the raw capabilities of the camera.  Astronomers use it all the time to get pictures of stars and galaxies which the hubble telescope just cannot see in focus, and the same can be achieved when improving the quality of cctv footage

what i'm saying is that it's always going to be a calculated guess...and the guessing should be done not soley by yourself or soley by the computer... a partnership is the best option... which is also done by astronomers (who hire artists)