Why is it?

Soldato
Joined
23 Nov 2004
Posts
8,024
Location
The Place To Be
When you take a photo off your digital camera, it is usually in the region of 2 or 3 MB. However, if you open this in paint, and then simply save over the top of it, it is reduced to say 700KB without any noticeable loss in quality?
 
Its a scam, the Camera companies have struck a deal with the memory stick people so that all images are inflated thus not giving you as much room on the memory stick. It results in you buying more memory sticks, thus getting your to part with some more cash :D

cough! Rubbish!!! cough!
 
There's a huge difference in quality when saved in paint, it discards a lot of the information in the image and makes it look crap!

If I save an image at max quality in Photoshop, straight from my camera it normally increases by a few hundred Kb! :p
 
I'd guess the camera has to save the images pretty quickly and has a lot less computing power than your PC - so just doesn't compress everything nearly as much.
 
There's a huge difference in quality when saved in paint, it discards a lot of the information in the image and makes it look crap!

If I save an image at max quality in Photoshop, straight from my camera it normally increases by a few hundred Kb! :p

really? I can't notice the difference between the two versions.
 
Send me the original and I'll try and get it to look better, ctrl+alt+shift+s = save for web which is normally the best way to get high quality with low file sizes.
 
Send me the original and I'll try and get it to look better, ctrl+alt+shift+s = save for web which is normally the best way to get high quality with low file sizes.

Ah I didn't know it could do that. This updated version looks much better.

cheers!
 
Yeah when you save an image as JPEG file on your computer the picture is compressed (via wavelets in JPEG's case) so that for a given resolution level, the picture is saved using the minimum number of coefficients.

After JPEG compression the actual amount of data you need to keep the image at a resonable level is much less than the raw pixel by pixel version.

As an example take this webpage, the majourity of it can be expressed as blocks of colour (due it's nature). Vastly saving on the amount of information that needs to be stored.

The system (or methodology) involved must obey a critical sampling ratio (in order to reconstruct the image you need to have a "look" at the imagine a certain number of times). I'm not sure of the exact setup in JPEG, but I think the wavelet basis used is a 2D bi-orthogonal Daubachies wavelet applied using high and low filters (based on a quadrature mirror filter setup).

Don't work with images though so there might be some else with more detail.
 
Back
Top Bottom