re: digital imaging

sonicboom wrote on 7/4/2002, 2:39 PM
i met someone yesterday who works for a company in nyc called Celartem technology
she claims her companies software can take any digital image -- regardless of how many mega pixels it has - and enlarge it without loss of quality
in other words, my 2.1 digital camera will come out as sharp as a 5 mega pixel camera
i can blow up the pictures as large as i want without any difference
i went to the website
www.celartem.com
i am going to download a trial version and see for myself
for the record, i don't work for the company or know anyone who does--with the exception of the person i met yesterday for the first time
i am curious as to what others think of this software
or is it a case of "to good to be true"
let me know what you think
thanks
sb

Comments

FadeToBlack wrote on 7/4/2002, 3:08 PM
BillyBoy wrote on 7/4/2002, 4:17 PM
Talking apples and oranges. There are many schemes that 'blow up' images ie making them larger without apparant loss of quaility. The fact is pixels are added through some method like bilinear and bicubic interpolation.

I think you may have heard what you wanted to hear: "...my 2.1 digital camera will come out as sharp as a 5 mega pixel camera" as opposed to what was said: "...take any digital image -- regardless of how many mega pixels it has - and enlarge it without loss of quality.

In other words larger isn't sharper. How much detail any image has is determined when the image is first created. Therefore a 5 megapixel camera will caputure five times the pixels of a 1 megapixel camera. True, you can blow up the 1 megapixel image in size, but you are not bumping up the quality. Loss may not be that apparent, but enhancement meaning a image with vastly improved detail is just so much hogwash.
Chienworks wrote on 7/4/2002, 6:01 PM
In my old Apple // days, i blew up a 0.05 megapixel image to a wall sized (6 x 8 foot) poster. Quality loss was absolutely zero. However, the pixels did end up being about 1/3" square. Up close it looked like a disaster, but the poster was hung high on the wall in an auditorium and from most of the room it was very striking.

... and it only took 11 hours and 3 ribbons on an ImageWriter II. wooooooooo!
SonyDennis wrote on 7/4/2002, 7:09 PM
There are a number of companies with similar technology. The only ones that can actually create sharpness where it wasn't before are using multiple images to create stills. They watch the movements between frames, upsample them and morph them all onto the same target locations. In the end, you end up with a pretty sharp image. Here's one I've never worked with, am not associated with, and I've only seen their website and videos:

[Video Focus from Salient Stills]
http://www.salientstills.com/product/product_intro.html

Your milage may vary, but I can say that upsizing a single image can only get so good, the information is not there (you can probably get 2x blowup with a few artifacts). However, using multiple images that have movement (camera or subject) CAN create more resolution for a still.

///d@
SonyDennis wrote on 7/5/2002, 12:06 AM
> if that were the case, then everyone can use VHS and produce HD TV shows.

Wouldn't that be nice? But it does remind me that a LOT of early HD content is going to be up-sampled SD content. Fortunatly, the SD content that studios have in 4:2:2 or better will upsample well enough on expensive upsamplers that most people couldn't tell the difference from real HD, but it's still a shame, and I hope people don't shy away from buying HD sets because the pictures aren't as sharp as they expect them to be, when they're really watching upsampled SD content!

///d@