Canon Rebel XT

Re: Canon Rebel XT

I had this point when I started my reply, but I forgot :hehe:
Most of today’s labs are automated so when there are different settings in different pics of the same role the lab machines just flatten out the all the pics with the same settings, sometimes actually killing the effect!

Re: Canon Rebel XT

Film still has a greate resolution. But as you all know it's not perfect. You get film grains, digitising what you want can be a pain in the arse. On the other hand a £1 disposable will give relatively good pictures.

Digital has its flaws aswell, comapred to 35mm small CMOS/CCD sensor. Uses bayar pattern which interplotaes the result in every digital camera minus triple CCD video cameras and Sigma DSLR who have developed a sensor for detecting individual colours for every pixel. So when you 16mp thats not quite true, as it's not capturing RGB per pixel but BW with a filter.

Re: Canon Rebel XT

but it does a damn good job of calculating that color.. tests have proven that Digital SLR cameras are more "color accurate" than film.

The best films have a Delta E (deviance from perfect color reproduction, lower is better) of around 11.. The best digital SLR's have achieved mean Delta E's of under 5!!

Re: Canon Rebel XT

Just posted some pics from my D70s in the Images2 section:
http://www.paklinks.com/gs/showthread.php?t=217291

Re: Canon Rebel XT

May well be more accurate but what I was pointing out that each pixel is not captured in RGB but a matrix of mono pixels in the pattern of RGBG known as the Bayer matrix. The results are then intrplorated so you loose a vast amount of detail at full resolution. Infact you could say to get a real resolution you need a single block of RGBG to get an actual pixel. So that means divde the reolution by 2, as it’s a matrix division by 2 = divide by 4. So that 16MP camera can only capture 4 million real pixels.

Some examples:

Perfect capture (RGB)

http://upload.wikimedia.org/wikipedia/en/7/71/Shades_aa.jpg

Bayer RAW Capture (RGBG)

http://upload.wikimedia.org/wikipedia/en/e/e6/Shades_bayer.png

After interploration and addition of white:

http://upload.wikimedia.org/wikipedia/en/b/bb/Shades_dc.jpg

Re: Canon Rebel XT

what but risc bhai jaan.. your first argument was that the bayer interpolation dosn't capture color. Well that's been refuted as DSLR have better color accuracy.

[quote]
The results are then intrplorated so you loose a vast amount of detail at full resolution.
[/quote]

ok.. so now about resolution.. the proof of a pudding is in the eating.. therefore setting aside all mumbo jumbo of the science behind it here's a very recent (and very timely for this discussion) test done with medium format film and medium format backs. They also included the 16MP Canon DSLR just for kicks to see how it measures up against the big guys..

You decide..

Top is a 39MP digital back (P45 from Phase One)

Middle is 4x5 format Velvia Film

Bottom is 16MP Canon 1DsMkII

Re: Canon Rebel XT

[quote]
what but risc bhai jaan.. your first argument was that the bayer interpolation dosn't capture color
[/quote]
I did not say that. Have a re-read.

Not exactly high res and as I outlined downsizing a picture shows less of the interploration. BTW 35mm film can be digitised to around 64MP. Unless you are having a photography display with large prints e.g. A2 size you wont notice much between a top end DSLR. However the science behind all but Sigmal CMOS/CCD is not perfect. Although I would buy a Canon 350D if I was going to buy a camera tomorow it's not perfect technology.

Re: Canon Rebel XT

risc bhaijaan these are 100% crops not resized.

with what? only the highest quality drum scanners? And trust me at that resolution it’ll all be grain and nothing else. If 35mm film was so ‘enlargeable’ there wouldn’t be any 4x5 or large format shooters out there.

The science may not be perfect, but it’s good enough to beat the crap out of film. Even the old 6MP models do.

Here’s another link to more information. Read up when u get a chance.

Re: Canon Rebel XT

guys, even IF 35 mm film would give better images, most photographers look at other things aswell. Like the price? How cheaper it is to use a DSLR instead of film? How easy it is to ps on the computer? 99 out of 100 pictures taken with DSLRS are processod on the computer before anything else.

Re: Canon Rebel XT

I still have a problem with de-mosaicing interpolation. I dont like the idea that 6 out of 9 pixels are calculated rather than captured. Until CCD simillar to Foveon become mainstream I’m not in a hurry to buy DSLR

http://www.foveon.com/article.php?a=67

http://www.dpreview.com/news/0202/02021103foveonx3preview.asp

http://www.dpreview.com/news/0202/02021102foveonx3tech.asp

Re: Canon Rebel XT

You're paying too much importance on how-it's-done than the how-does-it-look.

Rather than look at charts or scientific explanations, look at pictures.. Bayer interpolated sensor implementaion produce far superior results than the Foveon implementation. So far only Sigma produces one and sad to say they are the bottom of the barrel when it comes to picture quality.

At times, it's best to rest aside the measurebation and trust your own eyes.. afterall that's all photography is about.. taking pictures and viewing results, on screen or in print.. not siting around analyzing the bytes in the files. :)

Re: Canon Rebel XT

Science always plays a part. Just look at the difference in these two pictures assuming both are captured with a perfect lens and sensor.

Re: Canon Rebel XT

These are not pictures but illustrations. You want the real comparison between a Foveon and COMS/CCD then look at side by side comparisons from both. Real pictures from cameras that use the technology.

Re: Canon Rebel XT

Call it whatever you want but the science is still the same. You cant escape from that fact that the pictures are interplorated. Interploration looses detail.

Re: Canon Rebel XT

You can escape from it and everyday people create images that are a testimony to that very fact.

I really don't get your logic here. Just because of the processes behind the scenes, you're going to not believe your own eyes? Praising a technology even if it doesn't deliver is stupid IMO. I'm telling you having seen images from both technologies, the ones from CMOS/CCD with bayer interpolation are years ahead.

Foveon might be great on paper and in principle, but it hasn't given us anything in terms of improved picture quality or color fidelity yet. That's the intriguing part about science and practical life.. In theory a LOT of things shouldn't be the way they are.. but in real world applications, things are much different.

Re: Canon Rebel XT

Im not saying at the moment a Bayer matrix does not achieve better overall pictures. merely pointng out the flaws in design which you seem to take to heart. You cant deny purly from a scientific point of view theroretically it's a flawed design. Bayer gives good pictures but not perfect.

Foveon in comaprison to Bayer is a new technology. You have to remeber CCD was first invented by NASA for military applications over 50 years ago. Hubble for example has a 1MP camera. Bayer matrix has been used for time so developed much further. I still think the Sigma SD10 which really has ~ 3.4MP gives good reults in comaprison to any 6MP DSLR camera. Especially when you view the pictures at full resolution.

Re: Canon Rebel XT

Some extracts from dpreview:

the X3 sensor technology achieves amazing levels of detail and resolution pixel per pixel compared to the six megapixel sensor of the EOS 10D. Indeed it’s fair to say that the EOS 10D doesn’t truly exhibit any more visible detail than the SD10 (just a larger image

‘Single pixel resolution’, pixel for pixel much better than Bayer

Resolution of enlarged images virtually indistinguishable from six megapixel Bayer D-SLR

http://www.dpreview.com/reviews/sigmasd10/

Re: Canon Rebel XT

Also look at these sine patterns. Massive difference in level of detials where lines are clost to each other:

SD10: http://www.dpreview.com/reviews/samples/rescharts/sigma_sd10.jpg

D100: http://www.dpreview.com/reviews/samples/rescharts/nikon_d100.jpg

D100 anticipates a grey instead of a white next to the black pixels as it’s guessing.

Re: Canon Rebel XT

FOVEON has been out long enough..

why compare dead models like the 10D or D100??

A D70 NEF output rivals anything the SD10 could put out, the Canon 5D just decimates it.

You need to select bodies with 'weaker' AA filters to compare against FOVEON, not 10D or D100 which have heavy handed AA filtering.

Re: Canon Rebel XT

Foveon X3 released in 2002 so it's not exactly new in todays world of electronics. Compnay founded in 1997 not quite the over 50 years Bayer matrix been in development. Not even 10 years for Foveon.

Comaprisons are made by dpreview which I'm sure you know is a highly respected review site. The models you outlined where it's competition at the time. Foveon just happens to have been more successful in industrial imaging over retail so a lack of DSLR with their sensor.

BTW the D100 has a reolution twice of the sigma yet still captured less intracate detail. Just shows the flaws of Bayer matrix. Regardless of how much AA is made you cant have a picture withought interploration which is the heart of the problem as I keep on mentioning.