10d enlargements versus scannned 35mm

[greg_tan]

I was wondering what size enlargements can be done with the 10D? I

have heard some people doing 20 x 30 enlargements. How does this

compare to say 35mm Portra scanned with a Minolta 5400 dpi scanner?

I know puppy face has a lot of experience with the Canon 4000 dpi

scanner with film versus the 10d so this question is most directed

towards you. How much does Genuine Fractals help?

[Greg M]

I use a Digital Rebel myself, but the sensors are very similar, if not the same as the 10D. My 8.5x11 prints look great, and I've had 11x14 printed that also look really nice. I'm not too sure how much larger I would go but, if someone were to look at it from a few feet back, I would think 16x20 is presentable as well. What looks good to one is not presentable to someone else, though. All a matter of taste and expectations I guess.

 

Velvia slides scanned with the Canoscan FS4000 scanner at 4000 PPI (file size of 5888x4000) are REALLY clean and nice and can easily go 16x20....probably 20x30 too, although I have not tried it. I have scanned other negatives and slide film, but at higher ISO ratings the grain starts to get a little intensive, but useable if you apply a program like Neat Image. I've decided anything over ISO 100 I'm better off using the Digital SLR.

[lu_yin]

20x30 from 10D actually looks pretty nice after enlargement and sharpening. The print won't look ultra sharp with magnifying glass, but quiet good in normal viewing distance Scanned 35mm are OK up to 8x10, but would start to look really grainy when you printer bigger sizes.

[jonathon]

I think that's quite funny, two really opposite opinions as the very first two responses 8-}

[Greg M]

Not really.....

 

As I mentioned above, negatives and other positive film above ISO 100 (at least through my Canoscan FS4000) isn't nearly as good as using my Digital SLR- too grainy. I stand by what I said regarding the Velvia slide material , though. What Velvia slides I have scanned so far are very impressive.

[mormegil]

I sometimes worry about sharpness on the 300D/10D on enlargements.

 

Bob Atkins excellent article on Digital DoF http://www.photo.net/learn/optics/dofdigital/ states that at 8x12, you would need to have circles of confusion down to 18.45 microns on the 10D sensor.

 

The 10D sensor has got 3072 pixels horizontally. So at 22.5mm accross the sensor, that's 7.32 microns / pixel. That seems to be well under the 18.45 micron resolution required for "acceptibly sharp."

 

The problem is the Bayer pattern. Only 1/4 of the pixels on the sensor are Red, or Blue, and 1/2 are green. The pixels you get out of the camera have Red, Green and Blue values for each of the 3072 pixels accross. The difference is made up by interpolating (mathmatically guessing) the remaining values. In essence, you loose some sharpness in solid Red, or Solid Blue areas (less in solid Green areas). You get the sharpest detail in the neutral areas.

 

So that means you have a real resolution of about 29 microns for Red and Blue, and 15 microns for green.

 

So Red and Blue are above "acceptibly sharp" at 18.45 microns.

 

I guess that's why we apply USM

[PuppyDigs]

The quality of prints from film scans depend on the film scanned, the technical quality of the image (sharpness & exposure), the subject matter, the scanner and your PS skills. I find prints from my 10D and scans of Portra 160VC and Provia 100F are about equally excellent up to 8 x 12, printed on my Epson Photo or a local Fuji Digital Frontier.

 

For landscapes larger than 8 x 12, 10D images start looking soft--albeit nearly grainless. That is, small details in landscapes are somewhat indistinct compared to film. Fine grained film resolves fine detail better. For large subjects with blurred backgrounds--head 'n shoulder portraits--the 10D files can handle much more enlargement than those with landscapes. 12 x 18 and 16 x 20 prints look great at normal viewing distances. Still, my Portra 160VC scans show more detail and but a little more grain than the 10D files. However, at normal viewing distances the different isn't very apparent. I haven't made any 20 x 30 prints from my 10D, but my Provia scans look surprisingly sharp--no grain is viable at normal viewing distances.

 

Incidentally, my Portra 160VC scans show a much wider dynamic range than both Provia scans and 10D files. The Portra negs actually have more information than my poor FS4000US can record, so sometimes I have to scan twice, once for shadows and low mids and a 2nd time for upper mids and highlights.

 

I enjoy working with both film and digital and use them about equally nowadays. They both have unique "flavors" that I find enticing.

 

I don't use Genuine Fractals (I use the stairstep routine).

[paul_kurian]

From 10D, I am doing enlargements till a size of A2(42cm x 29.7cm) and the result is perfectly alright especially if I shot in RAW format and then saved as TIFF file. I am sure bigger enlargements are possible. Film scanning I didnt do so far.

[james_barber1]

I've had a couple of slides and Negs scanned Hi-res, and the first thing I noticed is how grainy they are. Granted, the slide was not velvia, but still, I was very surprised. Velvia does have a reputation for being superlative though (its just ridiculously overpriced in South Africa).

 

800 ISO negatives are ridiculously grainy, and you can really see it when you view the image full size on a screen that can handle it.

[fotografpeterlundberg.se]

All of you who scan film - make sure that the scanner software does NOT apply any USM by default. Maybe thats why the films starts to look grainy?

 

regards

Peter

[www.philwinterphotography.]

I use the 10D and the FS4000US scanner. While I've learned to scan pretty well, I'm getting better prints of every size up to 13x19 than I get when I scan. And that's using the same lens for both film and digital. I shoot both RAW and jpegs (can't tell much difference) and convert them to tifs to adjust them. I've not tried prints larger than 13x19, but I don't see why the 10D couldn't produce excellent prints at least 16x20 if the image itself is good (100 ISO, tripod, etc).

 

Phil

[scott_ferris]

Hi all,

 

I only use film seriously but do have a 5MP S50 so am not a

complete idiot on the digital front. I also have a 3600DPI scanner

for messing about with.

 

What I don't understand is why, if you have taken a good Velvia

or Provia slide why you would bother getting a hi res scan

instead of just getting the thing printed ? I regularly get 20X30

and 24X36 prints done of my Velvia and the whole thing costs

less than a scan! Now if its a wrongly exposed slide of a once in

a lifetime event I can understand the effort and the advantages

the computer can give you but short of that why bother? A low res

scan can give your printer all the detail it needs for 8.5 x 11's and

if you have a top flight printer you should be useing digital

cameras anyway, well at least not slide film.

 

Can anybody enlighten me ? Hopefully politely!! Take care, Scott.

[mark u]

I agree Scott - scanning film is almost bound to reduce final image quality compared with making a regular enlargement.

[daniel_taylor]

"I was wondering what size enlargements can be done with the 10D?"

 

Good 16x20 is easy to achieve with a 10D. I've seen pretty good 20x30, although that's ideally the domain of higher resolution digital, MF, or LF. Of course you tend to stand back more the larger the print, and standing back the 20x30 I've seen from 6 MP digital looked great.

 

IMHO, digital enlarges much better than 35mm film due to the lack of noise and the better acutance. You would need a film like Velvia or Provia 100F and an excellent scanner to even hope to keep up. Grain becomes more and more intrusive as you enlarge, and grain is an issue by 8x10 with the finest grained films I know of (Provia, Astia, Velvia, Reala, etc).

 

People talk about fine detail in 35mm film enlargements, but I just haven't seen it, either in my own work or in tests. The site below tested Provia 100F, NPS 160, and a 10D and claims to have found one spot in one test image where Provia rendered more detail in a patch of grass. I guess it did a little bit, but all three are a mess at that level of extreme enlargement. Nobody standing back from a huge poster print is going to notice an extra blade of grass and say "wow, that really makes the image." They will notice overall apparent sharpness (acutance) and noise or lack thereof. (Interesting to note that the author felt digital rendered the grass better than the neg film.) At normal enlargements, the extra blade of grass won't even show up, much less be noticed by anybody.

http://194.100.88.243/petteri/pont/Pontification/m_Aesthetics_Shootout/_Is_slide_better.html

 

"I guess that's why we apply USM"

 

USM has nothing to do with Bayer color interpolation and was actually a technique invented in the darkroom. The Photoshop routine just emulates the technique. We apply USM to increase contrast between edges, acutance, in order to make the details recorded stand out more. In my experience USM is always needed with film scans despite the fact that scanners do not use Bayer. It's needed more with certain lenses, less with others.

 

With a DSLR you tend to apply more USM because you can (lack of noise), yet if you compare images straight from a DSLR and a film scanner, the DSLR image already has more edge contrast. Quite frankly, DSLR's are starting to redefine "sharp", leading to some people claiming digital images are "too sharp". Velvia is about the only film that can touch a 10D on acutance in my experience. This is supported by MTF graphs, as a 10D sensor stays above 50% MTF at higher lpmm frequencies than the vast majority of films.

 

Bayer does not impact resolution in the way or amount that many people theorize. Luminance information is far more important to resolution and fine detail than color information, something that eye doctors and engineers have known and taken advantage of for a long time. Our eyes also work on this principle (or perhaps the principle exists because of how our eyes work). You can't just divide some numbers by three and say that Bayer means a sensor is not sufficient for a given enlargement.

 

"What I don't understand is why, if you have taken a good Velvia or Provia slide why you would bother getting a hi res scan instead of just getting the thing printed ?"

 

Honestly? Because digital workflow, including digital output, passed up the traditional wet lab a while back. This is especially true of slides.

[nick_hallbeck]

I use a 300d plus RAW and adjust everything in PS8s RAW mode. If I bump up the pixels to 6144x4096(+) and keep my 300 resolution could that help my enlargements. I also heard of using Unshape Mask. Anyone try those techniques?

[nick_hallbeck]

oops I forgot line. When using you usm what is the settings that gets the best results?

[mormegil]

On the mention of Bayer not affecting sharpness, and luminance being much more important...I'm a little unclear.

 

If you're taking a picture of something purely red, then only 1/4 of your sensor pixels are recording any luminance data.

 

 

 

I also mentioned application of USM. I wasn't implying USM was needed to fix problems with bayer directly, just sharpness issues (which, may stem from bayer).

 

Well, I've heard that Foveon sensors give more sharpness for less megapixels. If that's true, then bayer does loose something in sharpness.