The irony of film beating digital at its own game.....

[sperera]

This is my opinion as far as I see things based on what I've shot and compared: when reducing images to web site sizes like 650 pixels for marketing purposes, for example I shoot for online fashion shop etc etc the film shots without fail reduce INFINITELY better than any digital shots.....the images look crisper and hold detail wayyyyyyyyyy better.....

IRONIC

 

[Loris Medici]

First of all, I don't know anyone who will able to differentiate film and digital capture resampled to 650px (long dimension) - given they were downsized appropriately, including myself.

My usual downsizing method is to USM at original size (Amnt. 50 - 150, Radius 0.9, Threshold 0) then use bicubic interpolation to resample and then use 0.3 - 0.6 px high-pass sharpening at final size . Yummy!

What do you do while downrezzing? (Both capture methods...)

 

[Worker 11811]

I agree with you in principle that, when sampling from film's random structure to digital's Cartesian structure, there is a beneficial filtering effect which will make scaling appear to be better but, once the image has been scanned and stored digitally, all bets are off.

Thus, the only benefit you get is during the transition from film to digital. Once the image has been digitized you are back to square-one.

 

[Matus Kalisky]

I am wondering whether this has to do with something else. I remember seeing two scans of one 35mm slide somewhere on flickr (the subject was a rose) one made with Coolscan 5000 and one with Epson V700 (or similar). Even though there is a clear difference in real resolution - one would expect that downsizing the images to some 800x600 pixels there can not be any difference. Well - the difference between the scans was that the one made with the Coolscan was "crispier" with better micro contrast.

Now - if you realize that the information stored in a digital capture is about 60 - 70% of the pixel count because of the Bayer interpolation and anti-aliasing (digital never looks really sharp to me at 100% on screen) - and those 30 - 40% are simply LOST - no sharpening will ever bring them back. On the other hand - good quality film scan has the true detail and resolution very close to the sampling resolution (NOT true for flatbed scanners)

Now if you downscale the image - how should the software know what part of the image is you detail and what part is not? The differences between the images will become less obvious but the one that was truly full of detail at the beginning will most probably look the best.

Now I would love to hear form some experienced guys whether my theory has something to do with reality ...

Actually - I may try to check myself - I have Microtek F1 at home and some 35mm slides that were already scanned with a Coolscan 5000. I guess I will give it a try!

 

[Loris Medici]

The key is to process each image appropriately / individually, depending on desired subjective effect and subject matter. (Notice how much play there's in the workflow I describe...)

You can make the digitally captured image look pretty sharp, w/o that "in your face" bad taste; all depends on your talents in post-processing. (Detail / resolution is another issue BTW - I talk about neither here... *Addition: At screen size I highly suspect that any difference in minute detail will be perceptible - given the condition that the digital capture wasn't extremely inferior compared to film.*) OTOH, depending on your scanning equipment and technique, you may hit the grain wall with film. (Not necessarily, but that's a solid fact in many casual circumstances...)

 

[sperera]

thanks all for your comments.....as I said, this is all in my experience and based on visual analysis.....film shots are able to reproduce the minutest details where digital seems to fudge things or have a halo if oversharpened.....its the acutance and higher grain relationship compared to digital that creates this visual illusion I would say

 

[Ray Heath]

Yeh, but what a pain to process, all those extra steps to use film for purposes for which it was never intended.

 

[Worker 11811]

"Photographer" means "One who paints with light."

Process is irrelevant. It is the result that counts. Regardless of whether it is digital or traditionally made, a photograph is a work of art, in and of itself.

That does not justify using a circuitous method just to produce a simple image but, within limits, it is up to the artist's discretion as to the manner to produce that image.

Photography is a peculiar art form because it blurs the line between art and technology to the point where it borders on the verge of a performance art. This is the reason why many people get wrapped up in the equipment and the methods used to produce an image when they should be concentrating on the substance of the image they produce.

 

[ctscanner]

Randy,

I tend to agree with what you're saying. I think you hit it on the head, too much concentration of mechanics and tools, and not enough on what's being done with them.

Back to the main discussion for a moment. If you are attempting to compare a digital camera capture vs a film scan - it would seem that with the digital camera, unless you were shooting in Raw, that you would have no control on how the image is sharpened to begin with, therefore possibly starting behind the eight-ball as it were.

As regards down-sizing for the Web: I generally will take the image down in 5-6 steps to arrive at the file dimensions that I intend to upload. I normally apply sharpening twice, once when i have completed cleanup with the full-size scan, which will vary depending on the image, and once again just prior to upload and that will usually around 10-15% of the first application. I haven't tried to use high-pass sharpening yet - something to think about.

George

 

[Worker 11811]

That is the synopsis of the lecture I got when I turned in my final exam to the photography teacher I studied with in College. (Tom Petit - R.I.P.) He said I need to forget about technical things and concentrate on making IMAGES.

When shooting with a digicam, I always avoid JPEG and shoot in RAW if the camera I am using will support it. You are right. The RAW exposure and balance controls coupled with sharpening upon import will provide a superior picture. As much as possible, I try to use JPEG only as the final step in the process when I'm making the file which will be posted via the internet or sent to a client or other recipient. Better, sharper, clearer results will always be obtained, IMHO, when done that way.

According to the latest books (namely "Photoshop for Photographers" by Martin Evening) it is no longer necessary to use step interpolation to enlarge or reduce a digital image. If you are using the latest versions of Photoshop and you make sure you use the correct interpolation settings ("Bicubic Sharper" for reduction or "Bicubic Smoother" for enlargement) you can get good results from one single step instead of several smaller ones. He does say that many people still believe step interpolation is better but he says it is not strictly necessary, in his opinion.

As to whether a scanned photograph will beat a digicam image, I agree with a caveat: It all depends on how both images are imported at the outset. A crummy scan can not beat a good digital import from Camera RAW using a good quality camera but a good scan, done properly, can beat even a good image from a digicam if the user knows what he's doing.

The thing is that, once imported from analog to digital, a traditional photograph should be on equal footing with a digicam image. I do not believe that there is a magical quality in a traditional photo that makes it better than a digital image. The benefit only comes at the beginning of the process.

Like I said, I believe that the scanning process, itself, acts like a noise filter. I also think that a well done photograph can hold more information (detail and dynamic range) than a digicam can produce. The quantization process will capture more information than the digicam.

Does that mean that a photograph is always better than a digital image? No. It goes back to the image over process issue. The photographer has to ask himself what he's trying to accomplish and tailor his technical process to his goal. If he is trying to produce an ultra-sharp image that reaches out and slaps the viewer in the face, so to speak, he will probably need a traditional photography work flow. If he is producing graphics and images for a web page, a traditional work flow would be like shooting flies with a Howitzer.

It all depends on the image the artist is trying to produce.

 

[Ray Heath]

G'day Worker,

Why not post some images to illustrate these these contentions?

 

[Worker 11811]

Which? Step interpolation? Or analog import versus digital origination? Or both?

This could be a good thing.
I have a bunch of negatives shot on T-Max 100 and I've got some color slides on Ektacrhome. I could borrow the 10 MP camera from work and shoot some identical shots with each then scan the film and import the Camera RAW. I could shoot in JPEG to use as a control.

The only thing is that my scanner is a Canon flatbed. (Canoscan 8800F) It produces good quality but do you think it will be up to the task of comparison? I suppose it could still prove what the average user could expect. Right?

This sounds like a neat experiment. It certainly has my interest.
Any suggestions on how to carry it out?

 

[Matus Kalisky]

Worker, I would say that a flatbed will not be able to provide 10MP clean image scan from 35mm. With realistic resolution of about 1600 - 2000 spi it will deliver only some 4 - 6 Mpix of information and those will be hidden in a much larger scan (probably 2400 spi - cca 9Mpix). Still it may be interesting to see the outcome. A 3200 - 4000 spi Coolscan 5000 scan would be more fair to compare the DSLR to.

Actually - a comparison to result from a Foveon sensor would be also interesting - but I do not have any Sigma camera at hand.

 

[Worker 11811]

I have some very old 4x5 B/W negs that my father shot 50 years ago. There are some 2-1/4 inch B/W negs too. They are all very good shots but many of them show their age. I have gotten good images out of them.

Only thing is, I wouldn't be able to make any digital shots of the same subjects to directly compare with.

 

[ctscanner]

Randy,

Certainly would be interested to see what you come up with. I would agree with Matus, if you have access to a Film Scanner it would probably provide for a more meaningful comparison.

Looking forward to your results.

Georgge

 

[SilverGlow]

It is not true that film "beats digital at it's own game"...this notion is absurd at best. When one scans in a film negative, it becomes digital, and going forward is in fact a digital image. At 650 dpi, one cannot tell one from the other in regard to sharpness. As to dynamic range? Yes, film beats digital there.

 

[Worker 11811]

Just for info: I have some pics that I took with the digicam from work and some pics of identical subjects taken with Tri-X Pan. I have the film developed and the CR2 files imported. I just have to get them all ready to upload.

Or, would you think it better for me to just post the original files, unedited, and let them speak for themselves.

 

[Ray Heath]

Just for info: I have some pics that I took with the digicam from work and some pics of identical subjects taken with Tri-X Pan. I have the film developed and the CR2 files imported. I just have to get them all ready to upload.

Or, would you think it better for me to just post the original files, unedited, and let them speak for themselves.

no, just keep talking around the issue

lots of these statements are made re: film better than digital, this scanning better than that, this digital technique better than that one, blah, blah, yet no-one is ever game to offer visual proof

does it matter which mode of capture is "better", too many forget the basic fact that the imagery is what is important

 

[donbga]

too many forget the basic fact that the imagery is what is important

Ditto what Ray says.

 

[pschwart]

no, just keep talking around the issue

lots of these statements are made re: film better than digital, this scanning better than that, this digital technique better than that one, blah, blah, yet no-one is ever game to offer visual proof

does it matter which mode of capture is "better", too many forget the basic fact that the imagery is what is important

imagery is what is important
Thank you! There is a lot of cool technology at our disposal, and it is fun to play with, but I hope we don't get too fixated on hardware and software debates unless it is to demonstrate how it can make our prints and images more successful.

 

[ctscanner]

Add me to the list of those who agree with Ray's thoughts regarding this debate.

Speaking for myself, I don't really know if it would even be possible to address all the variables involved in speaking to the OP's original assertion. An interesting exercise perhaps, but to do it correctly would seem to be problematic. And, when all is said and done, hasn't this type of discussion and analysis been done before - particularly when Digital Cameras first started to appear, and then as they became more prevalent. And, speaking for me, well I have looked at a fair amount of images originating with a D-Camera, as well as those that were shot on film, and scanned to create a digital file, and for the most part, rather I liked or disliked them, had nothing to do with the method used to generate them.

So, in conclusion, what I think is this. That I personally would like to see a lot more images posted to generate discussion about how they were agreed. After all, I think the consensus was: The image's the thing!

Thanks for your attention, and please excuse this wordy post.

George

 

[sperera]

thanks all.....but we're missing the point....

the very simple point i was making is for some reason, in my 'hybrid world' where I shoot both digital and film and digitize both for commercial reasons when i have to market the images in web site, facebook and so on the film shots reduce better to 600 pixel size than digital shots.....that was all i was saying!!!!! heh heh heh

 

[Ray Heath]

well, perhaps some of us are missing the point that this is a visual imaging site, and when one of us makes a strong claim as to the visual strength of one technique over another, one could/should/possibly must give a visual example as to the truth of their strong assertion, hey hey

 

[pellicle]

well, perhaps some of us are missing the point that this is a visual imaging site, and when one of us makes a strong claim as to the visual strength of one technique over another, one could/should/possibly must give a visual example as to the truth of their strong assertion, hey hey

unless one goes to all that trouble and then gets it rammed down ones neck that you're a sham and faking the images.

easier to just type something assertive.

 

[sperera]

as I say, people are misinterpreting the statement.....Iappreciate the feedback from everyone though

It is not true that film "beats digital at it's own game"...this notion is absurd at best. When one scans in a film negative, it becomes digital, and going forward is in fact a digital image. At 650 dpi, one cannot tell one from the other in regard to sharpness. As to dynamic range? Yes, film beats digital there.