Film vs Digital: How much cleaner is it ?

[johnw63]

<p>Perhaps I can be more specific, since the post got moved from the forum where I put it.</p>

<p>Every time a new DSLR comes out, while you can read about other aspects of it's output, the BIG question is " How much noise does it have at XXX ISO ? " Lot's of zoomed in crops and comparisons to this DSLR or that one. My issues is that I don't HAVE any of the above, so the comparison is vague, and I don't zoom way in on my film scans from an inexpensive Epson V500 scanner, so the shots don't have a point of reference. It's all about high ISO noise for most these reviews.</p>

<p>The case in point, for me , is a review at DPReview with the new Nikon D7000 and the Canon D60. All I could say is, " Yes. They both got noisier as the ISO went up. " But, how does that noise compare to real world use ? How big could you enlarge before it's an issue ? How does that change as you may need to crank the ISO up for odd situations ?</p>

<p>What I am looking for, and some of you got it, was how does this compare to good film scans, or even medium format stuff. I know that if a good photographer used a medium format camera, they could blow up to 16 X 20 easily. It would sharp and vivid. What ISO on a crop body camera, like a D7000 ( or D300 ) could I still do the same ? </p>

<p>If the progress of digital bodies is to the point that the noise is so much better than even MF film, and large prints are not a problem, why is noise the camera test reference in so many tests ?</p>

<p>On a side note, just because I have been posting for some length of time, doesn't mean my question is already answered. Comparisons of modern DSLRs to film are not really made anymore. It's all about previous DLSRs or the competing brand's camera. All the posts are web page sized images. I don't recall any posts I read where some one says " That print was as good or better than anything I could get from my Hasselblad ! "</p>

<p> </p>

[bobcossar]

<p>John, when I finally decided to switch from film to digital, I made one last test: I very carefully made the best exposure that I could....tripod, MU.....everything, of a set scene with good light<br>

One was shot on my Hassy, and the other on a Canon 10D. Actual custom 16x20 prin ts were made....the film on an enlarger to maximise the full film potential....</p>

<p>The 10D image was better in every respect, and that day I changed aproach to photography. I should add that the 2 images were in my studio for a while, and everyone who looked at them felt the same as I had.....hope this helps a wee bit....regards, Robert </p>

[johnw63]

<p>Thanks, Robert.</p>

<p>I don't know the specs on the 10D, so, I'll do some research.</p>

<p> </p>

[nathangardner]

<p>10D: 6.1MP APS-C CMOS sensor</p>

[bobcossar]

<p>Nevertheless, Nathan.....the results were as I described them.....hard to believe in this day of high pixel counts.....but it happened as I said...Robert</p>

[johnw63]

<p>How then, do the sites that calculate the MP needed to approximate film of a given size, end up with so much HIGHER numbers, than 6MP ? After I looked up that camera, I started web surfing in some earnest.</p>

<p> </p>

[User_2000406]

<blockquote>How then, do the sites that calculate the MP needed to approximate film of a given size, end up with so much HIGHER numbers, than 6MP ?</blockquote>

<p>Just wondering here to what extent film flatness, or lack thereof, might sabotage the theoretical resolution of film? I've seen plenty of explanations as to why a given film should have such-and-such linear resolution and thus a particular megapixel equivalent for a given format, but to what extent does that actually pan out inside of actual cameras as opposed to in lab tests?</p>

<p> </p>

[StuartMoxham]

<p>You need 300 ppi to make a high quality print maybe less in some cases for large prints that would be viewed from a distance and maybe more if you like small prints viewed in good light quite closely. A 6mp DSLR (Canon 10D/ 60D, Nikon D70/D100/D40) has 3000x2000 pixels. It can make a 10 inch wide print at 300 pixels per inch.</p>

<p>Thats it once you go bigger the number of pixels per inch will be less. A 20 inch wide print from a 6mp DSLR will only have 150 pixels per inch. If you interpolate or upress you cannot add anymore details so you are just adding fake pixels to spread the file over a larger area.</p>

<p>Now thats not to say that the prints can't look great I don't doubt Robert saw for him better cleaner prints. People were rather blown away by just how big they could print these clean digital files and of course if the files were clean there was just no grain.</p>

<p>I've seen great 12x18 portraits of a baby from an old Canon D30 thats only 3mp. The 8x10 prints I've made from my D1h look great. I've made a few 12x18 inch wedding portraits from a Nikon D70.</p>

<p>Trouble is what is going on, there is no doubt that a shot from a Hasselblad on 120 film should easily have more details than a 6mp DSLR a piece of decent 35mm film should have more details.</p>

<p>Theres no doubt in my mind that my small hand enlarged B&W 3.5x5 inch to 5x7 inch album prints have more details and look better that a simlar sized prints from a digital minilab.</p>

<p>For me I think it's the large clean prints when the prints look cleaner they often do look better. Many see the large clean print from a DSLR and because it is clean they equate it with medium format quality even if it is not as detailed it does not matter it's clean.</p>

[dan_south]

<p>John, I'll attempt to answer your questions on noise. First, here is some background.</p>

<p>Digital sensors are susceptible to different types of noise. Digital images contain more noise when they are shot at high ISO values because the light must be amplified more. It's analogous to sound recording. If you have two singers, one with a loud voice and one with a very soft voice, you would have to amplify the soft singer's voice more in order to get a stong enough signal to record. This amplification boost noise in the recording. Cameras shooting at high ISO do the same thing. They boost the light that hits the sensor so the sensor can make an image with less light.</p>

<p>Then there is another kind of noise that shows up in long digital exposures. During the long exposure, the electronics in one pixel throws of electrons. Sometimes these electrons are mistakenly interpreted as light by neighboring pixels.</p>

<p>A third kind of noise (similar to the first) shows up in the shadows even when shooting at low ISO values.</p>

<p>Most digital noise problems are worse with a sensor of a small size. The pixels on a small sensor are smaller, less efficient, and closer together, all factors that increase noise, all other conditions being the same.</p>

<p>Film has no equivalent to noise, but film has grain, and grain increases as ISO increases.</p>

<p>Now, to you questions.</p>

<p></p>

<blockquote>

<p>Every time a new DSLR comes out, while you can read about other aspects of it's output, the BIG question is " How much noise does it have at XXX ISO ? "</p>

</blockquote>

<p>Yes, because in the last few years, digital cameras have become quite good at suppressing noise. One of the more noise resistant cameras shooting at ISO 1600 or 3200 will create images that look far cleaner and more detailed than film at this same ISO rating. Digital camera enthusiasts are now fascinated as to how much cleaner they can become, and improvement occurs with each new generation. That's why reviews spend so much time talking about noise.</p>

 

<blockquote>

<p>My issues is that I don't HAVE any of the above, so the comparison is vague</p>

</blockquote>

<p>Yes, but if you shoot high ISO film, your images will exhibit very obvious grain.</p>

 

<blockquote>

<p>and I don't zoom way in on my film scans from an inexpensive Epson V500 scanner, so the shots don't have a point of reference.</p>

</blockquote>

<p>Well, that's up to you. Without zooming in, comparisons in the reviews cannot be made convincingly.</p>

 

<blockquote>

<p>It's all about high ISO noise for most these reviews.</p>

</blockquote>

<p>No, it's not ALL about high ISO noise. That's just something that digital cameras tend to handle very, very well, so the reviewers talk about it. And there are some cameras that DON'T do as well. The Sony A900 is an example. So the reviewer needs to communicate whether a particular camera has a noise issue. Typically, these days, they don't, but there are always exceptions.</p>

 

<blockquote>

<p>The case in point, for me , is a review at DPReview with the new Nikon D7000 and the Canon D60. All I could say is, " Yes. They both got noisier as the ISO went up. " But, how does that noise compare to real world use?</p>

</blockquote>

<p>The point is, how grainy would a film image look at ISO 6400? Digital is going to beat it many times other for cleanness and detail.</p>

 

<blockquote>

<p>How big could you enlarge before it's an issue? How does that change as you may need to crank the ISO up for odd situations ?</p>

</blockquote>

<p>It may not be a problem, because noise reduction software can be very effective at minimizing the effect of noise, and cameras like the D3S have very little noise in the first place.</p>

 

<blockquote>

<p>What I am looking for, and some of you got it, was how does this compare to good film scans, or even medium format stuff.</p>

</blockquote>

<p>I've had a lot of film scanned. Drum scans are expensive, take time, and are far from perfect. Digital camera files are immediately available, paid for in the price of the camera, require no development costs, and can be very sharp if you shoot them and process them well. 20+ MP cameras are at least as detailed as any medium format film scans that I've ever seen.</p>

<p>Further, digital cameras have MANY other advantages - instant review of composition, exposure, and the effect of lighting. Instant feedback on the effect of strobes. Automatic and/or manual color correction in camera. Easy compositing techniques.</p>

 

<blockquote>

<p>I know that if a good photographer used a medium format camera, they could blow up to 16 X 20 easily. It would sharp and vivid. What ISO on a crop body camera, like a D7000 ( or D300 ) could I still do the same ?</p>

</blockquote>

<p>All of the same technical factors apply: tripod quality, external vibrations, focusing accuracy, lens sharpness. I have seen super sharp 13x19 inch prints from an 8 MP crop sensor camera. My 21 MP camera prints with brilliant sharpness at 20x30, and it's very detailed at 24x36.</p>

 

<blockquote>

<p>If the progress of digital bodies is to the point that the noise is so much better than even MF film, and large prints are not a problem, why is noise the camera test reference in so many tests </p>

</blockquote>

<p>Again, because noise handling is so amazingly impressive with today's digital cameras. I suppose that everyone is curious to see how much farther each new model can go into the high ISO universe.</p>

<p> </p>

[matt_needham]

<p>No amount of reading what different photographers' experiences are will amount to a hill of beans compared to seeing large prints hanging side by side. Do it. It's easy. Then you know for sure. Personal testing also takes into account all the variables that don't get talked about when only comparing gear specs. </p>

<p> </p>

[StuartMoxham]

<p>ella</p>

[StuartMoxham]

<p>There are a lot of sample images on dpreview that you can download to make printed comparisons. Give a few images a try and see what you think.</p>

[hjoseph7]

<p>Digital is definately cleaner since there is no grain to worry about, however for large reproductions, some photographers still stick to medium format and large format film. They say it holds up better under enlargement than digital.</p>

[Ed_Ingold]

<p>John,</p>

<p>Film doesn't have noise, it has grain. Film scans have noise, but only where the scanner is unable to penetrate the murk, which happens with some reversal film (notably Velvia), but never with negative color film. The maximum density of negative color film is well within the capability of either flatbed or film scanners.</p>

<p>Digital images have noise, which occurs more the higher the ISO setting or if you attempt to lighten deep shadows. Digital images also have pixels, which become visible if they are larger than 4-8 pixels/mm in a finished print. However you can resample a digital image by a factor of 2 or more to allow greater enlargement. It doesn't increase the resolution, but improves the appearance. If done well, the results have no noticeable digital artifacts.</p>

<p>You ask how a digital camera compares to medium format film. I took a series of photos about a year ago specifically to compare my Hasselblad and Ektar 100 film with a D2x and an Hasselblad CFV-16 digital back. The CFV comparisions are posted in my gallery in Photo.net. The following was not published heretofore.</p>

<p>The upper shot was taken with an Hasselblad 205TCC with a CF80/2.8 lens, using Ektar 100 negative film, scanned at 4000 ppi on a Nikon LS-8000 scanner. The 100% crop is at the original resolution of 8500x8500 pixels. You can just see the grain, but I would not consider it obvious nor objectionable. The equivalent print size at 300 ppi would be 28x28 inches, without resampling (you can resample scans too).</p>

<p>The lower shot was taken with a Nikon D2x using a 28-70/2.8 AFS lens - one of Nikon's best, also with a substantial tripod. For comparision, the original image was resampled on the long axis (4288 pixels) to 8500 pixels. The source of the 100% crop is outlined in red on each original image. The equivalent print size would be 18x28 inches.</p>

<p>Is this good enough? That's a matter of personal opinion, the subject matter and the manner of display. I am making my best effort to present an accurate comparison of a practical subject which represents a wide dynamic range with many levels of detail. Notice that the highlights are <strong>not </strong>blown in the D2x photo, despite bright sunlight on polished metal.</p><div></div>

[Ed_Ingold]

<p>As a postlogue, it should be obvious why I didn't include 35mm film for comparision (clue: twice the grain and half the resolution). I haven't been tempted to use 35mm film since my first digital camera, a Nikon D1x, acquired in 2003. Were it not for a digital back, I'm afraid my Hasselblad would be sitting on the shelf too. That's my opinion (and choice), of course, and YMMV.</p>

[Ed_Ingold]

<p>Another point... Assuming your monitor has a nominal resolution of 96 ppi, the 100% crop is equivalent to looking at a 28" print, at 300 ppi, with a 3x loupe. Those who complain about the "plastic" look of digital images wouldn't like prints from medium format film either.</p>

[johnw63]

<p>Thanks a lot Edward.</p>

<p>From my view, the scanned Hassy shot has more noise. </p>

<p> </p>

[johnw63]

<p>Dan,</p>

<p>Thanks for taking the time. I knew about most of the sources of noise, and that it was often most noticeable in the darker areas. Getting an idea of how big I can go and still not have issues, at a given mega pixel and/or ISO is really helpful.</p>

<p>In the end, I wanted an idea of what to expect. If I ever get a nice enough shot to blow up that big, I would hate to find out AFTER the fact that it was too noisy and I should have tweaked the ISO and shutter speed to ensure a good result. Just knowing the limits of print size, ISO and Noise is all I really was after. I sure took a long post in the Nikon forum and this one to get a straight idea.</p>

<p> </p>

[frank_menesdorfer]

<p>I don't care about sharpness at all. What a sharp or clean image worth when it's not giving you anything? Nothing.<br>

Anyway mosltly all images are copy or versus of images had been taken a hundred years ago. That is photography today. Mind you that the words most expencive photograph not even sharp or clean as you would say. :)</p>

 

[dan_south]

<p>John, there's a lot to learn. Digital photography is a different world. You're not going to learn to navigate that world effectively by reading responses to forum posts any more than you could learn how to swim or box or ride a bicycle by reading about it. You need to jump in and take the plunge. You need to take pictures, process them, make mistakes and figure out how to optimize the process for your own purposes.</p>

<p>Think about getting your hands on an inexpensive camera like a D7000 or a Rebel T2i or a Sony NEX camera and GET STARTED. If you still don't want to buy one, borrow one from a friend for a few days. Download a free trial copy of Lightroom or Aperture and try your hand at RAW processing. There's a learning curve and hands-on experience is the only way to shorten it.</p>

<p>Good luck!</p>

 

[scott_murphy_photography]

<p>Well Robert Cossar voiced his opinion, and we are all certainly entitled to our own opinion, but my opinion is that not only is his opinion very general but in fact <em><strong>very inaccurate</strong></em>. I don't know what he is using as a reference for his statements, but I can shoot my Hasselblad 500CM with Kodak Ektar 100 loaded in an A-16 back, scan the negatives at 4000 ppi in my Nikon Super Coolscan LS-800 ED and absolutely <strong><em>blow the doors off </em></strong>anything digital has to offer, except the new medium format digitals, which are outrageously expensive. 645 at 4000 ppi is equivalent to about 68 MP. The best any FX camera can muster up is 24 MP. It does not take a math whiz to figure out which can record more information. Film can generally record a greater dynamic range than digital in situations with wide EV's. It is also <em>far more tolerant </em>of overesposure than is digital. And if you are shooting medium format grain is not an issue either. But after all, it's grain, and this is photography, <em>IT'S SUPPOSED TO BE THERE</em>.</p>

<p>The same goes for the same setup hand printed on Kodak Ultra color paper. Richer, deeper blacks (on glossy paper of course), and more shadow and highlight detail.<br /><a name="00XhdD"></a></p>

[bobcossar]

<p>Don't know what to say, Scott.....the 16x20's were made and viewed by many. And the findings were as I said.<br>

Really, everyone should perform the tests that will satisfy <strong>them</strong>. In the end, facts are facts. It's often the interpretations that satisfy. Or don't.</p>

[dan_south]

<p>Scott, I admire your passion, but we don't need to turn this discussion into another film versus digital frenzy. If you say that 645 print film is equivalent to 68 MP, I'm sure that you have your reasons for arriving at this figure. However, I don't know of the existence of any 68 MP cameras, so it's a little hard to test the theory. Phase One makes a 60 MP back, but I have not had the chance to test one. Have you? I have heard from reliable sources that it exceeds that resolution of 4x5 film.</p>

<p>What I CAN compare is the many 600 MB, 16-bit, professionally prepared drum scans that I have had made of my own 6x7 chromes with the output of my 21 MP Canon 5D Mark II. I don't see ANY more detail in those massive scans than I do in the Canon's 27 MB RAW files. (I wish I COULD see more detail given how much I've paid to have these scans made!)</p>

<p>The image quality is very comparable. The Canon might even have a slight edge in some instances (perhaps due to more modern lens designs), and of course there's no extra money spent on scanning when I shoot with the Canon. That's as objective as I can be, because the shots were taken on the same tripod and head assembly by the same photographer (me) under similar conditions, and I have no axe to grind for either film or digital capture as I shoot both. If you want to pursue this in greater detail, let's perhaps begin a separate discussion.</p>

[jeff_z.]

<blockquote>

<p>If you want to pursue this in greater detail, let's perhaps begin a separate discussion.</p>

</blockquote>

<p>But really, who in their right mind cares if your latest, greatest digital wonders are somehow, someway, <em>superior</em> or not? ...What in the world are you doing with the <em>end results</em> that anyone but some kind of fanatic would ever notice?</p>

<p>As far as Mr. Cossar's test, well many people's taste in various matters differ... I've looked, and still do not see any digital work in galleries that has aesthetic qualities that would make me want to shoot digitally. On the front pages of major newspapers, I am often impressed at the work <em>on screen</em>, and it is almost certainly digital, but I can't say it's much different from well exposed and scanned film work, except perhaps for high iso images. It has its strengths I suppose, but definitely, so does film.</p>

[johnw63]

<p>Dan,</p>

<p>I know I will have a lot to learn. What I wanted to know BEFORE I handed over a good $1200 is what to expect. The noise close-ups told me very little because I had nothing to compare to. I was quite sure that the noise levels of today's cameras are better than bodies from say ... 4 years back. But, I couldn't get a feel for how that worked out on prints. Small shots on my computer screen, saved on the internet, I was pretty sure didn't tell the story. So, I needed some relationship to film enlargements. For some reason, I couldn't get anything concrete out of the Nikon forum. All I wanted was a " This camera at ISO XXX could be enlarged to such and such a size and still look great. If you tried to do that with a Nikon D40, you'd have to back the ISO down by X stops or the noise would be apparent. " That's really what I was looking for. I hoped I could get that sort of input from a few posts. That shouldn't be something I have to find out AFTER the fact, should I ? It's just a matter of specifications, not some skill I need to acquire.</p>

<p>The phrase "inexpensive" is sort of relative. I don't consider the D7000 to be inexpensive. It's more than twice the price of the most expensive camera I have ever purchased. However, I know what you mean. In the scope of professional gear, it is.</p>

<p> </p>