Is Returning to Film a Boon or a Mistake....?

<p>Kevin, your current and previous Canon and Nikon DSLRs have been capturing images at around 3000 - 4000 ppi. If you printed them at that capture resolution, they would be no bigger than postage stamps. It's exactly the same with scanned film.</p>

<p>Capture ppi = [print dpi] x [print enlargement factor]<br>

E.g. 4000 ppi = [300 dpi] x [13.3x]</p>

<p>What makes MF and LF film so rich in (potential) detail is that if (and that's THE BIG, endlessly contested IF) it can compete with digital on the amount of real detail it can capture within those 4000 ppi, it then wins because it has a lot more "i"s (inches) - a lot more multiples of 4000 p.</p>

<p>The degree to which THE BIG IF is true can be established by testing different films and different digital sensors under the same conditions - this is why I brought up the great advantage of modern 645-format SLRs which take both film and digital backs. [so too do some 6x6 and 6x7 SLRs, but the increasingly large film formats make it more a test of lens quality].</p>

<p>Tests like this seem to indicate that digital is at about a 1.5x linear advantage to film - e.g. a 37x37mm digital back compares well to 56x56mm [6x6] film in typical usage. I'll probably get flamed for daring to propagate such a value, because there are people occupying positions which are way off to either side of the shaky consensus: some would say that digital is easily twice as good (2.0x) and some would say film is every bit as good (1.0x)...and no-one can agree on what "typical usage" is either! But I favour attempts to bring some scientific method to the issue and looking at it dispassionately rather than idealogically.</p>

<p>Ray,</p>

<p>You will indeed get flamed, but not just for propagating that factor, but for blatantly failing to recognize the fact that film is not even "every bit as good", but much, much better. ;-)<br>

(It really is even after having been scanned).</p>

<p>Do bring in some scientific method to the issue, and find out yourself.</p>

<blockquote>

<p>I'll probably get flamed for daring to propagate such a value, because there are people occupying positions which are way off to either side of the shaky consensus: some would say that digital is easily twice as good (2.0x) and some would say film is every bit as good (1.0x)</p>

</blockquote>

<p>You could be right. I tend to think it's around 1:1 now as the laws of physics are constraining both mediums. Whichever figure is correct, I don't think either medium will get much better now other than in terms of increasing area. Easy with film but there are plenty of manufacturing challenges with digital sensors.</p>

<blockquote>

<p>So my question is: why is Nikon advertising 4000 ppi when, seemingly, it would never be advantageous to print at such a resolution...?</p>

</blockquote>

<p>It's quite simple. When Nikon says 4000ppi, they mean 4000 points per inch on the film. So for argument's sake if you scanned a 1'' x 1" piece of film you'd end up with a 4000 x 4000 pixel digital file. If you printed that file at 300dpi, your print would be 4000/300 x 4000/300 or 13.3" x 13.3".</p>

<p><em>"You could be right. I tend to think it's around 1:1 now as the laws of physics are constraining both mediums. Whichever figure is correct, I don't think either medium will get much better now other than in terms of increasing area. Easy with film but there are plenty of manufacturing challenges with digital sensors."</em></p>

<p>Indeed.<br />One thing, of course, is that the size advantage has always been, and still is, on the film side.</p>

<p>Another thing is that not the laws of physics are what are limiting digital (nor film).</p>

<p>That small sensor may boast 4000 'pixels' per inch.<br>

But what that doesn't say is that there is a Bayer (or similar pattern) filter, reducing the true resolution to half that.<br />What it also doesn't say is that there is a soft focus filter in front of that Bayer pattern and the sensor that is reducing resolution even further.</p>

<p>Sensor's pixel density figures are not the thing to base comparisons on.<br />Or if you do, you perhaps should also take unsharp masking, interpolation, and other tricks that can also be performed on scanned film, into consideration.</p>

<p>In short: comparisons of film and digital on the net have still a long way to go before they even begin to find a way in which like can be compared to like.<br />It's such a great muddle that we have gone from "digital can't deliver what 35 mm film does" a few years ago to "digital easily outperforms large format film" today, without any basis for such a change in view in advances in technology or in fact.</p>

<p>Scott, If someone needs to post a sign beside their images dictating how close you are permitted to get, then the quality is an issue. And no, the difference was easily visible at 2 feet. More so in close at about 18".</p>

<p>The 10D is by no means a match for 6x7 film at that size. Nothing I have seen for the better part of a decade has changed that.</p>

<p>FWIW, Bayer filtering doesn't apply to scanned film, which is what the prior few posts were discussing. Don't confuse an anti-aliasing filter with a Bayer filter. One is an optical low-pass filter (AA) and the other is a software filter. Save for the Foveon sensor, one-shot capture sensor arrays all use Bayer or another color interpolation scheme. But this Bayer filtering argument no longer holds much water because even with it we have been at the point for some time now where the resolving power of digital sensors is equal to or better than film at a 1:1 basis (particularly once film has been scanned, and those second generation losses are factored in). Too, not all digital arrays have optical AA filters (Leica, Pentax the new 645D, and most MF backs don't have them).<br>

The remaining resolution advantage to film is derived from it being easy to scale up in size, cheaply. When sensors can be made at LF sheet film sizes with the similar pixel density and quality control as is commonplace with today's DSLR-sized sensors , the argument will be settled. (There may still be compelling reasons to use film even then, such as cost, working style, and tradition, but ultimate resolution will not be one of the more compelling reasons. I'll like be among those using both film and digital past that point.) </p>

<p>Kevin, <a href="http://www.drumscanning.com/rates.html">this page</a> illustrates the relationship between scanning various negative sizes at different resolutions and then printing them out at 300dpi.</p>

<p>Read the text and not just the colored squares if the relationship isn't clear, but it's pretty straightforward. For example, a 6000x9000-pixel image would print at 20x30" at 300dpi (because 6000/300 = 20 and 9000/300 = 30). I'll not get into ppi/dpi/spi distinctions here; they're not super-important for this discussion.</p>

<p>The Nikon 9000 can scan film in one pass up to 56mm x 83.5mm, I believe (that's considered "6x9" for many camera manufacturers). A drum scan - as this place specializes in - can usually pull more information out of a piece of film and can scan larger film in one pass.</p>

<p>There also are places that can do Nikon 9000 scans quite affordably for you: as one example, <a href="http://www.digmypics.com/pricing.aspx">this company</a> charges $8 for 4000dpi scans of 120 film up to 6x9. (I've never used them and cannot address quality etc.)</p>

<p>Kevin, you have received a lot of good and interesting responses. I hope I'm not repeating anything previously said, but what historically has made 4x5 the "standard" for landscape photography is not just the size of the film but also the camera movements that can allow the subject to be altered. Most important is front tilt, for adjusting the plane of sharpness, and front rise, to allow foreground trees to be straight. Yes, some of these adjustments can be made in photoshop but, in my opinion, doing them in camera gives a much better result. For DSLRs and especially Canon, the TS lenses are simply fantastic. If you can afford a larger sensor digital back, then there are several medium format cameras which have movements (Fuji GX680, Linhof Technika, Mamiya RZ67 with adapter) and offer the versatility of film plus digital capture. Scanning 4x5 film is readily done, but requires some effort. I picked up a used Polaroid Sprintscan 45 and modified the film carrier for wet (liquid) mounted scans. Check out ScanScience and BetterScan for more info (I'm not affiliated with either). My total investment was under $200 and I get very good quality scans from 4x5 and 6x7. Large Format Photography website has a page devoted to scanner comparisons that shows how close you can get to a drum scanner with the less expensive alternatives. I agree with the previous posters that both film and digital have their place depending on the subject matter. I use Canon 5D MkII, Leica M8, Linhof 6x7, 4x5 and a 5x7 pinhole camera. Having options is a good thing. Good Luck. Rick.</p>

<p>Ivan,</p>

<p>No, digital sensors, with their Bayer (or other) colour filters, and anti aliassing filters, are still well behind film. Even well behind scanned film, by a factor of (give and take) 4.</p>

<p>Pixel density on sensors now equals scan resolution of the better scanners. Per inch, so the bigger film still has a clear advantage.</p>

<p>But (and this is important to remember) that is without taking the Bayer pattern and soft focus filters into account.<br />Do that, and even 35 mm scanned film will have a clear advantage over full frame 35 mm digital.<br>

<br />Now don't scan film, but print it optically, and the gap widens.</p>

<p><br />Re that "laws of physics" thing: neither film not digital are limited by the laws of physics.<br />Yes, increading pit density on sensors is running into problems. But those problems are still far from being caused by limits nature sets. There is still room for improvement.<br />Same goes for film, by the way.</p>

<blockquote>

<p ><a href="../photodb/user?user_id=528518">Dave Luttmann</a> <a href="../member-status-icons"></a>, May 21, 2010; 08:52 a.m.</p>

</blockquote>

 

<blockquote>

<p>Scott, If someone needs to post a sign beside their images dictating how close you are permitted to get, then the quality is an issue. And no, the difference was easily visible at 2 feet. More so in close at about 18".<br>

The 10D is by no means a match for 6x7 film at that size. Nothing I have seen for the better part of a decade has changed that.</p>

</blockquote>

<p>So I set up a test of two prints, one at 125ppi and the other at 720ppi, both of them sized to 4x6 inches so I could fit both on the same piece of paper.<br>

Here are the links to the photos<br>

<a href=" photo</a><br>

<a href=" photo</a><br>

When viewed at 24 inches I can't tell any differrenace between the two.</p>

 

 

<p>At 18 inches with a lot of careful looking I can just begin to see a difference, but it is so small that there would really be nothing to choose between the two.</p>

 

 

<p>At 12 inches the 720ppi print is clearly sharper but not to the point that the 125 look horrible.</p>

<p>At 10 inches the 125ppi print is looking pretty soft.<br>

</p>

 

<p>As it happens I like very high resolution photos and mine tend to have far more resolution then what a 6x7 camera could get achieve. But when it comes to others looking at prints most of the time this extra resolution is simply a waste, people just don't look close enough at large prints for the higher resolution at add impact. What makes a large print look sharp is not the detail so much but other factors, such as lighting, contrast and content.</p>

 

 

<p>I have made fair number of 12x18 inch prints straight from an 8MP camera and also from stitched images that have 20Mp of super clean sharp pixels, nobody but me notices the difference.</p>

<p>In a lot of ways I wish resolution mattered more in a print, I have a large collection of images that run between 100 and 200mp, but in the end I have to admit that the resolution adds very little over the same photo at 8MP or even 6MP.</p>

 

 

 

<blockquote>

<p>For DSLRs and especially Canon, the TS lenses are simply fantastic.</p>

 

</blockquote>

<p>Hi Rick,</p>

<p>Do you own one...? Or does anyone else responding to this OP have any experience with Canon's new tilt-shift lenses...?<br /><br />I must admit, I've been thinking very hard about possibly heading in this direction with my current 1Ds Mark III set up.</p>

<p>Kevin, I have the new 24mm TS/E II. It is a phenomenal lens, and I understand the 17mm is also (I haven't used the 17).</p>

<p>But unless you're going to use it for stitching, a TS lens doesn't really do anything to resolve your initial dilemma about printing larger.</p>

<p>You're simply going to need more pixels if you want to print big, whether you get them [1] from stitching several of your 1DsIII exposures together, [2] from a higher-resolution sensor (like the Pentax 645D), or [3] from shooting onto a good-sized piece of film (MF or larger) that you then scan.</p>

<p>Those are your three basic options.</p>

<blockquote>

<p>But unless you're going to use it for stitching, a TS lens doesn't really do anything to resolve your initial dilemma about printing larger.</p>

 

</blockquote>

<p>Yes, I'm definitely leaning towards stitching.....</p>

<p>On this note, any recommendations ongoing towards ptgui, CS5, or some other piece of software for the most effective stitching experience...?<br /><br />(I'm presently using Photoshop CS3... maybe I already mentioned that in a previous post...)<br /><br />Thanks again for all input!</p>

<p>

<p>I use PTGui for stitching and find it gives me great control over the whole process. One of the good things about stitching is it gives you the same control for perspective that a LF camera with movements does. For example in this photo the vertical lines are not converging, like they would be if this was taken with a normal camera pointing up.</p>

<br>

<a title="pan1c 08-22-07 by KonaScott, on Flickr" href=" src="http://farm3.static.flickr.com/2557/4296968591_eb3dbe93b0.jpg" alt="pan1c 08-22-07" width="500" height="382" /></a><br>

For a larger view click <a href="

<p>But as much as I like very high resolution photos I think you should take the time to print some of your nicer photos large and see what people think, say 20x30 inches, which cost around $10 at many places.</p>

<p>I think you may be surprised at both how good the prints look and how people don't seem to view large prints closely.</p>

<blockquote>

<p>But as much as I like very high resolution photos I think you should take the time to print some of your nicer photos large and see what people think, say 20x30 inches, which cost around $10 at many places.</p>

</blockquote>

<p>Hi Scott,<br /><br />Just a few days ago I went ahead and sent off an order for one of the prints to Mpix, printed at 16 x 24" (this was at just less than 240 PPI).<br /><br />I must admit that I was thrilled with the results. I have shown them to family members and friends alike and they are all raving about the results.<br /><br />So I may go for 20" x 30" on the next test print, but I'm sure this will place the PPI well below 200.</p>

<p>I think what you will find with a 20x30 inch print is that people just don't get that close to it. <br>

<br>

I use to be of the opinion that prints show all have at least 300ppi, but as I have watch people viewing larger prints I realize that this is way over kill for large prints. Even a good 12MP image is going to make a great looking 20x30 inch print, for that matter even a 6MP image will be good enough to hang on a wall. Sure a 6MP is going to look soft if you view it from a foot or less, but that is not how most people like to view prints. <br>

<br>

 

<p>

<p>

<p>

<p>

<p>As photographers it is easy to get too wrapped up about resolution and not enough about the content of the photo.

<p>

<p>

 

</p>

</p>

</p>

</p>

</p>

 

<p>

<p>

<p>

 

</p>

</p>

</p>

</p>

</p>

<br>

 

<p>

 

</p>

</p>

<p>Scott...there is a big difference between a native 6mp file...and a 6mp file from a downsampled higher rez file. I've got shots between the 10D and 5D2...no problem telling the difference at 16x24 from 2 feet.</p>

<p>I agree with the content issue. That said, I don't agree with the viewing distance. In galleries, people start from a larger distance...then they move themselves closer to immerse themselves in the detail. The 6 or 12 mp sensor at 20x30 will simply not have any fine detail.</p>

<p>If you want to improve your photographic vision, then stick with digital. Print size isn't everything - in fact it's nothing, as you can see from examining the tiny pictures here on Photo.net. Those that are great pictures remain great pictures; and those that aren't wouldn't be any better if they were shot on 20x24 film and blown up to the size of a house.</p>

<p>Digital gives you the great advantage of nearly instant feedback on whether you've captured a great image or a mediocre one. It also gives you a second chance if the exposure, lighting or camera position aren't quite right. With film there's an inevitable delay and the interruption and complication of yet another process to contend with. Just because something is more complex or more difficult to achieve doesn't necessarily make the end result any better.</p>

<p>Of course, if you're happy with the composition and visual impact of <em>every</em> picture you take at the moment, then you might want to consider the extra complexity that using film will add. But until that happy day, then I'd just push the boundaries of what you can do with digital and what digital can do for you. I think you'll find that those boundaries are quite expansive.</p>

<p>As a footnote to the above. If your printer can't get a decent 20x16 from a 21Mp file, then change your printer! I'm getting A3 sized prints (11x16) from a 12Mp D700, at a quality that puts most wet-darkroom prints from MF film to shame. I can see no sign of pixelation at this size, and the colour quality and overall "cleanliness" of the prints is superior to all but the most meticulously produced film shots.</p>

<p>Another example: I was in a restaurant the other day, and was impressed by some large sized prints (about 20x30 inches) hanging on the walls. I approached to have a closer look and the quality completely fell apart when viewed close up - fuzzy and obviously taken with an amateur camera. However, once I stepped back to normal viewing distance they regained their power to captivate, simply because they were great pictures, and I found myself able to easily forgive their lack of real technical quality. Whether they were shot on digital or film I really didn't care.</p>

<blockquote>

<p ><em><a href="../photodb/user?user_id=528518">Dave Luttmann</a> , May 21, 2010; 10:39 p.m.</em></p>

<p ><em>Scott...there is a big difference between a native 6mp file...and a 6mp file from a downsampled higher rez file. I've got shots between the 10D and 5D2...no problem telling the difference at 16x24 from 2 feet.</em></p>

</blockquote>

<p >Could you post these shots?, or at least enough of each for a test print?</p>

<p > </p>

<p >At 10 inches 300ppi is about the limit of what people can see, a photo from a 10D printed at 300ppi (6.7 x 10 inchs) is going to look about as sharp as it gets, when view from 10 inches. Now back up to 24 inches, scaling the image size up and we get a 16 x 24 inch print. </p>

<p > </p>

<p >Is it posible that your 10D photos are soft at the pixel level?</p>

<p > </p>

<p > </p>

<p > </p>

<p > </p>

<p>Scott, here is a test sheet you can print to evaluate the prints at different sizes:</p>

<p>http://www.photo.net/film-and-processing-forum/00WErk</p>

<p>Let us know what you conclude.</p>

<p>Thanks again for all of the wonderful responses on this forum! They are truly appreciated.<br /><br />Okay, setting aside the questions surrounding printing (i.e. print size), I did have the following image-sensor and optics question as it pertains to the difference between the sensors and optics in a digital camera versus the optics in a digital scanner:<br /><br />Is there any difference… or perhaps to state it more clearly, is there an advantage one way or the other… with regard to the two separate processes involved in the film-based landscape-photography world versus the digital-camera-based landscape-photography realm <strong><em>as it pertains to the image sensor and the optics involved?</em></strong><br /><br />Meaning, if you are using the new Mamiya 7 II film-based camera (6 x 7 cm, 55 x 69.5mm), for example, you of course establish the image on film and then (if you are working within Photoshop or some other photo manipulating software) you are utilizing a scanner (let's say the Nikon Coolscan 9000) and its image sensor and digital optics to take that film-based image and convert it to a computerized, pixelated form.<br /><br />Or conversely, if you are using the new Leica S2 digital camera, you would be relying upon the built-in image sensor (part of the body) and optics (associated with the lenses themselves).<br /><br />But both processes rely upon image sensors and associated optics (or at least somewhat common forms of technology and processes, from what I understand).<br /><br />However, with the Leica S2 scenario, one is using an image sensor and associated optics to take the image directly from "reality," as it were. It is a "direct from the scene" sort of situation, in my mind.<br /><br />With regard to scanning, you are first shooting the scene onto a piece of film, and then you are utilizing the image sensor and optics associated with the scanner to create the computerized/pixelated image.<br /><br />My point being: there have been several individuals who have contributed to this forum who are advocating film (and I deeply appreciate their input here!). They are noting that the overall results produced with film are better when compared with a digital camera's image sensor and optics. And yet, if I'm utilizing a scanner to drop the film-based image into Photoshop, I still need to go through a very similar image-sensor and associated optics process, am I not? (Meaning, something very similar to what you would originally get "on the scene" if you were utilizing a digital camera's image sensor and optics.)<br /><br />So with respect to the pixel-for-pixel quality, as it were, what's the advantage in using a scanner's hardware versus a digital camera's hardware to computerized/pixelate the image...?<br /><br />Yes, I do realize the aspect of the differences in overall image size and the total number of pixels; but within each specific inch of the image sensor's data-gathering capability versus the scanner's, is there a difference...?<br /><br />Or am I thinking about this whole situation/scenario incorrectly...?</p>

<p>Again, I'm not thinking about the subject of the overall size of the file, the print size, etc. Rather, I'm getting back to the heart of the matter between shooting originally with film or shooting originally with a digital camera's image sensor.<br /><br />Thanks again!</p>

<p>You are thinking correctly.<br>

And yes, there is a difference.</p>

<p>It already starts during capture. (Forgetting about the size advantage film has for now)<br>

Not hindered by Bayer and anti-aliasing (a fancy name for soft focus) filters, there is lots more captured by film then there is by a digital sensor.<br>

All that extra info can be retrieved by the scanner, which is not hindered by those filters, not by restrictions set to sensor size.<br>

But what the sensor didn't see, is lost forever.</p>

<p>Dynamic range? Film (negative) is no slouch in that department.<br>

A scanner can be tweaked much more, and much more effectively than a camera's sensor.</p>

<p>Film and scanning have disadvantages too.<br>

It takes a lot of time.<br>

And film has that thingy called grain, which is faithfully captured by the scanner, and which makes a digitally captured image look cleaner.</p>

<p> </p>

<p>Hi Kevin, if I read your latest question correctly, you're asking "Isn't digital more of a direct-to-file process than film (which has to be scanned, effectively requiring the photographer to work with a 'second-generation' image), and thus digital would be superior?"</p>

<p>To my mind (I can't speak for others), you're right in the sense that in a digital vs. film faceoff when the <strong>recording surface area and ISO are equal,</strong> most (but by no means all) photographers in 2010 would opt for the digital image.</p>

<p>In other words, if it's a choice between a 21mp, 24x36mm sensor and a 24x36mm piece of film, both shot at, for example, ISO 200, the resulting digital file will generally be judged superior. Same with a 30x45mm digital sensor vs. a 30x45mm piece of film. (Erwin Puts captured a tiny bit more on ISO25 b&w 35mm film than a Leica M9 sensor, but in practical use most photographers aren't going to shoot that film.)</p>

<p>But most of those in this thread who are suggesting "film" are not claiming an even-for-even advantage for film. They're saying that a 6x7 or larger piece of film, well scanned, can often best a smaller (e.g., 24x36-sensor) digital image, just as a sheet film image can often best a (smaller) medium-format digital image.</p>

<p>Thus for single-image capture (no stitching), <em>especially with regard to convenience,</em> a medium-format digital camera might be seen as optimal. If you can swing the cost of medium-format digital, there's a lot to be said for it (luminous-landscape.com and getdpi.com probably have more discussions of the $25,000-and-up MF digital options than anywhere else).</p>

<p>But the price of medium-format digital is pretty steep, and not very many non-pros can afford it. Therefore a lot of photographers who can't spend more than the price of a new car on a MF digital camera choose instead to shoot medium- or large-format film, which has a much lower entry price (think "hundreds of dollars" instead of "tens of thousands"), and then they do high-resolution scans of only their best film images (not of all of them).</p>