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

<p>Thanks Q.G. and Ralph for the info!</p>

<p>A few things need to be put straight though.</p>

<p>Comparisons of digital to film are almost without fail performed by blowing up the digital file to where you just can't see the individual pictures, and then enlarging the film image to match the size of the subject.<br>

I.e. comparisons are always made of digital at its limit to film, without ever wondering if film may be strecthed way beyond that before it reaches its limit. And it can.</p>

<p>But if, say, an 5x7" printed from 35 mm full frame digital looks as good as a 5x7" printed from 35 mm film, that apparently means that 35 mm full frame digital is as good as 35 mm film. Period.<br>

Which is, of course, a nonsensical way of thinking.</p>

<p>So yes, Ralph, there is an "even-for-even advantage" for film. But it would appear that noone of the let's-compare-my-digital-to-that-old-fashioned-film camp ever bothers to investigate it.</p>

<p>Another thing: high quality digital is not cheap, no.<br>

But that is not why many people use film. Many people still using film do so because they like what film has to offer over digital.<br>

Like it so much that they are willing to put up with the messy fluids bit, willing to put up - when going down the hybrid road - with the lengthy scan process.</p>

<p>In the above: "to where you just can't see the individual pictures" should of course read "to where you just can't see the individual pixels"</p>

<blockquote>

<p ><a href="../photodb/user?user_id=282122">Q.G. de Bakker</a> <a href="../member-status-icons"><img title="Hero" src="http://static.photo.net/v3graphics/member-status-icons/hero.gif" alt="" /><img title="Frequent poster" src="http://static.photo.net/v3graphics/member-status-icons/1roll.gif" alt="" /></a>, May 23, 2010; 06:31 a.m.</p>

<p >A few things need to be put straight though.</p>

</blockquote>

 

<blockquote>

<p>Comparisons of digital to film are almost without fail performed by blowing up the digital file to where you just can't see the individual pictures, and then enlarging the film image to match the size of the subject.<br />I.e. comparisons are always made of digital at its limit to film, without ever wondering if film may be strecthed way beyond that before it reaches its limit. And it can.</p>

</blockquote>

 

<p>The fact is that digital can go to just about any resolution you might want. I am a resolution junky, I like high-resolution photos. I am also fairly near nearsighted, so I have a bad habit of viewing print with my glasses off and my noise right in the print. What as this is find for small prints, say 8x12 inches, for larger prints viewing the close you loose the total image. In watching other view large print what is see is them going close enough to comfortable see the whole image at once, i.e. you make the print twice as large and normal people will view it from twice as far away.</p>

<p>As an example of very high resolution is the photo below, it is the same number of pixels as if you scanned a 6x7 MF frame at 4000ppi<br>

<a href="

It did not cost much to take and it only took 3 minutes to take the photos that it is stitched from. That photo has far more resolution then every be needed in any print. You could print that at 6 x 5 feet, stand 12 foot away and it would still look very sharp. But who is going to stand 1 foot away from a print that large?</p>

 

 

<p>When I argue that super high resolutions are not needed for large prints it is not because high resolution is hard to achieve with digital, it is that the resolution is simply not needed.</p>

 

 

<p>BTW one problem going after super high resolution with either film or digital is you end up with a smaller depth of field. Whereas this is fine for some photos others suffer from it. I have been backing off my resolution, not because getting the high resolution is hard, it's not, but because I like more DOF over the ultimate in resolution.</p>

 

<p>Film is great stuff, and I'm shooting as much of it than ever. Just not buying for a moment Q.G.s argument that it's 4X better in resolution than digital sensors at a 1:1 basis. Millions of professionals would scoff at digital if this were so. In actuality, the majority have abandoned 35mm film en masse (I count myself among them).<br>

Those professionals continuing to use film in larger formats, are largely doing so because it scales up far, far less expensively at ultra-high resolutions. Or because one-shot sensors are not available beyond 645 in size, yet. Or they’re simply continuing to use it for other reasons like film’s great latitude and dynamic range. Or they’re using it to achieve a particular look, or find it simple and uncomplicated, etc. <br>

Oft overlooked in these discussions is that while film can indeed record amazing amounts of resolution, the subject gets complicated beyond the point of capture because to do anything with a film image capture typically requires another step. So, unless viewing a slide directly on a lightbox, or making Polaroids or tin-types or Daguerreotypes which are direct and the original image themselves, there will be additional system losses due to optical degradation.<br>

Transmissive light must pass through processed film (diffusion) either to print it traditionally or for the image to be converted from analog to digital with scanning. (Yes, even with drum scanning, there are transmissive optical losses). By the time the image is put to use, it is at minimum a second generation image. Digital captures when copied aren't subject to these generational losses.<br>

I also have an inkling where the 4X number might be coming from. To not add additional system losses when scanning original film, over-sampling is necessary. Nyquist Theorum posits that a 2X over-sampling in a linear dimension is required to capture the native resolution. Assuming this is true, to match resolution between a digital capture and film, a 4X larger file is created for the scanned film image. Same resolution.</p>

<p>Scot,</p>

<p>The stitching thing is of course neither here nor there. You can stitch film, and scanned film too. So there's no point of bringing that up.</p>

<p>Ivan,</p>

<p>No: professionals have embraced digital when all it could do was produce wax faced portraits, and the like, i.e. there was and is a strong attraction for professionals that has nothing to do with image quality.<br>

Professionals who value quality above all, or have clients that value quality, still use film in large numbers. (They are also the ones that did very little with miniature format and the small prints that could sustain).</p>

<p>And no: that 4x thingy has nothing to do with what you think. That is, not in the way you think.<br>

It's quite simple. Sensor technology is at a point where pixel density on sensors happens to be the same as scan density. So on a 1:1 comparison you would be forgiven to think there is equality.<br>

But only for a brief moment: Bayer patters and soft focus filters employed in digital cameras limit what the sensor gets to see. A factor 4, compared to what could have been if such devices weren't necessary, is a conservative estimate of the degradation caused.<br>

You can forget about that Nyquist nonsense when scanning film is concerned. You're not overlaying a pattern with another pattern. The pixel density of the scan sensor gets used to the full.</p>

<blockquote>

<p>Bayer patters and soft focus filters employed in digital cameras limit what the sensor gets to see.</p>

</blockquote>

<p>Hi Q.G.,</p>

<p>It was noted to me a couple of years ago that there were a select few number of digital camera manufacturers (I believe Leica was one of them) that did not have anti-aliasing and other automatic filtering mechanisms within their digital sensors; Bayer patters may have been another one of these aspects that were not included as well.<br /><br />The individual who noted this was a longtime professional wedding photographer, one who used Leica equipment primarily. He said he chose the R system specifically for this reason.<br /><br />Do you know if this is true...?</p>

<p>I went ahead and downloaded the tech spec sheet on the new S2, but I cannot see any information regarding these aspects (maybe I'm blind).</p>

<p>Kevin,</p>

<p>The only other way of capturing colour would be the Carver Mead/Foveon trick, of putting the three wells below each other, instead of in a pattern next to each other.<br>

Else, a Bayer pattern (or similar device) is needed, and thus used.</p>

<p>A few, more expensive digital devices do not have anti-aliasing filters, yes.</p>

<p>The Leica S2 has no anti-aliasing filter (it uses software in post-processing to do what would be necessary). The Kodak sensor has a Bayer pattern.</p>

 

<p>The anti-aliasing filter is necessary to stop the sensor from getting moiré artefacts in the image when high frequency data in the image gets higher than the frequency (pixel-pitch) of the sensor.<br>

The filter effectively blocks any high - low contrast patterns close to the pixel pitch of the sensor at the expense of some crispnes in the image.</p>

 

<p>What has been only lightly touched on in this thread is the look of film vs. digital prints, and left uncontroverted is whether "normal people" go up to a large print or are content to view it from afar, whether film or digital. Normal? Who's that?<br>

Granted, Photonet may consist largely of more or less well-adjusted, regular folks; but whether "our" own habits as photographers will differ widely from a presumed "normal" mode of enjoying photographic images is questionable in my own observations. And what about the customers' viewing habits, people who are not just casually interested? And not only on first viewing, but later on if they feel like immersing themselves in the image more deeply, as they would after paying for a print. Speaking only of myself, I know that I first look at an image from wherever I may first encounter it whether from close up or afar, and if it's interesting, go in for a closer view or viewings later on. This is most especially true in the case of large prints, because viewing it close up helps to "enter" the space being represented on the flat surface of the print. I'm not referring to technicalities of resolution, but to what it looks like subjectively. Sharpness is important, but I don't feel that is necessarily all there may be to it.<br>

More having to do with the look of the print, and maybe more to the point of the subject of the film vs. digital print: do inkjet printers render blacks with the saturation of photographic paper, in continuous tone? I don't know if the best of them do, so I'm interested in what opinions there are on that, and whether it may be considered a decisive factor for those who like Kevin may wish to explore film photography. </p>

<p>Follow-up question regarding interpolation:</p>

<p>From a few of the reviews I have read concerning Genuine Fractals 6 and Qimage (among others), it appears that some of the more recent interpolation programs are becoming extraordinarily good at enlarging digital images.</p>

<p>Here are my 2 follow-up questions:</p>

<p>Given that most of the reviews of Genuine Fractals focus upon smaller digital files being enlarged by 5 or even 10 times, I'm wondering if there is a difference in the overall quality of the enlarged image when it is enlarged by only 50% - 100%?</p>

<p>Meaning, if I'm shooting with the 1Ds Mark III and I'm already able to print at 16" x 24" at around 230 PPI, if I want to print at 20" x 30" (and yet increase the resolution to maybe 250 PPI) -- because this enlargement is only in the range of about 50% (I haven't gotten out my calculator, so I'm just going on a mental image right now), is the enlargement (using Genuine Fractals or Photoshop or Qimage) going to be much less "problematic" as opposed to trying to enlarge something by 500% or 1000%...?</p>

<p>I hope that makes sense.</p>

<p>Does anyone know if Photoshop CS5 has improved their underlying algorithm associated with enlarging (or even reducing) photo sizes? Meaning, because I'm upgrading to CS5 next week, I'm wondering if there is something new in CS5 in relation to a better underlying algorithm for producing enlarged digital files...?</p>

<p>Thanks in advance for any and all responses!</p>

 

<p>The problem with interpolation is that it is making up data.<br>

No algorithm can do that and produce convincing images beyond an enlargment factor of about 2.</p>

<p>The result will be the same as those we had when the top pixel count number was about 2 or 3 MP, yet people already wanted to use such digital machines for magazine page size images: a lack of detail made everything look like it was made of plastic.</p>

<p>In short: if you want to have 'large' images that still look good, get a camera that produces such large images, and don't put your hopes in post-processing.</p>