Film vs Digital - Color Rendition

[mauro_franic]

<p>The Coolscan 9000, in my opinion was close to perfect for $2,000.</p>

<p>It just needed a few more years of R&D and upgrades and it could have been scanning 6,000 dpi in half the time with roll feed. Timing was against it.</p>

[rishij]

<p>The Minolta DSE 5400 is really a feat of engineering. Just the optical module itself is beautifully done, from the collimators to the strip of IR LEDs... 'grain dissolver' is weak at best though, & depth of focus is just as dismal as the LS-4000 though. Some good ideas... just not good enough.</p>

<p>For the record though, there's more of a dropoff in contrast at the limiting resolutions with the Minolta in comparison to the Imacon, meaning that the Minolta could still benefit from an increase in resolving power. So could the Imacon, given the contrast I see with my eye on the film under a light microscope at 4.75.</p>

<p>BTW, does anyone know how the diffuser in the LS-9000 works?</p>

[mauro_franic]

<p>I didn't understand the lentil patties comment but I can give you a good offer for your 1Vs if they are in like new condition.</p>

[mauro_franic]

<p>I'm afraid to open my scanner to see since they are sort of scarce. </p>

[rishij]

<p>Scott, as far as I've read I, for one, think your input is valuable. I like dissent b/c it either re-affirms one's own opinion or forces one to look at it from a different angle & learn.</p>

<p>Good point with the 'maybe it was Nikon & Canon's plan...' :)</p>

<p>Another gripe I have with Imacons & drum scanners: because they roll the film around a drum, geometric distortions unique to each frame (depending on the exact alignment of the film) make it darn near impossible to stitch 3 bracketed shots together into a HDR because alignment software can only correct for translation & rotation, not complex geometric distortion... trust me I've tried. The same 3 frames of film scanned on a Minolta/Nikon? No problems with alignment. Just problems with edge-to-edge sharpness due to film flatness problems (unless I break out the holder I built for my Minolta).</p>

<p>Still, given the choice, most of the time I'd shoot digital b/c of the convenience & b/c of the enhanced latitude in comparison to slides (I just don't like negatives b/c of the increase in noise once you decompress the tonal range that was compressed into the limited tonal range of the negative... though maybe I should give it another try b/c of its latitude). </p>

<p>However, if there were a desktop scanner that were as high-res as an Imacon, didn't show pepper grain, held film flat for edge-to-edge sharpness (I'm talking the kind of edge-to-edge sharpness you get out of a dSLR!), & really saw into the shadows of Dmax regions of Velvia, then, yeah I'd be much more enthusiastic about shooting Velvia. Right now I shoot it more as a novelty & rely on my 5D for the heavy lifting (well that's a lie: I'm waiting for my 5D to be fixed... shorted it in the rains/waterfalls of Oregon).</p>

<p>Rishi</p>

[scott_ferris]

<p>Mauro,</p>

<p><em>"I can give you a good offer for your 1Vs if they are in like new condition."</em></p>

<p>Thanks but I couldn't, hand on heart, describe either of them as in like new condition! The low count one is close but spent a few years, even when I went mainly digital, being a fallback to familiarity backup camera. Still got all the boxes, paperwork, instructions and receipts somewhere, I think they are in a loft in England!</p>

<p>Rishi,</p>

<p>Thanks for the acceptance, and the intelligent and unemotional but authoritative input. Shame we couldn't cull the first hundred or so comments in the thread down to this last bit. Interesting point about drum scans and HDR, another factor I had not encountered having never relied on scans.</p>

<p>If we are not on exactly the same page, we are definitely on the same chapter and reading the same book. I would almost certainly buy a good reasonably priced desktop scanner that worked and that realised the true potential of my 135 film, I'd pay good money for it if it went up to my 6x9 as well. My first half a dozen or so digital weddings I took a full compliment of film bodies and stock to use as backup, I enjoyed using them side by side.</p>

[rishij]

<p>Yeah I kinda stopped reading this thread somewhere in the beginning b/c I had initially entered it under the impression that this was going to be some sort of objective test of the color capability of film vs. digital (I still intend to do that test I mentioned earlier & that you thought might be useful Scott, though in general I hear it's difficult to make profiles for negative film b/c of the complex TRC... I'd assume a LUT based profile should do the trick but maybe there aren't enough patches on an IT8 chart, I dunno). I quickly lost interest when I found out it wasn't & no one wanted to partake in the more technical talk (no offense Mauro -- it was still a fun competition for some & certainly a magnanimous offer on your part re: the prizes!). That's just me. Else I would have more knowledge about why people kept knocking you.</p>

<p>I think the opinion of someone such as yourself who's done film/darkroom in the past & now shoots digital is invaluable. Because you've tried both mediums in your hands & so your results, though not necessarily the best obtainable, are very relevant to the conversation. I say 'not necessarily the best obtainable' here b/c I'd like you to perhaps realize that perhaps you could've obtained better results with film had you scanned them in a capable-enough scanner & then post-processed them (sharpening, noise reduction, etc.). Because your previous quoted results seem to show film underperforming more than it should. For example, I've made some gorgeous 16x24 prints off of 135 Velvia... but the only way that happened was by scanning on an Imacon, doing some very selective & subtle noise reduction, & then selectively sharpening the heck out of it (using a lot of Bruce Fraser's philosophy) & then using intelligent upsampling algorithms. That's the sort of stuff you can't do optically in the dark room (AFAIK), even though I realize staying in the darkroom means you don't experience any of the resolution/contrast loss introduced by scanning. </p>

<p>Thing is, digital capture is also significantly sharpened by the RAW converters & in post-processing software. So I think there is still a space for high-quality film scans + post-processing... much like DLT mentioned.</p>

<p>Anyway, yes let's keep the conversation unemotional & civil, please. People are simply stating their experiences. That doesn't mean they're wrong or lying, & their experience is valuable information to all of us.</p>

<p>Rishi</p>

<p>P.S. Does anyone want to take a guess at what the contrast ratio between black & white might be on the printed ISO chart Mauro shot & I provided all those scans of? Just a ballpark guess even! I'm really curious as to what contrast of the chart I'm quoting when I say the Imacon scan of 135 Velvia yields 20MP...</p>

[rishij]

<p>Or, at least, someone please answer this simpler question:</p>

<p>To get the contrast ratio of my chart under my lighting, print a file on the same Epson printer that printed the chart that has half of the image black & the other half white, then put that up right next to my chart (under the same lighting) & meter the black vs. the white using a spot meter (or my camera's meter for that matter for a rough value)?</p>

<p>Thanks in advance,<br />Rishi </p>

[scott_ferris]

<p>Rishi,</p>

<p> The standard ISO chart is specified to have a minimum contrast ratio of 40:1, and is typically printed closer to 80:1.</p>

<p><a href="http://www.imatest.com/docs/testcharts_SFR_SVG.html">Link</a></p>

<p> </p>

[rishij]

<p>Thanks Scott. Wow it just seems unbelievable that my meter would show 6 stops (2^6 = 64) difference between a big black patch vs. a white patch under my indoor lighting... But Norman Koren does say that a photographic print can have a 100:1 luminance ratio... I need to invest in a spot meter :)</p>

[scott_ferris]

<p>That link is a good read :-)</p>

<p>They do point out that the iso 12233 is generally accepted to be much too high a contrast ratio for accurate testing of digital cameras and suggest 10:1 is better for a relatively high contrast target. Even the iso agree the 12233 chart has design flaws for current testing purposes.</p>

<p>Have fun, Scott.</p>

[rishij]

<p>Great, I'll finish reading that & then perform some resolution tests using lower contrast charts to maybe get a better 'real-world' resolution for some of the films I use (Velvia, Ektar, Portra). Thanks Scott.</p>

[StuartMoxham]

<p>Do we really need to drag every last bit of detail out of a piece of film to make a pleasing print. Once we get to a certain size with a given format the grain becomes the bigger objection for most than what details they could see under a microscope. If you need bigger prints then shoot a larger format instead of trying to drag the last bit of detail from 35mm film. Even if you could scan 35mm film at 10,000ppi does anyone really think that in general the prints would be nicer than scanning 6x7 at a lower resolution?</p>

[rishij]

<p>Grain is random in appearance, unlike most elements being imaged. Noise removal software takes advantage of this very fact. Oversampling to retain detail of actual imaged elements while also resolving more grain is therefore still advantageous especially if you use grain removal algorithms. In the end you'll preserve more image detail. It's hard for me to see why anyone <em>wouldn't</em> want that. For MF, you'll also resolve that much more detail.</p>

<p>As for switching up to the next higher format, easier said than done when it comes to film. On extended backpacking trips, for example, the versatility of the 35mm format is unparalleled. Of course a lot of these discussions will become moot with the release of higher resolution 35mm dSLRs...</p>

<p>I have really been considering stepping up to MF for fun anyhow... is it just me or are lens options with MF more limited compared to the plethora of zoom/wide-aperture lenses Canon offers for their 35mm EOS line? If so, why?</p>

[Dave Luttmann]

<p>Stuart, I think you hit the nail on the head. I use 35mm film for about 75% of every wedding I now photograph. I don't use film because of superior resolution. Fuji Pro 400H and Portra 400 are not resolution champs....especially not in 35mm. </p>

<p>My reason for using them is that when the 400 speed film is exposed for shadow areas, with a rating of 160-200 iso, I get incredible skin tones. In fact, I've had brides comment about how some wrinkles and blemishes have been rendered nearly invisible. I think a point that has been missed here is about color accuracy. Pleasing color is not necessarily "accurate" color. Nor does high resolution always make for a better photograph. </p>

<p>For me, in terms of on print, a 5D2 offers a resolution advantage over 35mm film. Not always in absolute terms, but in lower contrast situations, it does. Does it provide for the color I want? Nope. It can't. There is no way to overexpose the 5D2 (or7D or any DSLR) by 1-2 stops, while metering for shadows, without blowing out highlights.....and I still don't get the color I want from film. I can get close....maybe 85-90% of what I want, but not without a lot of work. The film workflow in that respect is more efficient. I shoot it, send it Richard Photo Lab, and then download the shots....done. Pretty much no PP is necessary.</p>

<p>There is more than one reason to use film. </p>

[mauro_franic]

<p>Same as Dave and Stuart here, even with a single MF frame scanned with the Nikon I have a lot more detail than my Epson 7880 can print at 24x32.</p>

<p>Pleasing (to my personal liking) colors, tones, and organic look are the main drivers why I shoot film. Also the fact that I can obtain a certain color rendition (not accurate in absolute terms) consistently and predictably. I don't shoot film only bcs I need to print large. With color negative I refer to color gradation or color definition advantages (difference in colors are crisp not smudged) but not color accuracy. I almost never shoot a gray target outside the studio.</p>

<p>On size, as part of my workflow I always make 8x10s, 8x8s or 4x4s first to green light a print and adjust framing lines. Some times I make 4 4x4s on a single sheet with different framing for evaluation. A print is judged small before I consider it for either small or large output. </p>

<p>I do like digital for some applications. For most uses I choose film. Although I wanted to use digital more when I got my 40D, after several months my personal preference pushed me back to film. In most cases it is just that, a personal preference.</p>

<p> </p>

[mauro_franic]

<p>I often shoot 35mm alongside 6x7 and I never hesitate to use a 35mm shot over a 6x7 shot for a large print if my instinct tells me so.</p>

<p>I sent Edward 24x36 prints from 35mm (along with the film, scans and raws from this color exercise). With pushed 35mm, grain may be visible at close inspection on prints that size but it never bothers me and I sometimes seek it. With all the grain removal tools I have, I have never made a print in my life where I had chosen to remove its grain.</p>

[StuartMoxham]

<p>I enjoy the workflow also with B&W film and I find it easier to get something that I like. I can develop B&W film myself which saves me trying to find a good lab.</p><div></div>

[tony_leinster]

<p>Just to go back to the original title of this thread (sorry guys lol) but we are talking supposedly about film vs digital COLOUR RENDITION. So for what it's worth and I know I'll be wearing teethmarks later,<br>

the 40D seemed to do very well, the outdoor shot was probably too saturated but still pleasing. The wood tone was at least like wood. The studio shots were both good with what I assume was the grey background looking neutral, maybe a very slight hint of yellow but about as good as it gets. Skin tone was, to me very good.<br>

Velvia was OK on the daylight but the studio shots seemed very blue and the skin tone was therefore cold. Also it did seem quite grainy for the ISO it is.<br>

Ektar, daylight was awful unless the wood really was that pink! Studio not too bad at all but still grainy for 100 ISO.<br>

Portra 160 and 400 to me look slightly washed out and the white flowers are blown, again both are grainy but the 400 is acceptable for that ISO.<br>

TMax All I can say is why is this in a post about colour rendition, not being funny, just don't understand it.<br>

I have to say that all these can probably be improved in post production but from the examples shown I would rather start from the 40D results in this instance. I fully understand that for medium format this may not be the case, well scanned film should, at the moment , be better.<br>

Since we are talking qualifications just for interest I worked for many years for Pentax, Nikon and Contax/Yashica and have owned two pro photo labs dealing mainly with commercial printing. I was in the photo industry for about thirty years, now retired. I use both digital and film although nowadays mainly digital for colour.</p>

[James G. Dainis]

<HR> Note from moderator<P>"...he too often presents excellent work in stupidly, and unnecessarily, misleading ways."<P>

 

As the moderator of this forum, I don't like to see the word "stupidly" used in that way when referring to another forum member. Differences of opinions are welcome but let us be a little more careful of <I>ad hominem </I> attacks. <BR>

 

<HR>

[rishij]

<blockquote>

<p>The standard ISO chart is specified to have a minimum contrast ratio of 40:1, and is typically printed closer to 80:1.</p>

</blockquote>

<p>Gave this site a read, which says that the ISO 12233 chart is printed closer to 80:1. But this confuses me. Tonight I used my camera's spot meter & a 50mm lens w/ extension tubes to measure the black vs. white on my print of the ISO 12233 chart under pretty uniform indoor lighting. I got 4.333 stops of difference, or a contrast ratio of 20:1 (2^4.333). Which makes sense, right, since prints, being a reflective media, are typically only supposed to have up to 4-ish stops of dynamic range?</p>

<p>How does one even get that a print of that chart to show 80:1 contrast? That's over 6 stops...</p>

<p>Also, the site has two charts available at contrast ratios of 20:1 & 2:1. But isn't the final contrast ratio dependent on your printer, paper, ink, & lighting of the print?</p>

<p>I feel like I'm confused & missing something here... please help.</p>

<p>Thanks,<br>

Rishi</p>

[mauro_franic]

<p>Rishi, I generate my charts on a dye printer instead of my Epson printers. Dyes are bad for archival but are higher resolving and produce better contrast than pigments on the same paper.</p>

[rishij]

<p>Thanks Mauro. Did you physically measure the contrast of the chart you shot with a spot meter?</p>

[mauro_franic]

<p>I didn't but next test I run I will. </p>

<p>The 4+ stops you are getting are pretty good though. I've run test with target printed on matte paper that wouldn't surprise me if they had just 2-3 stop range.</p>

[rishij]

<p>Yeah but I hear it's more accurate to use resolution charts with lower contrast b/c real world detail typically has less contrast. E.g. a 1:20 or 1:20 contrast ratio chart. I believe we covered this in an earlier thread. Do you recall what contrast ratio best simulates real world data? As Daniel Lee Taylor has pointed out a # of times in threads over the last few years, shooting such high resolution charts for comparison may be why film sometimes outresolves digital in these tests, yet underperforms in prints compared to digital SLRs of formidable resolution (the other huge reason being poor film scanning, which I think you & I agree on).</p>

<p>Also, sometimes the final reported value of the limiting resolution looks at where the response of the media has dropped to 50%. Or perhaps 10%. Your method of sharpening & increasing contrast before determining the limiting resolution of the film potentially overestimates the resolution of the film in comparison to the numbers reported when some of these sites test digital sensors.</p>

<p>I guess that's why it's important for the same person to perform some of these tests using the same exact methodology. I'm curious so I intend to.</p>

<p>Btw, interesting <a href="http://www.imx.nl/photo/Film/page169/page169.html">site</a> showing Orthopan film clearly outresolving a 18MP Leica M9. That looks like good methodology to me...</p>

<p>-Rishi</p>