I looked at Velvia 50 under a microscope today

[swilson]

<p>If the dots were square and uniform then a 10 micron square dot has its first null in the MTF curve right at 100 lp/mm. But it is closer to a circle and a circle with a diameter of 10 micron will have its first null at around 140 LP/mm. Due to a whole lot of other factors you are not going to see this limit when photographing but it does say that a 10 micron dye cloud is not at odd with just barely resolving 80 lp/mm. A 3 micson spot is far smaller then what is needed to resolve 80 lp/mm.</p>

[will_frost]

<p>Mauro Franic,<br>

You begin by saying "This post is not a competition with digital, it is a discussion on Velvia dye clouds."<br>

You then compare film to digital: e.g.: "A lot less resolution than just MF film and at a joke of a price."</p>

 

[mauro_franic]

<p>Will, my point was that an analysis of the dye clouds is far more interesting than the usual resolution of film vs digital (which I tested myself extensively). I also meant the capabilities of each medium are well understood - not a lot of room to see it as a competition.</p>

[mauro_franic]

<p>Scott, 3 microns gives you at least 2 clouds per line to resolve 160 lines/mm.</p>

[mauro_franic]

<p>Specifically to the OP's question:<br>

"</p>

<h1>I looked at Velvia 50 under a microscope today</h1>

<p ><a name="00WrgO"></a><a href="../photodb/user?user_id=5214737">Thomas Lozinski</a> , Jul 14, 2010; 04:31 p.m.<br>

At 100x magnification the grain or grain clusters? were about 10 nanometers (0.01mm) in diameter<br /> Counting each grain or cluster as a dot would give 2540 dpi and would give 24mm x 100dpmm x 36 x 100dpmm = 8.6 million dots (grains, grain clusters). I really don't think you can get more resolution than this out of the film"</p>

<p>The answer is obviously a) the clouds are much smaller than measured by the OP, and b) of course one can get higher resolution then 8MP from Velvia 35mm.</p>

<p> </p>

[j_marrs]

<p>This is such a pointless argument. If you want a film vs. digital debate, then says so in the beginning! But since that's where you're going, you can see that film is better for landscapes & people for big prints:<br>

<a href="http://www.twinlenslife.com/2009/05/digital-vs-film-real-deal-nikon-d300-vs.html">http://www.twinlenslife.com/2009/05/digital-vs-film-real-deal-nikon-d300-vs.html</a><br>

&<br>

<a href="http://www.kenrockwell.com/tech/filmdig.htm">http://www.kenrockwell.com/tech/filmdig.htm</a><br>

& this is from the Lightjet professional lab @ westcoast imaging. No digital outresolves 4x5 film from their printlabs.:<a href="http://www.westcoastimaging.com/wci/page/info/FAQ/faqprintlab.html">http://www.westcoastimaging.com/wci/page/info/FAQ/faqprintlab.html</a><br>

<img src="http://www.westcoastimaging.com/wci/page/info/FAQ/uploadfaq/images/megapixel-print-size-chart.jpg" alt="Megapixel Print Size Chart" width="570" height="857" /></p>

purple is supeb, dark blue is excellent, orange is only fair........ Looks like nothing digital has to offer can outresolve film.

<p>To see this chart in its entirity, visit <a href="http://www.westcoastimaging.com/wci/page/info/FAQ/faqprintlab.html">http://www.westcoastimaging.com/wci/page/info/FAQ/faqprintlab.html</a><br>

I've used westcoast for my prints from 35mm & Medium format & they are incredible. I love their canvas prints......<br>

As you can see 35mm is equivalent to a 15-22mp camera & 69mp can only match 6x6 film in the superb range. And nothing touches 8x10!</p>

[swilson]

<p>Resolution is only a small part of what make a photograph, but if you really do want resolution then digital is going to win every time.<br>

4x5 is not used for much for resolution but rather to get an image that is free of grain.<br>

<br />But if you really are a resolution junky then film is simply not going to give you your fix in this day and age.<br>

If you really want high resolution how about a 320 MP image, which is what you would get scanning a 4x5 film at 4000 ppi.<br>

<a href="http://sewcon.com/largephotos/pan1d%2009-12-08.jpg">http://sewcon.com/largephotos/pan1d%2009-12-08.jpg</a></p>

<p>But resolution simply losses meaning after a certain point. Whereas I think the op is off on the size of the cloud dyes and that due to this film can and does resolve more then what he has estimated how good an image looks is only somewhat determined but the resolutin captured.</p>

[art_thomas1]

<p>"Kodak Ektar 100 color negative film test ( updated d.d. 15-01-2009)" </p>

<p>This test was shown at a website cited early in this thread. The results that a Sony A900 looks better than either Velvia or Ektar were obvious. </p>

<p>However, the test was flawed. The test target was screen-printed, as shown in some of the downloads. Because neither the lines nor picture were solid, but screened, the fuzzy logic in the rather clever A900 converted them to what it thought the picture should look like. The film, on the other hand, had to deal with those pesky little white spaces in between bits of information on a screened print. </p>

<p>That means to me that no conclusion can be drawn from the test. It neither proves nor disproves Ektar vs. Velvia vs. Sony A900. </p>

<p>It does prove that the A900 provides good looking results, even if they aren't exactly what was photographed. </p>

[beamoflight]

<blockquote>

<p>"Kodak Ektar 100 color negative film test ( updated d.d. 15-01-2009)"<br /> This test was shown at a website cited early in this thread. The results that a Sony A900 looks better than either Velvia or Ektar were obvious.<br /> However, the test was flawed. The test target was screen-printed, as shown in some of the downloads. Because neither the lines nor picture were solid, but screened, the fuzzy logic in the rather clever A900 converted them to what it thought the picture should look like. The film, on the other hand, had to deal with those pesky little white spaces in between bits of information on a screened print.<br /> That means to me that no conclusion can be drawn from the test. It neither proves nor disproves Ektar vs. Velvia vs. Sony A900.<br /> It does prove that the A900 provides good looking results, even if they aren't exactly what was photographed."</p>

</blockquote>

<p>I am the maker of that test you are referring too so negatively. To be honest, I have refrained from responding to most of these negative reactions I have seen passing by, since I think it is generally pointless to respond.</p>

<p>I have to say though, that the thing that bugs me, is that most of the people that respond like this, haven't mostly ever done even a half decent test themselves and published it accordingly. In most cases, if anything is shown, it is badly executed, with photos not even taken from tripod... I am not saying my test didn't have some issues, it was a learning experience for me too, and I would definitely do some things differently next time. But to say that "the test has some flaws, and therefore no conclusions can be drawn at all" is just plain stupid. I think I have shown that there is still lots to learn, and from other responses I know quite a number of people agree with me.</p>

<p>Yes, the dot screen printing pattern was an issue, but it was still a magnitude smaller than what the a900 and and films could truly reveal. Only the direct flatbed scan of the test chart, truly revealed it.</p>

<blockquote>

<p>"It does prove that the A900 provides good looking results, even if they aren't exactly what was photographed."</p>

</blockquote>

<p>This remark made me laugh... I hit the release button, light enters the camera to expose the film or digital sensor, and now all of a sudden, the result "isn't what was photographed??!</p>

<p>In what 10th dimension world are you living in that I am not?</p>

<p>Sorry, but this is all about scale. Would you say the same thing if for example you photographed a model. You photograph her from a distance from feet to head up. Now you get in with a different lens, a macro lens, and photograph a small piece of her skin, revealing all the hairs on her arms, and even the skin pores.</p>

<p>Would you now all of a sudden conclude that the first photograph of your lovely model "isn't what was photographed"?????????? because "it doesn't reveal all the hairy and blotchy unevenness" of her skin?</p>

<p>Sorry, but this remark just doesn't make sense. Even if the screen pattern was close to be revealed, it still "is" what is photographed, and still does tell a lot about how all the contenders did in the test. In fact, the mere fact that all of the tested media were capable of revealing so much, is a compliment to all of them! People also don't realize how big the actual test chart was, it is NOT some small post card, but a giant 60x90 cm print.</p>

<p>If you think you can do better, why not do it and publish it properly... after two years, I am still waiting for a real good substitute for my test, not only executed well, but published and documented accordingly up to a level that I did. I haven't seen it...</p>

[j_marrs]

<p>If you do a test, at least follow the scientific method. One sample isn't going to cut it. You need a control group, and variables. I think we can all agree that digital is better than film @ high ISO, unless you want a lot of grain. </p>

[will_frost]

<p>Marco Boeringa,<br>

I found your comparison test for Ektar vs. Velvia vs. Portra vs. the A900 quite useful.<br>

J Marrs,<br>

In that test, I see a control (the flatbed scan), and four experimental groups. I think Marco Boeringa demonstrated a distinct difference in apparent resolution between the groups, and a distinct difference in apparent noise between groups. I encourage you to repeat the experiment, or to design and conduct a better one.<br>

Everyone else,<br>

If anyone reading this thread is interested, the test is located <a href="http://www.boeringa.demon.nl/menu_technic_ektar100.htm">here.</a><br>

Will</p>

[swilson]

<p>J Marrs,</p>

<p>Just what are you talking about with your control group and variables, the test was simply to compare a number of films and the A900 and I believe it did a good job at that. A control does not entry into this kind of testing.</p>

[art_thomas1]

<p>"If you think you can do better, why not do it and publish it properly... after two years, I am still waiting for a real good substitute for my test, not only executed well, but published and documented accordingly up to a level that I did."</p>

<p>I have.</p>

<p>"I haven't seen it..."</p>

<p>I'll take your word for it.</p>

<p>My test was a little different. I did Velvia 50 (original) , Ektar 25 (a little over the pull date even back then), Reala and the then top-o'- the- line Canon digital (with less MP than an A900). Back then I used a microscope (Yes, only on the film), a 4000PPI Nikon scanner , a 2900PPI Nikon scanner, bought 8000PPI drum scans and borrowed the local camera store's Canon digital demo. Lenses were my Contax/Carl Zeiss 135mm, F:2.8 at around F:8.0 and my Contax/Carl Zeiss 50mm, F:1.4. The Canon had the shop's best 135mm Canon lens, I think F:2.0 and their Canon 50mm, F:1.4.</p>

<p>I used both a Kodak test target (screen printed) similar to yours, and a USAF 1951 2'x3' test chart from Edmond Scientific. I also shot painted graphics on a truck about 200 yards away. I used a sturdy, heavy wooden (absorbs ground vibration better than metal) tripod with a weight hung from the middle post.</p>

<p>My results and conclusions:</p>

<p>1. The Ektar out resolved the Velvia 50 by a slim margin. The Velvia out resolved the Reala a fair amount. The Reala out resolved the Canon a fair amount. The Velvia out resolved the Canon a LOT. The Velvia had the smallest grain of the films.</p>

<p>2. The best looking result on the computer screen was the Canon for the screen printed target, with Reala the worst. On the USAF chart and the real world tests, Velvia on the drum edged out the Ektar but more to grain issues. Both Velvia and Ektar had more resolution than a 4000PPI Nikon as was seen on the 8000PPI drum. The drum did nothing for Reala except produce a file 4x the Nikon 4000PPI. Detail increased on the Reala between the 2900PPi and 4000PPI Nikons.</p>

<p>3. Digital prints were made of cuts from each as if the total scan would be 5'x7.5' on an Epson rated at 1440/2880 DPI input at 360PPI (it can take several dots to represent a pixel). Velvia off the drum looked best, the Canon looked the worst. There was a LOT more detail with the Velvia and Ektar than the Canon digital. Up-sizing was bicubic as Genuine Fractals was not available to me as I had thought it would be at the time.</p>

<p>No, I do not have the results ready to post at your very command as I am almost 3000 miles from home and will be for some time. The results have been posted on the net previously. Some may still be there. I know some are not as I did not re-up my site "paid membership".</p>

<p>I'll stick to my statement:</p>

<p>"This test was shown at a website cited early in this thread. The results that a Sony A900 looks better than either Velvia or Ektar were obvious.<br />However, the test was flawed. The test target was screen-printed, as shown in some of the downloads. Because neither the lines nor picture were solid, but screened, the fuzzy logic in the rather clever A900 converted them to what it thought the picture should look like. The film, on the other hand, had to deal with those pesky little white spaces in between bits of information on a screened print.<br />That means to me that no conclusion can be drawn from the test. It neither proves nor disproves Ektar vs. Velvia vs. Sony A900."</p>

<p>.........and further state that results from a screen printed test target do not prove resolution but can make computer generated digital camera output look pretty good.</p>

<p>But, being more precise in my use of the English language I'll revise my comment:</p>

<p>"It does prove that the A900 provides good looking results, even if they aren't exactly what was photographed." to say:</p>

<p>It does prove that the A900 provides good looking results, even if does not show exactly what was photographed. </p>

<p>P.S. <br>

I think you did us all a service doing your test and sharing the results and I thank you. </p>

[jamie_robertson2]

<p>This discussion beggars belief.</p>

[beamoflight]

<p>Sorry Art if I was a bit harsh on you yesterday. To be honest, it was not my best day...</p>

<p>We're probably not going to fully agree on the usefulness of the test I did, but that doesn't matter to me. Point is, I have at least made a serious attempt to be thorough and put a lot of work in it, and I don't like to be "dismissed" based on remarks made in "thin air" as quite often happens. Or based on some uncontrollable mythical "expert knowledge" that some people claim to have... Like saying "Oh, I heard two decades ago from a now deceased person at company X that lens Y can reach 200 lp/mm resolution with film Z. They have published it in an internal report that isn't accessible..." Yeh, sure!</p>

<p>Anyway, I would still love to see better documented results if anyone has them. Based on some of the other stuff I have seen passing by, I think my results were quite in line with others.</p>

 

<blockquote>

<p>1. The Ektar out resolved the Velvia 50 by a slim margin.</p>

 

</blockquote>

<p>I am not surprised by this. If you look at my results (which is Velvia 100F), the results of Ektar and Velvia are quite close too. I can imagine a flip between them in some cases.</p>

<p>In addition, Velvia 50 and Velvia 100F are two distinct films. I also shot a role of Velvia 50 after I published the test, but didn't include it as I found that the scans results didn't have finer grain than Velvia 100F (the F variants of Fuji films use the latest "grain technology" according to Fuji). In fact, according to the official Fuji film datasheets for Velvia 50 and Velvia 100F, Velvia 100F even has a marginally smaller RMS granularity, indicating a slightly finer grain:</p>

<p>RMS granularity for Velvia 50 (RVP50): 9<br>

http://www.fujifilm.com/products/professional_films/pdf/velvia_50_datasheet.pdf<br>

RMS granularity for Velvia 100f (RVP100F): 8<br>

http://www.fujifilm.com/products/professional_films/pdf/velvia_100f_datasheet.pdf</p>

<p> </p>

<blockquote>

<p>This discussion beggars belief.</p>

</blockquote>

<p>;-), yes, we should all be out shooting film, instead of discussing it till it's dead...</p>

 

[mauro_franic]

<p>As usual people are testing the scanner when they think they are testing film.....</p>

<p>Velvia outresolves Ektar as well as a 4000dpi scanner.</p>

<p>This is a tiny little crop from a test scanned with a Coolscan 9000 at 100%.</p>

<p>http://shutterclick.smugmug.com/Other/Ektar-TMX-Velvia/6616619_YJEwK#471880876_V8LEa-O-LB</p>

<p>Under the microscope (and also with the 8000dpi scans I posted), one can easily observe that Velvia outresolves the scanner. Ektar and the scanner have very similar limits.</p>

[art_thomas1]

<p>"Sorry Art if I was a bit harsh on you yesterday. To be honest, it was not my best day..." </p>

<p>A strong and healthy defense based upon hard work is a good thing. I certainly took no offense. I did however want to defend my experiment and work done. </p>

<p>"Point is, I have at least made a serious attempt to be thorough and put a lot of work in it" </p>

<p>You certainly have. Additionally, one thing it did point out is that digital cameras do tend to make irregular surfaces look more pleasing to the eye. This is a benefit to many digital shooters. </p>

<p>"1. The Ektar out resolved the Velvia 50 by a slim margin." </p>

<p>That is because it was the old Ektar 25 vs. the old Velvia 50. Comparing slides shot with a good lens in the real world and looking at detail in the distance, unscientifically, I feel the new Velvia 50 out-resolves the current Ektar 100. Even with the old Ektar vs. the old Velvia, microscopic comparison of resolution was hampered by the larger grain of the Ektar 25. At the point of my test, the Ektar 25 had both been frozen and was probably a couple years past the pull date. Ektar 25 has been thought to be one of the few films that is negatively altered by freezing. Freezing is supposed to increase its grain, among other things. I had neither the equipment nor inclination to test the Ektar 25 vs. the Velvia 50 when I had both films current and refrigerated, but unfrozen. Perhaps someone out there has done another test. </p>

<p>"and I don't like to be "dismissed" based on remarks made in "thin air" as quite often happens" </p>

<p>I'll second that. </p>

<p>Again, I thank you for your well-thought-out and meaningful test. You, of course, did do your own work, use some thoughtfulness, maintain consistency and standards, and came to a rational decision. Although I disagree with a part of your conclusions, no one could rightfully say that you did not materially and unselfishly add to our knowledge base. Again, I thank you. In my mind, you join the ranks of contributors like Les Sarile, et.al. who contribute meaningfully to photonet and the enthusiast as well as the professional community. </p>

[steve m smith]

<blockquote>

<p>As usual people are testing the scanner when they think they are testing film.....</p>

</blockquote>

<p>That is the problem with nearly all digital vs. film comparisons. They are really digital vs. scanner tests. A fair test would be a good print made from a fully digital process compared to a good print made from a fully film and optical print process.</p>

<p>Another problem with these test is that they are often biased towards the reviewers personal preferences.</p>

[david_henderson]

<p>It depends what you're trying to find out. For most people, making prints from film will involve scanning somewhere in the process. So there's every point IMO to assess a print from a film scan vs a print direct from a digital photograph on the same paper and using the same printer type. </p>

<p>Just as there's every point in comparing prints made the same way and to the same spec using different scanners ( I never can tell what a good scan looks like on a screen).</p>

<p>And comparing prints from the same scan ( or digital photograph) using different labs and printer/paper combinations</p>

<p>and so on.</p>

[syd]

<p>I'm still shooting film ... 100% Velvia, scanning it and loving it. No test required for that ...</p>

[Dave Luttmann]

<p>Best response yet Simon. I use Velvia and Astia in MF and 4x5. Don't need a test to see I like the look of films better than a lot of digital work.....and 35mm is fine for virtually grain free 16x24.</p>

[swilson]

<p> In the end I think each person needs to do their own testing. In 2005 I was questioning the need to continue to shoot film so I did some tests to see if there would be cases where the film would be a clear winning in image quality. For my cameras, lenses and scanner I could get more detail with film of high contrast subjects, but not by a huge amount and the other issues with film made its overall image quality not as good as what I was getting using my digital camera. These tests to help me decide if there was a point for me to keep shooting film, there was not. But other people with other gear really need to test for themselves.<br>

<br>

I have also done tests with friends where we go out and shoot the same scenes, I shooting digital and them shooting film, and then we compare the prints. This has helped them get an idea of what they would get moving from film to digital, given the film camera, film, lens and scanner they are using.</p>

 

[danielleetaylor]

<p><em>Scanners - like the Coolscans and others, don't have bayer issue like DSLR sensors. You should consider this in your response to my first question about your 12MP digianything.</em></p>

<p>It's evident from your own test samples that 12 MP DSLRs (i.e. the D2x) out resolve desktop scanned 35mm film when it comes to low contrast, color detail. I've demonstrated that the 18 MP 7D can out resolve 35mm Velvia, scanned on an Imacon, on low contrast color detail. So what "Bayer issues" are you referring to in relation to resolution?</p>

[danielleetaylor]

<p><em>OP, as Edward indicated 35mm Velvia resolves 160 lpmm (this is over 4,000 true dpi) on a high contrast target. This is materially (about twice as much) higher than any 12mp camera for high contrast and fine detail.</em></p>

<p>Come on Mauro, you know better than this. Velvia only records 160 lpmm on 1000:1 lab tests. You never, ever find 1000:1 contrast in fine detail in the real world. You would have to shoot the sun through a picket fence to achieve those kinds of lab conditions. Note that 1000:1 is greater than Velvia's dynamic range, which means you get no tonality, just black and white in those tests.</p>

<p>Velvia's resolution is roughly 80 lpmm for real world high contrast detail. Velvia on an Imacon or drum scanner will just barely out resolve a D700 on high contrast detail, but that doesn't mean it will necessarily look better. It would take a lot of work to make the Velvia scan as sharp and noise free as the D700 shot. (Then again, it can take a lot of work to give a D700 shot the same color as Velvia, which may be preferred by the photographer.) As I've demonstrated in past threads, an 18 MP 7D RAW shot is a pretty good match for the very best 35mm Velvia scans and exceeds Velvia in low contrast resolution, while pretty much matching it on high contrast resolution.</p>

[will_frost]

<p>Just out of curiosity, are there test charts for "medium contrast resolution"? Something like 18% grey bars on a white background and a black background? That would be a helpful standard, wouldn't it, since you'd be pretty much guaranteed to run into that sort of contrast in the real world most of the time?</p>