Jump to content

I looked at Velvia 50 under a microscope today


Recommended Posts

<p>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. I'd say my 12 megapixel camera looks better when blown up 100x (to 150 inches on my computer and looking at 100%) than the velvia did. I'm not trying to start an argument but just wondering if anyone else has actually gone and done this simple test themselves and would agree with me or correct me.<br>

I know there are some finer grained films than velvia but not many color films. Now I'm seriously doubting the higher claims of film resolution. </p>

Link to comment
Share on other sites

  • Replies 102
  • Created
  • Last Reply

Top Posters In This Topic

<p>Yes, 10 micrometers, thanks for the correction. (can't believe I made that error.)<br>

The microscope was on 10x objective with 10x eyepiece but that is irrelevant since I was using a micrometer slide placed over the slide (slightly had to adjust focus) but I could get a good approximation of grain size. </p>

<p> </p>

Link to comment
Share on other sites

<p>David, in general yes that is what I'm saying with a few caveats. My measurement was very quick and my micrometer only went down to 1 micron (although if anything the dye spots were a bit bigger than 1 micron.) Scanning at 4000 might give a slight edge since the 2540 ppi aren't going to be in exactly the same spot that the sensors array are in. I'm not taking into account bayer array sampling since the dye spots don't seem to only be one color (I may be wrong here.)</p>

<p>This was just a quick very simple test and I encourage others with microscopes to spend 5 minutes and see what they see. I've spent way too much time reading different points of view on what is better and I should have just done this quick test myself instead.</p>

Link to comment
Share on other sites

<p>I'll examine a few Velvia slides under a microscope next week. I did it before, but I didn't take any notes. All I can remember is that I could see details that were not resolved in 4000dpi scans. I think that your conclusion wrong and that Velvia and other modern colour films hold a lot more information than your numbers suggest. Basically, dye clouds in colour film are not aligned in rows like cells in sensors (pixels, if you like), but are distributed randomly in different layers, hence counting them under a microscope, which has limited depth of field, in order to calculate the equivalent pixel count does not make sense, or at least is not as straightforward as one may think. That said, I am unable to tell at this time how exactly you are wrong, and it is only my opinion that you are. I hope I'll be able to find out based on my own observation.</p>
Link to comment
Share on other sites

<p>Measurement issues aside, film grain is not linear (grid) and cannot be counted as such. It "clumps" more around denser exposure than light/no exposure. A digital imager is a completely linear grid. You simply cannot "count" mega pixels from film; you would get different numbers with different exposures.<br>

In my opinion, this is largely responsible for the "film look" that digital cannot emulate. It may be why it is easier to get a pin sharp image from a digital.</p>

 

Link to comment
Share on other sites

<p>Lee<br>

When I said that my 12 megapixel camera looks better when blown up 100x that is exactly what I meant. I am not saying that the resolution is better. The velvia at 100x looks like a bunch of dots and isn't pleasing to look at. The digital image resized to 150 inches wide and viewed at 100 percent did look better, maybe because tones were smoothed out from photoshop interpolation. I can't say for sure that the camera has more resolution (especially considering only 3mp for blue, 3 for red). I would further say that the resolution of the film appeared to be less than the dot size as would be expected. I'm trying to be as objective about this as possible; I'm not taking a side. I shoot both film and digital.</p>

<p>Yes, the microscope has a limited depth of field, but the actual emulsion is extremely shallow (at least compared to the thickness of the film.) </p>

<p>The following much more in depth article agrees with my measurement for size of dye spot (10 microns) although they indicate resolution is slightly better at 8 microns. They also indicate that Velvia is capable of 4000 ppi (almost double what I estimated in my 1 minute experiment.) Note that they are doubling in the conversion from lp/mm to ppi because of Nyquest) If at any point in your workflow you are converting from film to digital (to scan or to print) you are going to have to take this into account.</p>

<p>http://cool.conservation-us.org/coolaic/sg/emg/library/pdf/vitale/2009-10-vitale-filmgrain_resolution_v24.pdf</p>

Link to comment
Share on other sites

<p>I can't contribute to the questions of physics and testing and what you have to do to compare film scans and digital photographs.. What I can say with considerable certainty is that Velvia over a couple of decades was used by a great proportion of landscape and nature photographers to whom detail and sharpness of detail have been very important. I'm just struggling a bit with the thought that all these people did this despite the fact that the resolution of their film turns out to be not very good at all. I'm also struggling with a view that its taken nearly twenty years for a guy to have a eureka moment which demonstrates that all these picky landscapers and tree photographers haven't actually been able to produce large detailed prints from scanned Velvia. Most of all I'm looking at some pretty detailed 36" sq prints I have here from scanned Velvia and wondering how my lab managed to make these prints (which don't seem to rely on distance for their impression of sharpness) from a film that won't scan at more than 2500 ppi?</p>

<p>I guess I think that sooner or later you'll find a flaw in your measurements, or in the conclusions that you're drawing from them. I'm also pretty sure that I've seen threads on here about how much useful data can be extracted from film</p>

Link to comment
Share on other sites

<blockquote>

<p>They also indicate that Velvia is capable of 4000 ppi (almost double what I estimated in my 1 minute experiment.) Note that they are doubling in the conversion from lp/mm to ppi because of Nyquest)</p>

 

</blockquote>

<p>The author of the linked article uses MTF 30% as the native resolution of film, that is without using a lens, from the Velvia datasheet. He is very clear about it and it has nothing to do with Nyquist. The MTF 30% for Velvia according to the author is 80 lp/mm or 4064 ppi. It is unconventional to use MTF 30% as MTF 10% is considered the resolution limit. Using MTF 30% leads to lower resolution for film+lens (on page 16 of the linked article) as opposed to MTF 10%.</p>

Link to comment
Share on other sites

<p><em>I'm just struggling a bit with the thought that all these people did this despite the fact that the resolution of their film turns out to be not very good at all. I'm also struggling with a view that . . . all these picky landscapers and tree photographers haven't actually been able to produce large detailed prints from scanned Velvia. Most of all I'm looking at some pretty detailed 36" sq prints I have here from scanned Velvia and wondering how my lab managed to make these prints (which don't seem to rely on distance for their impression of sharpness) . . . .</em></p>

<p>David, I think you are considerably overstating the degree to which Velvia is being criticized, if indeed it is being criticized at all; I do think the OP is probably roughly correct about its practical, real-world resolution limits (which imply that 35mm Velvia has about the same resolution as a typical Bayer-sensor DSLR of about 14 MP).</p>

<p>Let's put this into a little perspective. Suppose we were to say that you get at least some useful image detail out of Velvia 50 scans out to 2800 ppi or so. If those 36-inch-square prints are from a 6x6 (as are most square prints, I suspect), then a 2800 ppi scan means you have about 171 ppi of real image detail to work with. With many subjects, 171 ppi of real detail is quite sufficient for a pleasing print. Indeed, the most widely-quoted limits of human vision of which I'm aware (0.5 minute-of-angle) correspond to about 380 ppi at a viewing distance of 18 inches, and that's for extremely-high-contrast black-and-white detail. For practical human vision of ordinary pictorial subjects, the Velvia print may not be leaving behind to much detail <em>visible to a viewer with average eyes</em>.</p>

<p>Also, high sharpness, even at moderate resolutions (ever notice how Fuji's Velvia 50 MTF response graph is <em>above</em> 100% from about 2 lp/mm to about 25 lp/mm?) can create a fairly convincing impression of detail.</p>

<p><em>You simply cannot "count" mega pixels from film . . . .</em></p>

<p>Absolutely correct--they're too different. Any claimed equivalency is at best an approximation for a given set of conditions.</p>

 

Link to comment
Share on other sites

<p>The normal human eye with 20-20 vision can resolve detail as small as one minute of an arc (1/60 degree)<em></em><em></em><em></em>. At 18 inches the human eye can resolve 3.76 lp/mm or 191 ppi, not 380 ppi. At 10 inches the eye can resolve 6.76 lp/mm or 343 ppi.</p>
Link to comment
Share on other sites

<p>C. Sharon, our issue is not math, it's whether the limits of normal human vision are 1.0 MOA (you) or 0.5 MOA (me). I'm not sure exactly how 20/20 is defined, but I've known several people who were better than that (20/15, 20/10). I guess it depends how far out on the bell curve you want to go whether you want to say the limit is 1.0 MOA, or 0.5 MOA, or what.</p>

<p>Obviously if your (corrected or uncorrected) vision is poor, <em>that</em> will be the limit of the detail you can see. But for my brother who is 20/15 or a little better in one eye, well, he can see detail finer than 20/20, if it is there to be seen. I'm pretty myopic, and even with my contacts, I only get to 20/25 or 20/30.</p>

 

Link to comment
Share on other sites

<p>Dave: From Wikipedia - In the term "<strong>20/20 vision</strong>" the <a title="Fraction (mathematics)" href="http://en.wikipedia.org/wiki/Fraction_%28mathematics%29">numerator</a> refers to the distance in feet between the subject and the chart. The <a title="Denominator" href="http://en.wikipedia.org/wiki/Denominator">denominator</a> is the distance at which the lines that make up those letters would be separated by a <a title="Visual angle" href="http://en.wikipedia.org/wiki/Visual_angle">visual angle</a> of 1 <a title="Arc minute" href="http://en.wikipedia.org/wiki/Arc_minute">arc minute</a></p>

<p><a href="http://en.wikipedia.org/wiki/Visual_acuity">http://en.wikipedia.org/wiki/Visual_acuity</a></p>

<p>Of course there are individuals with better than 20/20 vision but 20/20 is the statistical standard for "perfect vision" hence it is used to calculate what the human eye can resolve.</p>

Link to comment
Share on other sites

<p>C. Sharon,<br>

I thought they were using nyquest from the math . . . 83lp/mm converted to dpi<br>

83lp/mm * 25.4mm/inch = 2108 line pairs per inch, unless a line pair is 2 dpi, one for the positive line one for the negative line? </p>

<p>Les,<br>

That is the main point of my post. People should test stuff out and see what they find. That is a very useful test you did on the crayons (although it doesn't say anything about the dynamic range in ONE print.) You have shown that if you really screw up the exposure you can get a normal looking photo, but you haven't shown that ektar can capture highlights and shadows at the same time in your very evenly lit scene.) Still a very good test. It looks like you could have even gone farther with the Ektar especially since reciprocity would help you another few exposure values up. I'll try repeating your test with my cam and see what I get but I have a feeling it won't beat the Ektar</p>

<p> </p>

Link to comment
Share on other sites

<p>Well I'm looking at a recent slide under a scope right now and am a bit surprised that you could make out the grain at only 100X. I'm looking through at 400X right now and would be hard pressed to say where one "grain blog" ended and the other started. That being said I'm not very good with a scope but will try to take some photos through the eyepiece later if desirable. <br>

However, I also feel that my digital capture (D80) of the image had superior sharpness, but not color. I do attribute this though to a great difference in equipment used. The digital was taken with a D80 and a 35mm FL. The film was taken with an old OM-10 manual focus with 50mm FL. Same aperture (don't know what it was) but of course the smaller lens will give better depth of field and the digital was certainly better focused. That has led me to the belief that to do a really accurate comparison one needs to shoot not only the same lens but also the same image sensing area. </p>

Link to comment
Share on other sites

<p>Can't say that I agree with you. I have looked at slow slides with a high-end microscope and high-contrast detail was resolved (by the microscope lens) much better than from a 4000 ppi scanner. I'd guess that to get a really clean scan you'd have to scan at 20000ppi and then downsample to your target requirement to avoid aliasing of the grain and to get all the detail crisply as it is in the film. The detail I looked at was a backlit spider web which is of course a very high contrast subject; with low contrast detail film can be really crappy.</p>
Link to comment
Share on other sites

<p>Thomas,</p>

<p>I don't have an 100x microscope with a micrometer stage, but by any common sense measure, your estimate of the grain size of Velvia at 10 microns seems very large, perhaps by a factor of 10. The dye clouds will overlap and perhaps run together in certain parts of the image - clear blue sky, for example - which could be misleading you.</p>

<p>If you wish to measure resolution, measure it directly, using an high-contrast resolution target and non-limiting optics. On second thought, that's already been done. Fuji publishes the results for Velvia, with the value of about 160 lp/mm for such a target, and a more modest 60-80 lp/mm for conventional subjects. The MTF50 resolution for an high-resolution target is still 45 lp/mm (comparable to your estimate).</p>

<p>The best obtainable resolution (160 lp/mm) occurs when the density of the lines is barely discernible from the density midway between lines. This corresponds to an uncertainty of about 3 microns, placing an upper limit on grain size. Not all grains are the same size, and there is a significant amount of chemical diffusion and light scattering within the emulsion, so it's safe to say average size is on the order of one micron or less.</p>

<p>A 12-13 MP FX DSLR looks pretty good by comparision because the MTF50 resolution is close to 60 lp/mm, which is what really counts in a print.</p>

Link to comment
Share on other sites

<p>This is as bad as it gets.Have no idea how the image from a 12 MP is as good as Velvia.You don't have to measure anything just look at them side by side.I know no fine art photographer who uses a 12 MP camera for his work. </p>
Link to comment
Share on other sites

<p>I'm sorry but, after reading so many postings like this one i have to ask...What is the point and what are you trying to prove?? Some people go on and on about micro this and micro that and trying to compare film to digital. If i'm that bored that i have to look at slides under a microscope and count grains i would think i need help of some sort. In my opinion, film and digital are not the same, therefore you cannot compare them. </p>
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...