October 02, 2014, 01:22:58 PM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Policar

Pages: 1 ... 19 20 [21] 22 23 ... 28
EOS Bodies / Re: FIRST Video comparison Canon 5D mkIII VS 1D X
« on: June 23, 2012, 05:41:09 PM »
I have analyzed the full res file.... the 1dx is also much sharper! mayby is true 1080P  ;D

It is possible that the focus was not spot on during the test shot with the 5D3. What would you define as true 1080p resolution? 1920x1080 or 1920x1080? LOL get it?...  ;)

What would I class as 1080? 1920x1080 in all 3 colours. Not just 1080 bayer filter lines.

So you don't consider the Alexa and F3 to be 1080p cameras?

EOS Bodies - For Stills / Re: IQ difference of FF vs APS-C?
« on: June 20, 2012, 09:30:36 PM »
The relative increase in sensor size between a compact and APS-C camera is much greater than that between APS-C and full frame.  And everything will look good in good light.  In low light there's a big difference, and there's more microcontrast and shallower depth of field (feels a bit like the difference between medium and large format film).  There's a substantive difference, but a good lens makes a bigger difference.  The 5D III is more forgiving of bad lenses, though; the 50mm f1.8 is great on it, even wide open, since the pixel density is less.

The best landscape photography is still shot on Velvia, which has four or five stops of DR.

Thank you, that gave me a good laugh!

Let me guess, you like HDR, UWAs, and instagram?

It's not that important unless you're in a rush and can't wait on light.

The best landscape photography is still shot on Velvia, which has four or five stops of DR.  If need be you can always use grad filters or multiple exposures, anyway.

Third Party Manufacturers / Re: Film is still hard to beat
« on: May 20, 2012, 08:05:40 PM »
Completely off topic, since we're drifting there anyway, Loren Eiseley was another scientist with a strong artistic side. His essay, "How Flowers Changes the World," should be mandatory reading for any human being.

Drifting further off topic (why not?), for anyone who hasn't heard of him, check out the works of Ernst Haeckel, especially Kunstformen der Natur (Artforms of Nature). I wish I could take macro photos half or even 1/4 as good as his 120-year-old drawings.

Slightly back on-topic, when shooting film, it all depends on how slow film you use to take shots, how good your scanner is, etc. Lots of people making comparisons (not just here, but on the interwebs), are forgetting that they're scanning slides shot 30 years ago on glass that didn't have as fancy coatings as now (although some old lenses are sharper than current counterparts, i'm looking at you, Super Takumar 50/1.4 vs EF 50/1.4).

Aside: i've just thought of a test, seeing as I've got an EOS 3 now (now there's one way film beats digital, Eye control AF), and a roll of Velvia 50 lying around, i'll go ask my friend if he's bought his 5D3 yet and do a direct same-shot-same-lens comparison one day.

Anyway, let't think mathematically.
5D3 is 5760x3840 = 22.11 x10^6 pixels.
Let's scan a 135mm film to 4800dpi:
36/25.4*4800 = 6803x4535 = 30.85 x10^6 pixels.
OK, so scanning all but the finest-grained film is not going to give you the same sharpness as a D800. So let's downscale the scanned film a bit later on.

Now, as is my understanding, an iso100 film has the same grain size, whether it's deposited on a 135, 120, 4x5, or 8x10 negative. (if i'm wrong on that, better stop me here).
So let's scan in a 645 (cropped to 3:2 ratio) film, 56x37.3mm, at 4800dpi.
56/25.4*4800 = 10582x7055 = 74.65 x10^6 pixels.

OK, that's about what I was expecting (i was typing as I worked, so i'm reading the results now too).
Basically, scanning 135 film gives (near enough) equal to the best FF digital in terms of MP (D800), just as scanning 120 film gives equal to the best MF back (IQ180). That's assuming that scanning film to 4800dpi gives just as good results as the digital equivalent, which is a test for later. But the results should scale, if I can only scan a 135 Velvia 50 at 2400dpi (or more likely, to 4800dpi then downscale), I'll get 1/4 the res (7.5mp), and scanning 120 film to the same will give 18MP.

Anyway, that's just from an MP perspective. DR and colours is a whole different perspective, and we could go all day, but they're just not directly comparable (vinyl vs CDs vs MP3s anyone? or transistors vs tubes?).
The best thing about film is the non-linear response curve, so highlights don't get clipped as easily, even if you overexpose a bit too much (if you're good and process your own), you can recover a bit by underdeveloping. With digital, once the photon-bucket is filled, it's filled, no going back (although you can instantly see results and shoot again, another plus on digital). So there's never going to be a direct number-to-number comparison. Knowing your tool is more important than the tool itself in this case.

The best film will always beat the worst digital.
The best digital will always beat the worst film.
I think that's the only thing we can say with certainty...

Except that megapixels mean nothing as a metric and mtf means everything.  And both scanners and sensors are supposed to be designed to oversample by a factor of at least 2 (although they don't) to prevent aliasing.

A D800 shot will put Velvia 50 to shame every time, mm per mm.

EOS Bodies - For Video / Re: Help
« on: May 20, 2012, 03:54:34 PM »
I used some weird free software for OSX and it recovered most of my files for me (.mov and .jpg).  Forget what it's called but it runs in the terminal, kind of weird.  Lexar has software that does the same thing and isn't free.  Just don't shoot any more with it.

Third Party Manufacturers / Re: Film is still hard to beat
« on: May 20, 2012, 02:34:10 PM »
Correct me if I'm wrong, but this "digital vs film" quality issue doesn't really kick until you start printing at large size (beyond 4x6in and magazine spreads).

I'm guessing that the majority of photographs are consumed (by the public) at small sizes (magazine size or smaller), and often on digital devices. At those sizes, is there a technical difference between digital and film?

Smaller than 8''x10'' anything, including 135, is more than good enough.  APS-C digital will hold up almost flawlessly at 11''x17'' (and probably acceptably at any size) unless you're super picky.  The 8x10 film prints I saw were 80''x100'' and at that point you do get a distinct advantage from large format, however.

Third Party Manufacturers / Re: Film is still hard to beat
« on: May 20, 2012, 01:19:18 PM »
That's more of a condemnation of your slide projector (or eyes) than anything...  And it's just not true; my 1080p projector is much softer than the slides I've shot.  Also, I've yet to find a digital camera that's even close to "per-pixel" sharp.  Bayer interpolation and the olpf knock resolution down by at least 30% linearly (and in theory should knock it down by 50% to satisfy nyquist, but of course no one does this).

In terms of extinction resolution 135 and FX are really pretty similar.  I have 25 megapixel scans of Velvia that have detail almost down to the pixel level.  The only issue?  Contrast is quite poor there and it's very grainy.  Subjectively, 135 looks closer to 4 or 6 megapixels than it does to less than one...  You can print up to 8x10 with 135 no problem; you can't from a web cam.

Large format, while no more detailed than any other given medium at normal apertures (due to diffraction), still has the best "look" by a very large margin (excepting maybe those 80 megapixel MFDB backs) just because the lenses are so darned good.  But it takes maybe 20 minutes and six dollars to shoot a photo and $150+ for a high quality scan, so you get what you pay for.  But 4x5, and particularly 8x10 (which I've never shot, only seen prints from) is still a worthy format, even if 135 and 120 are on the way out.

Third Party Manufacturers / Re: Film is still hard to beat
« on: May 19, 2012, 10:50:38 PM »
Modern SLRs (even the F4, which I briefly owned and quite liked) behave a lot like dSLRs except the ISO isn't adjustable on a per-shot basis, of course.

In terms of user interface, I prefer manual focus SLRs (135 and 6x7) to dSLRs and even to modern SLRs because the interface is so much simpler and you don't need to replace batteries.  Set your stop, meter your scene externally (spot or incident as appropriate) and decide on an exposure, set your shutter speed appropriately, focus, take a photo.  There are only three variables:  focus, f-stop, and shutter speed.  With a dSLR it's like using a computer, so complicated and there are so many modes for everything.  I still have no idea how to change focus setting appropriately with my 5D III and no idea what P mode does.

Large format is difficult enough to shoot that it's materially more painful than shooting digital, but 135 (either on a modern SLR or an older one) is nothing to be afraid of except that there's no longer any real reason to shoot it except nostalgia or fun.

Third Party Manufacturers / Re: Film is still hard to beat
« on: May 19, 2012, 08:36:28 PM »
You're surprised to find tech geeks on a camera forum?

I guess I shouldn't be, but it seems so simple to look at the results from two cameras and decide which is better and so complicated to try and interpret the science of human perception and the science of image recording (neither of which any of us here really understand to a significant extent) and then apply that. 

But in the sciences everything has switched to digital...and so for scientific purposes...yeah, it's by far the best.  And FX digital probably trounces 135 in general.

Third Party Manufacturers / Re: Film is still hard to beat
« on: May 19, 2012, 08:25:45 PM »
I'm just genuinely curious about the scientific difference. I'm not tyring to figure out which is better from a Whitmanesque "mystical" point of view. It should go without saying that amazing beautiful pics can be taken with BOTH. That isn't really the core question, imo.

Why would you care about the scientific difference?  Isn't the point entirely how it looks?  How do you even measure that scientifically?  Color gamut?  Acutance?  Resolution?  At how many lp/mm?  Zeiss lenses were designed for good micro contrast but worse resolution...is that better or worse and in what contexts?  Is grain good or bad?  What colors (if any) do you like?  How do you feel about false detail (aliasing), is it nice and sharp and crunchy or disgusting to you?  Is a lack of DR good or bad?  (This is a tricky question--printed images only have four or five stops of contrast at best so a capture with more DR than that looks flat when printed, but one without enough loses detail--the answer is of course subjective and it's based on the subject and light and how the image is developed.)

As a scientific instrument, digital is way better.  Astronomers (appropriately enough) switched from film to CCDs in the 1970s and have not looked back.  The simple answer here is that digital is WAY better in general.

In terms of signal/noise ratio, digital is just way better.  Way, way better.  No argument from anyone.  But some people like film grain because it looks more random and smooth.  The 5D Mark III I have found has pretty ugly noise, imo, while some digital cameras have amazing beautiful noise with a great texture--so that's a whole other subjective discussion.

In terms of resolution it's complicated.  Velvia (a very sharp color film) has mtf curves that resolve without aliasing to about 60 or 80 cycles/mm (at like 30% mtf).  The D800 has about 200 pixels per mm or 100 cycles/mm.  As per the nyquist sampling theorem that means 50 cycles/mm without aliasing and that's not even taking into account bayer interpolation.  So it sounds like digital is much worse here, but it's not!  Velvia drops off from >100% mtf to <100mtf around 20 cycles/mm but bayer sensors resolve to about 100% mtf until almost 70% of their stated resolution.  I think.  Without an antialiasing filter, the D800E might resolve 100% mtf (>100% mtf once sharpened) until 70 or 80 cycles/mm.  Of course there might be aliasing, which is a problem....except that aliasing looks subjectively like detail, so you might get the appearance of >100% mtf until or extinction with the D800E.  The best measure of subjective sharpness is the area under the mtf curve, which in that case would be dramatically larger for the D800E than for a slide of Velvia.  Even though, in theory, the Velvia can resolve to a higher resolution without aliasing.  Of course you need to scan film and even a drum scanner will knock off quite a bit of mtf from the system.

Imo, the combination of reduced noise/grain and increased mtf puts state of the art digital at twice the linear resolution of film.  Digital printing's vast superiority to darkroom color printing tips the tables even way further.  APS-C looks like 645 to me.  Full frame looks like 6x7.  But I prefer how 6x7 Velvia looks to how the images from my 5D III look by a pretty enormous margin.  Even though I can't explain why and even though others don't.

In terms of DR, it depends on which film (black and white negative can have easily way in excess of 10 stops, all of which are usable if you dodge and burn; Velvia has five stops maybe) and how you measure it (how much noise/grain is too much and if a soft highlight rolloff that doesn't contain recoverable detail but still looks nice counts as real DR).

In terms of color gamut, digital is more accurate but film can have a wider gamut in theory.  Once scanner...doesn't matter as much, the gamut is squished.  And some films have more vivid colors because the spectral sensitivity curves reject more colors than the weak bayer filters on digital SLRs.

But yeah, digital wins for a given sensor size by far.  Large format film (Velvia 50, specifically) is by far my favorite in terms of aesthetics, but the price is high and you need to be very careful about light due to the limited DR and how easy it is to blow an exposure.

It's just amazing that a 135 format sensor, now, can outresolve and outproduce in some situations formats such as medium and in your case, large format film...  Film was great but film was noisy... Wanted to shoot above 2000 ISO, forget about it... it would be so grainy you could barely make out what you were shooting, plus the cost of getting such a delicate film would not be worth the outcome.  I understand people are comparing the D800 to digital medium format now... but then again digital medium format now is out-resolving large format back then...  Looking at that perspective, it's mind-blowing to see where we have come from.

Absolutely, it really puts a lot of today's complaints in perspective.  Too noisy at high isos?  Not enough resolution?  Try shooting on film...  (That said, black and white film has a nice texture to it so you can get away with big enlargements.)

All the same, the photos I take with my Nikon F, which has a broken meter and broken slow shutter speeds, are consistently much better than those I take with my 5D Mark III, even if they're worse technically and I wouldn't print them past 8X10.  So sometimes a surplus of riches isn't exactly what you need.  I'd love to get one of those MFDB tech cameras but the prices are so ridiculous.  View camera lenses have a certain magic to them.

Not to complain too much.  The 5D Mark III is amazing.  At low ISOs, I would treat it like slow 6x7 film in terms of estimating print size.  I've seen much larger prints that look okay, but not up close.  The D800 is probably somewhat better, but that's pretty trivial compared with how completely amazing both are.

There's a lot of comparisons of the D800 and medium format, but to me, 135 cameras were pushing medium format with the intro of the 5d2.

5DII digital prints and 6x7 cibachromes (off Velvia 50) looks surprisingly similar in terms of sharpness.  I would give the advantage to the 5DII print, but a good scan of 6x7 printed digitally might still look a bit better (more detail but also much more grain).

I think people are comparing the D800 with MFDBs, not 6x7 film.  Most large format landscape photography doesn't really exceed 20-40 megapixel sharpness because the stops at which you're shooting (f32-f64) knock resolution down due to diffraction.  With a good tilt/shift lens, I could see the D800 resolving as much detail as large format, easily.  That said I wouldn't expect the tonality to be quite as smooth as 8x10 and view camera lenses have their own advantages.

It should be stutter, not flicker, that bugs you at 24fps....most theaters project with three or four blade shutters, so the flicker itself is above the flicker fusion threshold (72/96hz).

This comparison is useless since we don't know the lenses or other settings.  The Scarlet should have done better.  The Red MX and Epic look fine at 1200+ ISO.  That said, the 5DIII is a true low light monster.  If you turn highlight tone priority off (which results in clippy highlights), you can shoot 10000 ISO and it looks fine.  With f1.4 lenses, it's pretty incredible.  The availability of affordable 24mm f1.4 primes (≤17mm f1.4 primes for APS-C are harder to find) make the 5DIII an incredible camera for low light narrative.

if the data is binned before readout by low level sensor circuits as some have suggested, I doubt a firmware induced process change on the post binned data will help resolution.

The resolution isn't really that bad, though, if you look at resolution charts.  Something like 800-900 lines, which isn't terrible.  It's the apparent softness that's an issue, and that's compounded by the bad sharpening algorithm that you can't turn up at all or you get halos.  It's not an elegant solution, but sharpening in post really does help.  But sharpening also brings out compression artifacts.  If Canon or ML could improve the camera's sharpening algorithm (and add focus peaking and zebras) I would be thrilled with this camera for video.

There's little chance of more resolution out of the camera, but either less sharpening or a higher bitrate (though 90Mbps seems pretty high as is in terms of SD card buffer) would make a big difference.  Clean HDMI out would be fine, but at that point, just buy an FS100 instead.  The point of dSLRs is their size and ease of use; external capture can be a pain.

Pages: 1 ... 19 20 [21] 22 23 ... 28