« How Many Megapixels? | Main | 53rd Annual World Press Photo Winners »

Tuesday, 16 February 2010

Comments

Slightly OT because comments are blocked on the "Too many Mega-pixels" post.
In that post, either you have not given enough info or not allowed us to add info. e.g. sensor size, it could make a difference.

Cheers Rex

I wonder how how many LF lenses will be able to do the film justice? I suspect it is a small club.

Side question: Is the current Ektar the same RMS granularity as the 25?

Patrick

Although I won't be using this particular film, I am very happy to see that Kodak continues to support large format. A lot of us film shooters live under the shadow of waking up one day and finding out that our favorite film has been discontinued. (It's happened to me a couple of times in the last few years.) Hopefully this means that Kodak is making money in the analog business and will keep making their other large format films, too.

Rex,
I'm saying you get to specify EVERYTHING about the camera. Whatever you want, you got. Then just tell me how many MP it has. That's all.

Mike

If I recall correctly, Velvia 50 tops out around 80 lp/mm, and presumably the Ektar is in the same ballpark. That would get you (at best) something like 1-2 Gigapixels for an 8x10 sheet. Not that you'll get that in practice of course. Check out gigapxl.org for an interesting discussion of some issues in really trying to get 1 billion pixels (in the technology section).

This is great news!

Well, it's great news that there's a new and possibly excellent color 4x5 film from Kodak, how great the reality is will depend on how easy it'll be to get the stuff developed.

Time to haul out the LF cameras and lenses!

"Eventually, as film technology continued to improve, Kodak didn't see any disadvantage any more in a version that was four times faster."

Correct me if I'm wrong: The original Ektar was 25. ISO 50 would be twice as fast, 100 three times as fast. Or in other words: twice faster than the original, not four times faster.

Apologies for the nitpicking :o)
Alex.

"World's Finest Grain"... in sheet film. Hmmm.

And we're arguing about a few crummy megapixels in a new digital camera...

Are they insane? Don't they know film is dead?

Onya Kodak, even though I much prefer Ilford and Fuji. However, as a budding medium format user I might be tempted to try that E100 for landscape work. (would love to have used the 25, but then I always was an obstinate SoB).

In other news, I chose the 12-14 option in the poll. MUCH would have preferred a 10-12 option. Now, when can I pick up my camera?

(Translation - 'onya' is Austalian-ese for 'Good on you')

Good news. I just purchased a Graflex so I'm looking forward to seeing what the sheet version looks like.

I agree with Rex re: the megapixel post, that is, sensor size makes a difference here. I voted "12" because that's the *density* I'd want on a 35mm FF sized sensor, which seems to be 1.0 on the "normalized" sensor size scale. However, what I'd really want is a 6x7 sized sensor and therefore 48 megapixels or whatever it works out to. But I didn't want to imply that I'd want 48 megapixels on a smaller sensor. (And similarly, on APS I think 6-8 is about right).

Alex,
No, 25 to 100 is two stops, or four times as sensitive. (A stop is a doubling or halving.) Thus four times as fast.

Mike

Looks like about 1 Mpix/sq inch to me, based on a very rough analysis of some hi-res scans I found via Google

http://www.boeringa.demon.nl/menu_technic_ektar100_grainstructure.htm

I was quite pleased to see that I was within an order of magnitude of Ctein's number.

(you'd think EK would quote an actual grain size if they were going to go around claiming "the world's finest grain").

Dear David,

80 line pair per millimeter is a plausible number for what you'll get out of Velvia in an untuned-up 35mm camera, but the film's limiting resolution is 160 line pair per millimeter. Ektar 25's resolution was in the neighborhood of 200 line pair per millimeter (in-camera tests could squeeze out better than 150 line pair per millimeter, which is pretty damned amazing). Your average low- and medium-speed color film today hits 125-150 line pair per millimeter, so I would not expect Ektar100 to do any less than that. It might do more.

You'll never actually realize anything like that in a view camera set up; this is just FYI.


~ pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 
=====================================

Scott Sheppard has an interview with Scott DiSabato from Kodak about this new release on his podcast, at www.insideanalogphoto.com. You can download it in iTunes or from his RSS XML feed, if you don't do iTunes.

HOT DAMN THAT'S GREAT NEWS. It's way too punchy for me, but I bet it'll be just about right with single coated lenses. ;-)

If we're talking in bayer array pixels, then an 8x10 sheet will have approx 300-400 Mpixels. However if we're talking Foveon pixels then Ctein's 150-200 is more like it. In other words you'll get 150-200 pixels of fine line resolution but you'd need 300-400 pixels to match the beautifully accurate high frequency colour detail. I spent a little time thinking here... http://www.timparkin.co.uk/blog/large-fomat-resolution

Ctein, I'm no expert here; just curious. So, when I read an article like the attached...http://www.luminous-landscape.com/essays/Cramer.shtml ... what, if anything, can or should I conclude in relation to your featured comment? Or, is this apples and oranges?

If you shoot near wide open with the best lenses on 8x10, there is an outside chance you could approach the 80 lp/mm number. If you could, that would be 16256 x 20320 line pairs. Because you need 2X the samples to resolve a frequency (Nyquist), you would need a "perfect" sensor (think Foveon) of 32512 x 40640 pixels. That's 1.321 Gigapixels.

Of course, if the sensor has a Bayer pattern, you can count on needing at about 1.5X the number of pixels to actually achieve that many line pairs, which means you would need a 2 gigapixel digital sensor to equal today's 8x10.

The cool thing about analog sensors is that they're cheap, disposable and replaceable.

Dear Tim & Jeff,

Just want to be clear I did mean Bayer array. (Foveon's a dead-end, niche technology; it doesn't even enter into my future-think any more.)

But mostly, Jeff's right-- it's apples vs oranges. I'd not even have brought up the subject if Mike hadn't opened the can.

Charlie Cramer's article (which I hadn't been thinking of when I made my comment) is pretty consistent with what I wrote. There are a couple of factors which make extrapolation complicated. First is, as Charlie's examples illustrate, there are some kinds of detail that digital renders better and others that film renders better. If you demand that one medium equal the other for all types of subject detail, you end up with an overall mismatch. That is, if you set your pixel standard as being the one at which digital renders fine detail as well as film, for the kind of fine detail the digital is worst at, then you wind up with a situation where overall the digital file exhibits much higher resolution than the film.

Putting it succinctly, analyzing the worst-case scenario isn't a good guide to determining the comparable case.

In any case, Charlie notes that overall his 39 megapixel digital file wasn't quite as sharp as his film file, but it was close. You can be pretty confident that if you doubled the number of pixels, overall the digital photo would be at least as sharp if not sharper.

The second confounding factor is that the number of resolved picture elements in a film photograph doesn't scale with the area of the format. It's closer to the square root. That's just a mechanical engineering rule of thumb, all those resolution-confounding factors I talked about in the featured post. So, for example, one of my 6 x 7 cm negatives probably contains twice as much image detail as my 35mm ones, even though it has four times the area.

Similarly, 8 x 10 view cameras will record something more like twice the number of film "pixels" as 4 x 5 view cameras, not four times the number.

Lots and lots of windage involved in these estimates, but you put it all together and Charlie's results are pretty consistent with my assertion that 150-200 megapixels puts you in the same realm as quality 8 x 10.

Incidentally, this apples and oranges problem isn't unique to digital versus film. Kodachrome had unusually high cyan-layer acutance. Far above what you expect from its resolution. Consequently, long after other films had eclipsed it in overall resolution, you could still find Kodachrome partisans who claimed it was the sharpest film around, based on photographing things like bare tree limbs against blue skies. They were correct, in a very limited sense. But it didn't invalidate the overall judgment that the newer films were sharper and higher-resolving than Kodachrome 25.


~ pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 
======================================

Whats'da matter with you guys. Given the choice of giganormous megapixels and you're settling for 16MP?

Having gobs and gobs of pixels means that all the faults along the optics path becomes a useful low pass; the actual anti-aliasing filter can get dropped. Gobs and gobs of pixels also means that the (already small) resolution loss due to the Bayer mosaic becomes truly inconsequential.

Gobs and gobs of pixels means that we can go back to debating the truly important - why my Industar 50 is just as good as your Leica.

Regardless of the claims on the box, any new film release is good news. I'm sure the types of people using this film won't think twice about the marketing department's claims to 'World's Bestest Most Sharpest Colour Negative Film Made by Kodak Since December 2009...EVER' etc. It looks like Ektar in all forms is going from strength to strength; shame it's still not available in the antipodean land of Oz.

If we are really nitpicking, 100 is four times "as fast as" 25, but three times "faster". It's like the news items that lazily say something that is now 1.2x what it was before, represents "a 120% increase". It's an increase of, or by 20%. It's also an increase to 120%.

(Bah mode off - hurriedly scans own text for grammar and and spealing)

Always be asked this question to digital photographer when I am doing my 8x10 shoot.

May be I summerise the basic agreement so far is that the raw max resolution is 1.3 Gpixel (3 color per pixel calculated is like the Fevon sensor). But that assume a lens up to 80 lp/mm. The lens is limiting factor and if Ctein is right you can assume that may be because the lens give you 20 lp/mm over the 8x10 area, i.e 1/16 (due to square calculation) of 1.3 Gpixel i.e. about 100 Mpix (3 color).

In real life of course, the looks of Velvia 50 8x10 and a Black and White 8x10 (Contact Print) is so different even my digital SLR friends never want to or never have touched any film found my poor sample of these is very hard to compare with theirs. It is really apple and "dragon-eye". The issue as most of people have never seen a 8x10 Velvia 50 slide, just like many of you have never seen any "dragon-eye" fruit. It is beyond most people for this comparison.

Also, these day even carry around my D300 is too big (compared with my M8) and I saw many people starting to buying GF1, you have to in it for something for taking 4x5 and 8x10 or even taking your Hassey and Rolleiflex TLR. Some picture really want to ask for a large format but some for a camera next to you, even just an iPhone. It is a waste of time to compare.

But as one said, it is lovely to know that you have an option.

The key is that the new film give you an option and thanks Kodak for Ektar, Fuji for Velvia 50 and IIford for HP5.

@ RL: "Whats'da matter with you guys. Given the choice of giganormous megapixels and you're settling for 16MP?"

Er, yep. I gotta 12" X 18" print here from a 14 MP camera. Sharp as yer like! With giganormous megapixels (and reasonababbly sized sensor sites) comes da giganormous heavy camera, which I would have to lug around. Like a sheet film camera, (Yep, I'm nearly on topic) or a Zenit B.

Okay...There`s been a lot said here that that does not translate to what a finished print is going to look like to me. Currently I`m using a D-700 and 13x19 printer for color work. B&W is usually 5x7 neg. enlarged no more than 3x. My question is, what kind of print quality could I expect from an Ektar 100, 4x5 neg scanned on an Epson 750, then printed at 13x19, compared to the prints from the D-700 and the 5x7? While the D-700 is impressive for what it is, medium format digital would likely be more suitable for what i do. Just can`t afford it (sigh)

Cool! I've never shot color in sheet film before, maybe I should give it a try (the one film camera I've kept around, other than decorative antiques, is the old Omega D 4x5; I've used it very little, but it does enough things differently that I've felt it's worth keeping around).

Nice to see that film is not quite dead yet in the day and age of digital, being a "filmer" myself I'm quite happy to see a new film introduced. I primarily I shoot shoot in black and white with Kodak's relatively new T-Max 400, which I think is a great film.

"Foveon's a dead-end, niche technology"

That's too bad; I really, really like some of the results I've seen.

Four times faster, four times as fast, four hundred percent faster. . . who cares?

If we keep it to camera-speak we can all get along:

100 is two stops faster than 25.

:)

Dear Robert,

Oh, I'm not saying it will go away; don't worry! It's just that I'm pretty confident the technology will never progress to the point where you'd be seeing 100+ megapixel Foveon sensors. Which removed it from the 'future-tech' considerations, here.

pax / Ctein

The comments to this entry are closed.