what is the best theoretical maximum dpi for film?

[jimbobuk]

Just wondering if anyone knew any research into this? Obviously this will vary between film types, so lets say we're talking about the finest grain film currently available (i'm not sure what that is, but i assume its around ISO50 🙂)

My epson 4490 is clearly going to be the bottle neck on some scans, i now scan at 3200 dpi as this is mostly very viewable at 100%, it could probably benefit by being the next resolution down as in viewing there is rarely any NEW information added to an image from around 50-70% compared with 100%.. don't know if thats scientifically valid or not.

Anyways it'd be nice to know what the alleged physical maximum of the grain of modern film.. i've heard of rumours of 35mm performing liked the MF of old, and also that 35mm offers around 20 megapixels of good quality pixels... is that about right?

Cheers

Jim

 

[sf]

Theoretical? You mean what is the physical limit for the grain, regardless of what is marketable? Who knows. They know, but they'd never tell us.

Velvia 100 is very high. Bluefire Police film is even higher resolution. What is the value of this information to you?

You also have to consider the limitations of lenses and bodies and development. To get the lowest grain, you're going to have to give up other image quality characteristics. And you're going to have to pay through the nose for scans that capture that resolution.

 

[jimbobuk]

no its purely an exercise for my mind in absence of my wallet getting involved.. I have the Epson 4490.. I am adopting 3200 dpi as my scan limit at the moment, yet mostly things look better viewed at 70% of that. I was wondering if anyone knew roughly purely to put a perspective on things.. scanner manufacturers clearly lie about their maximum resolution, yet the tone of reviews seems to suggest that dedicated film scanners lie less so (and typically they offer a lower maximum resolution) ..

Knowing the ultimate maximum of the media itself (in its highest performing mode of say ISO 50.. i'm not saying i'd even consider switching to ISO 50 for that reason.. 400 seems a good fit for now for the slower lenses of the RF645), even roughly puts a perspective on things as far as going for a scanner in the future that sports say a true 4800 optical resolution..

I guess getting some scanner test charts and performing the scans would help solve this problem but i just thought i'd ask.

Cheers.

 

[BrianPhotog]

I can't find the link now, but there's a bio-film (based on engineered bacteria) in development by some company out there. Due to the size of the bacteria, it has a resolution on the order of hundreds of times greater then current film or digital offerings.

That would be the known limit of film resolution. Theoretical? Are we involving Quantum physics or sticking to the strait Einsteinian & Newtonian stuff? 🙂

 

[jimbobuk]

Basically i'd like to know when in a hypothetical future can i stop wondering if the scanner is not performing at the resolution it claims, rather the media itself having been exhausted of anything meaningful as far as the contained image is concerned.

Urgh.. scanning complete, far too late.. bank holidays aren't just made for wasting.. sleep for me i think... Laters

 

[philbond87]

I run a Linotype Hell 3900S drum scanner and I have run some tests of my own using RDP III (Fuji Provia).

Now, I have heard various claims that one can get useful data out of film at upwards of 6000 dpi. As that makes for enormous file sizes and scanning times (especially when I'm scanning my 4x5s) I thought I'd run my own tests to see at what point I could discern an appreciable difference in image detail.

As was previously stated, a lot depends on the quality of the lens used. But based on my meager (50 yr old!) equipment, the difference in image quality between 5000 and 6000 dpi was not discernible to my eye. In fact, the difference I could see between 4000 and 5000 dpi was almost negligible.

Nowadays I scan all of my film at 4000 dpi.

Again, this is for a Hell drum scanner and I can't speak to what you can get out of your own CCD scanner. What I would recommend, however, is that you do what I did and run your own tests.

Regards,
Phil

 

[jimbobuk]

I meant regarding the current films available.. though uber bio film sounds impressive... that really would kind of herald 35mm == Large format of old... could digital even consider attempting to fight off large format in that media, when its already lacking in pure resolution with existing film technologies at that size...

 

[Jason Sprenger]

Norman Koren put some of his thoughts on a similar subject on his web site.

http://www.normankoren.com/Tutorials/MTF7.html

Mr. Koren's data set is dated but his calculations and argument might prove to be food for thought in this area. His benchmark for film in his comparisons was Provia 100F. Not the highest resolution film but an excellent resolution color slide film.

Personally, I find a 4000 dpi scanner adequate to retrieve images from my color film photos right down to the grain. To be practically better, film technology would need a new leap ahead. Not necessarily in terms of raw resolution but rather in terms of resolution versus apparent grain.

 

[ClaremontPhoto]

If you're viewing onscreen it is pretty much always 72dpi. Personally I size to 480px wide, although some people say it's too big and some people say it's too small.

If you're printing on paper at home then 300dpi is fairly standard. So for a 150mm print you need about 1800 dots wide on 35mm film. Gets confusing doing inches and metric doesn't it? So say about 1200 dpi.

If you want bigger and better then you've still got the negs you can get whatever you want in a lab. I mean, can you do ultra large at home on your printer? A friend of mine can go to 60" (damn non metric again) but her printer cost about the same as her car.

 

[sf]

no its purely an exercise for my mind in absence of my wallet getting involved.. I have the Epson 4490.. I am adopting 3200 dpi as my scan limit at the moment, yet mostly things look better viewed at 70% of that. I was wondering if anyone knew roughly purely to put a perspective on things.. scanner manufacturers clearly lie about their maximum resolution, yet the tone of reviews seems to suggest that dedicated film scanners lie less so (and typically they offer a lower maximum resolution) ..

Knowing the ultimate maximum of the media itself (in its highest performing mode of say ISO 50.. i'm not saying i'd even consider switching to ISO 50 for that reason.. 400 seems a good fit for now for the slower lenses of the RF645), even roughly puts a perspective on things as far as going for a scanner in the future that sports say a true 4800 optical resolution..

I guess getting some scanner test charts and performing the scans would help solve this problem but i just thought i'd ask.

Cheers.

Your scans look better at 70% because flatbeds are invariably soft compared to the actual sharpness of the film. It looks better at 70% because the resolution rating of the scanner is way beyond its optical capability. You're resizing it to maybe where it belongs.

In all honesty, you have no use for super high res film, or even to care about film grain size because you're not scanning with a scanner or printing with an enlarger that cares. You will see no difference in grain between Neopan Acros and Tri-X.

Go to my scanner test to see what I'm trying to say :
http://www.shutterflower.com/scanner comparison.htm

Flatbeds just don't care about grain. Look at the Epson 3200 on Velvia and Fuji 160s. There is a BIG grain differnce in reality, but the scanner doesn't see that far down.

 

[varjag]

Zeiss recently demosntrated that best optics with finest emulsions gives resolution beyond the current crop of scanning or projecting equipment capabilities. You can enjoy full quality of that combo only under microscope.

Which just underlines that beyond certain quality level all that resolution talk gets fairly pointless.

 

[sf]

Zeiss recently demosntrated that best optics with finest emulsions gives resolution beyond the current crop of scanning or projecting equipment capabilities. You can enjoy full quality of that combo only under microscope.

Which just underlines that beyond certain quality level all that resolution talk gets fairly pointless.

You can only get so much resolution on a film.

FOr instance, if I can see the grain in my Velvia 100 on my Multi Pro, I've reached its maximum useful resoltuion. Beyond that point, I'm just going to get magnifications of the grain structure and no further detail in the image. This is the real issue here. This and the fact that flatbeds are very limited in their capacity to harness anything but tonal range - and even that is limited.

If you're scanning at 5400dpi or with a drum scanner, you're going to get down into the grain of the very finest emulsions within the reach of any photographer. SO, really, the best optics and best film is within the reach of the fiew with the money to use that resolution. I think that someone with a Minolta 5400 or Nikon 35mm scanner or a drum scanner is going to be able to get the very highest resolution film and actually make use of its resolution.

 

[varjag]

shutterflower, copy films were doing 600 lp/mm decades ago. I don't understand why you hold Velvia as a gold standard for resolution.

http://www.zeiss.com/c12567a8003b58b9/Contents-Frame/098729cea9ef2428c1257118005d1d20

 

[Socke]

varjag, some like colour and others want greys in their B/W 🙂

 

[varjag]

Well, the thread was about maximum theoretical resolution on any film, if i got that part right. If you care to follow the link in my post you'll find out that Zess exceeded 320lp/mm on continuous tone grayscale reproduction 🙂

 

[Bertram2]

Just wondering if anyone knew any research into this? Obviously this will vary between film types, so lets say we're talking about the finest grain film currently available (i'm not sure what that is, but i assume its around ISO50 🙂)
My epson 4490 is clearly going to be the bottle neck on some scans,
Jim

Sharpest film:
There are several things mixed up here: Kodak TP makes about 380 lp/mm at 100% contrast, not the right solution for all who want to have some greys too in their photos.

Scanner:
The 4490 is indeed the bottleneck, if you are not contended with the performance buy a decent dedicated film scanner like a Nikon V ED. It is not solely all about res tho, it is also about HOW the info is recorded.

That's not really necessary tho for monitor pics, excepted you are a pro and your website is your business card.
I scan all negs at the full res of 4800 and then resize it to the monitor size with PS later.

FP4 and 4800 , Epson 4490

Printing from scans is another issue, take the best film scanner you can afford.

Films:
Nobody needs more than he can see, the human perception ends at about 120 lp/mm on a 20X30cm enlargement . If you have eyes like an eagle. For some it all ends at 80 lp/mm already.
Most of the films have from 80 good 400) to 200 (good 25ISO B&W).

Now make your own conclusions ! 😉

bertram

 

[jimbobuk]

Thanks for all the links and the discussions guys, really nice food for thought. I'm happy to stick with my 4490 for the near future, perhaps i'll get the v700 or v750 in a little while as I do plan to experiment with some large format film at some point so the flat bed is neccessary to leave that option open.

I'll get my 1st medium format home developed black and white images online soon.

 

[jimbobuk]

shutterflower, nice review there.. I remember reading about you doing the review, didn't realise the results were up.

The Epson 4870 is superior to my 4490 isn't it? It doesn't look like its even competing in your test, I didn't realise the potential gain was possibly so great. I will be good, and hold onto my wallet 🙂