Page 2 of 15 FirstFirst 123412 ... LastLast
Results 11 to 20 of 142

Thread: Top-end digital concerns

  1. #11

    Join Date
    Feb 2002
    Posts
    783

    Re: Top-end digital concerns

    > I spoke with Schneider LF technical people, and it was their opinion that at 22 megapixels, one could use traditional LF lenses and not expect to gain that much by "upgrading" to the Digitar type lenses. (I'm thinking in particular of the Mamiya 22mp digital back.) They emphasized that a lot could be accomplished, size-wise and otherwise with a back like that. Above 22p, they recomended more sophisticated lenses.


    IMO, this is excellent advise, and a VERY valid point. But it must be a MF sensor, not smaller 35mm as mentioned. However, keep in mind, that LF lenses will never deliver the resolution that the Digitar lenses will.... so you will still have gains with digitar lenses...specially if the subject has very little depth, in which case you will have big gains, as the Digitars are optimized in the 5.6 - 8 range, vs. LF lenses in the f16+ range (dependent on fl). These details can make a big difference, if your type work allows it... Of course the real value of the MF back on a view camera lens, is the ability to stitch... now, you have the best of both worlds, assuming the subject is cooperative.



    > I've not found any Canon lenses to get me excited either.


    IMO, the value of DSLR is FAST lenses. Try the 200mm f2.0 if you want to have some fun. You can produce images not possible with LF. However, if DOF is your gig, LF is still the cats meow.... and large DOF is the death sentence to DSLR...it will never be optimized for such. Instead, its sweet spot is ultra fast lenses which are diffraction limited.... only the best lenses today will make these highly dense sensors produce amazing imagery.

    For some, it becomes cost effective to buy some really expensive digital set up, such as the 60MP MF back with digitar lenses....but you have to shooot a lotttt of film for this to be true. This is the beauty of LF IMO.... for a few dollars, you can match the output of a $80k digital set up....


    > I use an 8x10 camera and when I drum scan the image I have a 568 megapixel file.


    I don't mean to start a war.... and another poster above has already touched on this..... but a 568 MP scanned file is very deceiving to many readers. Your statement gives the illusion, you would need a 568 MP digital camera to match the image quality of your 8x10 scanned film.... I am sure this was not your intent, and there is a lot of variables to consider, so I won't comment on your unique application, but to put things in perspective for others..... I offer the following....using 1/R, If we consider diffraction limited lenses, at 100 lp/mm MTF for high resolving B&W film, at these f stops, the max. resolution of 8x10 film is...

    f32 - 204 MP
    f45 - 125 MP
    f64 - 72 MP

    And in the rare situation, where you can infinity focus at f22 - 468 MP


    Of course, with 8x10 color film the numbers are significantly less...using 60 lp/mm MTF of the film...

    f32 - 139 MP
    f45 - 92 MP
    f64 - 54 MP

    f22 - 204 MP

    For those interested..... 4x5 color film at 60 lp/mm

    f16 - 58 MP
    f22 - 44 MP
    f32 - 30 MP
    f45 - 20 MP


    These MP values represent the IQ of the film, NOT the file size. Often, to squeeze every last drop of resolution from film, the scanner must over-scan to grab all the resolution, creating file sizes that are enormous...... this is normal in all forms of reproduction work... and can be classified as "inefficiency" when going to a 2nd generation image. This value can vary from 1.3x to 3x based on the quality of the scanner.


    Anyway, the point is here, the "devil is in the details" when comparing digital vs. film.... there is no cut n dry comparisons.... unfortunately...


    As for the theme of initial post.... IMO, the MP wars are hitting brick walls.... next it will be lens wars to gain the benefit of those smaller pixels. A perfect example is the 50d... unless I use a near diffraction limited f2 lens (shot wide open) on it (which only one or two even exist, certainly not the 85 1.2) .... I don't gain much final rez over the previous 8MP or 10MP sensors. This is a perfect example of how most MP's are becoming marketing MP's vs. useful MP's. This makes perfect sense, and it demonstrates that the only sensible move from here, is to enlarge the sensor size, then it can accommodate smaller apt. settings, to gain more final rez AND allow smaller apt. settings. This was the move Leica recently made....I wonder if Canon and Nikon might some day end up in this camp, with a new line-up of lenses of course....


    As others have pointed out....the next big break in digital, is MF sized sensors in the 40 - 60MP size that are reasonably priced. It sure is a mystery if that will ever happen...but if it does, it will be lethal hit to film.

  2. #12

    Join Date
    Oct 2005
    Location
    Santa Cruz, CA
    Posts
    2,094

    Re: Top-end digital concerns

    Quote Originally Posted by bglick View Post
    >
    > I use an 8x10 camera and when I drum scan the image I have a 568 megapixel file.
    I don't mean to start a war.... and another poster above has already touched on this..... but a 568 MP scanned file is very deceiving to many readers. Your statement gives the illusion, you would need a 568 MP digital camera to match the image quality of your 8x10 scanned film.... I am sure this was not your intent, and there is a lot of variables to consider, so I won't comment on your unique application, but to put things in perspective for others.....
    I don't want a war either, and you are correct, I did not intend to say that anyone needed anything at that size. I'm enjoying it, but that is different from need.

    I think the most important point I wanted to make is to try it both ways. One can rent equipment, etc.

    I do expect (and hope) that when the megapixels of excellent backs get between the 200 and 468 that I can retire my film camera. I wonder about this. I have learned more about film real estate this year than I care to know.

    In the meantime, however, the 8x10 is performing fabulously. Personally, I am not that interested in sharpness. I am more interested in depth of field and very interested separation between the midtones. In this respect, the numbers mean little - or maybe the numbers I'd like are not being calculated anywhere and aren't available.

    Lenny
    EigerStudios
    Museum Quality Drum Scanning and Printing

  3. #13

    Join Date
    Feb 2002
    Posts
    783

    Re: Top-end digital concerns

    > And in the rare situation, where you can infinity focus at f22 - 468 MP


    Sorry, this was a typo.... f22 - 329 MP (not 468 MP)

    Anyway, my point was, your 568 MP file is probably equiv. to 1/2 (or less) of this value, in "real world resolution"........ The pure pixel count of a scanned file, converted to MP's is vooodo math. Scanned files are bloated with inefficiency.

    IMO, one of the most annoying issues I have with scanned film is this file size issue. With digital capture, you often produce file sizes half the the MegaBytes vs. scanned film, but yet, they both contain the same resolution. that's the reality of...

    1st generation digital capture vs. 2nd generation scanned film, bummer...

  4. #14

    Join Date
    Nov 2004
    Location
    Orange, CA
    Posts
    973

    Re: Top-end digital concerns

    Quote Originally Posted by Clement Apffel View Post
    But actually, if I start such a topic, it is because I need to deliver files for up to 2.5m large (close to 100 inches) prints. And at digital speed: meaning very short amount of time...And my quality exigency forbids me to use a stretched 35mm digital file.
    Assuming you have already ruled out a digital scanning back, and any kind of film solution is too slow, then you really have very few options here.

    I would suggest first determining what level of print resolution you can accept in terms of DPI. You can then calculate the required capture resolution in terms of megapixels. West Coast Imaging (one of the major fine art digital print labs here in the U.S.) assembled a print resolution chart correlating capture resolution with print output quality. There is of course a lot of subjectivity here, but they tried to accomodate the needs of what they call "the average photographer". Click on the following link, then click on "MegaPixel to Print Size Char":

    http://www.westcoastimaging.com/wci/...qprintlab.html

    Using the baseline or midpoints in the resolution ranges they give, WCI defines levels of print quality as follows:

    "Superb" 200 DPI
    "Excellent" 175 DPI
    "Better" 125 DPI
    "Good" 90 DPI

    Of course a lot of fine art folks here on this forum would be appalled at the idea of a 90 DPI print being considered "Good", but I think we've all seen mural sized enlargements made from a 6 megapixel camera such as a Nikon D70 which would seem serviceable to many folks (albeit rather pixelated).

    Apply these print densities to an 80 x 100 inch print, and you would need the following capture resolutions:

    "Superb" 16,000 x 20,000 = 320 megapixels
    "Excellent" 14,000 x 17,500 = 245 megapixels
    "Better" 10,000 x 12,500 = 125 megapixels
    "Good" 7,200 x 9,000 = 65 megapixels

    "Good" is equivalent to enlarging a D70 image to a 22 x 32 inch print size, which would be passable for certain commercial subjects but certainly not arty in any way.

    If you settle for "Good", then theoretically the newly announced 60 megapixel Phase One P65+ can nearly achieve the required resolution in a single shot (56 megapixel backs from Leaf and Sinar are also options). But this product is not shipping yet, and would require an investment of $50K+ to purchase new (renting presumably would be the best option). If "Good" is unacceptable, or the P65+ is not a viable solution for you, then you'll have no choice but to stitch. If stitching is feasible, then there are a whole bunch of possible solutions to choose from.

  5. #15
    David J. Heinrich
    Join Date
    Oct 2008
    Posts
    575

    Re: Top-end digital concerns

    Quote Originally Posted by bglick View Post
    As others have pointed out....the next big break in digital, is MF sized sensors in the 40 - 60MP size that are reasonably priced. It sure is a mystery if that will ever happen...but if it does, it will be lethal hit to film.
    Thank you for your informative post on LF resolution. But even if reasonably priced MF sensors 40 - 60 MP in size are produced, that wouldn't mean the end of large-format cameras, due to the possibility of movements. MF backs can of course be designed to accomodate this, but it doesn't seem to have happened much yet. One possibility would seem to be producing digital backs for LF cameras. I know I've read there's problems with this, but can't technology take care of it? (and also, couldn't the light recepticles be large on LF, due to the maximum actual resolution of LF film, thus not needing recepticles as small? this would also reduce noise).

    I imagine large digital sensors, with large individual light sensors.

  6. #16

    Join Date
    Feb 2002
    Posts
    783

    Re: Top-end digital concerns

    Eric, great post.... reverse engineering the problem is the most sensible approach. I viewed the WCI link, a few comments about the link, and your well constructed post...

    1) It's interesting the chart shows 6x7 film exceeding 63 MP back. This is a BIG stretch.... I think WCI is also confused about scanned file size (MB's) vs. digital captures (MP's).

    2) Obviously, there is a lot more to the equation than MP's, as the amount of "recorded" data is highly dependent on the lens quality, f stop used, DOF, pixel size, etc. etc. The resultant "recorded resolutions" can vary by a factor of ~ 3x. So by no means is such a chart to be relied on. But for the avg. photographer, the chart is a good starting point to get a grip on how much resolution the captured image might contain, and how he can use this to evaluate print sizes.

    3) The big missing factor, regarding the 100" print the OP has requested.... is viewing distance. A normal viewing distance is equal to a prints diagonal dimension. In which case, if you can make a print in the SUPERB range, at 16 x 20" (as an example), than each time you double the print size, you double the viewing distance. The perceived resolution will still be in the SUPERB category. This is why a Billboard can look tack sharp, even though the image was shot on 35mm film, as the image size from your viewing distance is the same as viewing a 4"x6" drug store print at arms length. When print becomes 100" long, I think it foolish to worry about close inspection, as you must be ~ 150" (13 ft diag.) away from the print to take the print in (visually). This "diagonal viewing distance" represents approx. a 55 deg FOV (Field of View), a very comfortable viewing angle for humans.

    A viewing distance of half the print diagonal will produce a 110 deg. FOV. A human has approx. a 100 deg. stereoscopic FOV without moving the eyes or head. So I would suggest the viewing distance should be somewhere between the print diag., or if you really push the resolution envelope, than 1/2 the print diag. Any closer than half the print diag. the person can not see the entire print without scanning their eyes / head.

    This provides a more realistic representation for the starting point of "how much capture resolution" is required for a print with a known diagonal. The numbers above are based on prints with aspect ratios of Square (1:1), to 1.25 : 1 (4x5). Adjust accordingly for different aspect ratio prints.

    > One possibility would seem to be producing digital backs for LF cameras. I know I've read there's problems with this, but can't technology take care of it? (and also, couldn't the light recepticles be large on LF, due to the maximum actual resolution of LF film, thus not needing recepticles as small? this would also reduce noise).


    MF backs already work with view cameras.... often with smaller view cameras that are higher precisioned for digital work... such as the Linhof 6x9, Toyo, Sinar, etc. You get the benefits of movements, if the lens has a big enough image circle, which will surely be limiting, at best. Many of these backs can also be retrofitted to fit 4x5 cameras, IIRC, many already have adapter plates for 4x5. I doubt they will make a digital back the size of 4x5, which I think was your question.....the reason is.... you will have to use longer fl lenses, which require higher f stops, limiting the potential resolution.... Anything is possible, but considering stitching is an option today (by moving the back around the image circle), it only leaves a single application for what you describe, non-static scenes. In which case, the single shot 60MP MF back comes close enough to 4x5...and who knows what next years MP will be? So I can't see any motive for such a product...but it sure would be nice....


    This is pure speculation, but if high resolution photography follows the lead of high rez astro photography, where resolution is REALLY at a premium.... then possibly the next generation of high rez still photography will consist of combining 4 digital sensors into one back. There will be 4 lenses, one for each sensor - all capture a different part of the scene and automatically stitched together as one output file. The shutters will be LCD shutters built into the lenses, so sync will be perfect, as its all electronic. An LCD shutter would appear like an lens element, it will produce no aberations. With power applied to the LCD glass, it will block light, or allow it to pass through, just as a mechanical shutter does. The value here is, since its electricity, the response times are ultra fast. This is required for accurate sync. These shutters are in use today, mostly for studio photography where fast shutter speeds are not used.


    The benefits are, you can still have large clean pixel sizes, high resolution, AND very fast, and very sharp lenses (due to their shorter fl's), even when DOF is required. Think of this as, "auto stitch". So this puts 100 - 150 MP into reality, with NO new advances in digital sensors.... considering Sony sells a 24MP camera body for $3k, the sensor can't be more than half this.... as you can see the possibilities are there. The missing link is.... a camera maker flushed with cash to R&D such a project. I think this is out of the scope of the Canons and Nikons market, so unless its a Leica, Zeiss Hassy or Phase One, I doubt we will see it. The real kicker is the LCD shutters, while they are fast and reliable, they are still a bit slow, IIRC, only 1/125 th max. shutter speed, and they draw a decent amount of power...but they are at their embryonic stage of development. A better solution is the sensor acting as the shutter..several companies are working on this now....then mechanical shutters will be history. Back to earth.....

  7. #17
    David J. Heinrich
    Join Date
    Oct 2008
    Posts
    575

    Re: Top-end digital concerns

    If the sensor acts as a shutter, then theoretically, each pixel could act as a shutter; you could have an HDR image quite easily.

  8. #18

    Join Date
    Apr 2006
    Posts
    135

    Re: Top-end digital concerns

    Quote Originally Posted by dh003i View Post
    If the sensor acts as a shutter, then theoretically, each pixel could act as a shutter; you could have an HDR image quite easily.
    I'm in love with the S5, I think they use two photo-sites for each pixel to get good HDR with "unsurpassed color fidelity". Any thoughts about that design ?

  9. #19

    Join Date
    Apr 2006
    Posts
    135

    Re: Top-end digital concerns

    Quote Originally Posted by bglick View Post
    > With digital capture, you often produce file sizes half the the MegaBytes vs. scanned film, but yet, they both contain the same resolution.
    Do you see any value in comparing the two by ever shooting cy/mm targets and comparing results ?

  10. #20

    Join Date
    Apr 2006
    Posts
    135

    Re: Top-end digital concerns

    Quote Originally Posted by bglick View Post
    > .. The pure pixel count of a scanned file, converted to MP's is vooodo math...
    SOoo... You have proved this because the Signal to Noise Ratio of both confirms this ?

Similar Threads

  1. The LL Digital Field Camera Experiment has Ended...
    By Eric Leppanen in forum Digital Hardware
    Replies: 11
    Last Post: 28-Aug-2007, 23:41
  2. Existing Light Guide available for download
    By al olson in forum Announcements
    Replies: 12
    Last Post: 7-Dec-2006, 17:27
  3. Why digital?
    By paul owen in forum Digital Hardware
    Replies: 26
    Last Post: 27-May-2002, 11:45
  4. Digital Darkroom Needs
    By John Miller in forum Digital Hardware
    Replies: 8
    Last Post: 15-Aug-2000, 01:30
  5. 4x5 best optics w/ Scheider HIGH END BACK sharper than 8x10?
    By Bill Glickman in forum Cameras & Camera Accessories
    Replies: 11
    Last Post: 17-May-1999, 04:31

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •