Leaf backs at 28MP, 35MP

[BJL]

I have not found anything at Leaf's site yet about those new 28MP and 35MP backs, but this seems to mean that Dalsa is joining Kodak with new "645" sensors of higher resolution, in the sense of closer pixel spacing.

Dalsa is my guess this time since Leaf is currently using Dalsa sensors, and the pixel counts are a bit different than for the forthcoming Phase One backs, which will use Kodak sensors. 35MP sounds like 7 micron spacing with Dalsa's current largest 36x48mm format.

Does anyone else wonder how well 645 format lenses will keep up with this reduced pixel spacing, given their historical design for 20lp/mm or 30lp/mm and generally lower measured MTF at 40lp/mm than 35mm format lenses? That 7 micron spacing corresponds to 50lp/mm or more. Maybe Fuji in particular anticipated this when designing a whole new series of lenses for the Hasselblad/Fuji H series?

[dazzajl]

I have not found anything at Leaf's site yet about those new 28MP and 35MP backs

I was talking to someone at Leaf this week who told me they would be announcing some new products next week so perhaps some details will come out then.

However, Leaf are not great for getting info up on the web so it may be a little longer untill there is some web resource to look up.

I would certainly agree with your concerns about current lenses being able to "keep up' though.

[izaack]

http://www.pdnonline.com/pdn....1221306

[BJL]

izaack,
    thanks for the Photo District News link. Those incorrigable medium format sensor makers keep forcing 4:3 shape on us!

dazzajl,
    I am actually not so much "concerned" (cannot vaguely afford a MF back!) as just curious to see where the lens-imposed resolution limits will fall. Currently I am betting on "no more resolution than a 20-40MP Bayer CFA sensor can give", except perhaps in some wide angle extremes. Pixel counts might get somewhat higher for other reasons, like avoiding moire by "oversampling".

[michael]

The word on the street (what Kodak and Dalsa are saying privately) is that medium format digital backs will likely top-out at about 50-60MP some time in the next two years. That implies around 5 micron pixels, which is as small as current technology can go without serious noise issues.

The real problem is that most lenses aren't up to the task. I was stunned recently when I saw what a Schneider Digital lens can do compared to my Zeiss glass on the Contax. It is increasing becoming apparent to me that the limiting factor is no longer imagers, but lenses.

This applies to 35mm in spades. The 1Ds MKII at 16MP is pushing Canon glass to the limit. There are only a handful of Canon lenses that are a match for it. If Canon goes to 22MP next year (which some people are guessing they will) this is also likely to be the limit, due to pixel size, and also due to the inability of current glass to keep up with the sensors.

Michael

[BJL]

Michael,

   thanks for the industry inside comments. For Kodak, that 60MP fits with their chosen "digital 645" format of a bit over 36x48 and their current smallest FFT CCD pixels, the 5.3 microns in their 8MP FourThirds format sensor. So I can imagine they are already working on that roughly 60MP, 5.3 micron sensor design, and do not plan further pixel size reduction for the big format sensors.

Above all, thanks for the confirmation about lens limits on usable sensor resolution. Getting significantly more resolution on the final print than given by say 20MP seems solely the territory of high quality primes; nothing more than a return to the traditional situation of high resolution photography, be it in large, medium, or 35mm format. Zooms are only for the other 99% of us!

By the way "significantly more resolution on the final print" to me means something like going one step up the traditional print size scale, like 8x10 to 11x14 or 11x14 to 16x20. Or A4 to A3, etc. That is, about doubling the print size with the same lp/mm of print resolution so the bigger print sustains the same close scrutiny. And so needing roughly double the pixel count. Is there a significant visible resolution difference to be seen between 12.5MP and 16.5MP, or between 16.5MP and 22MP? I guess not, outside the realm of pixel peeping.

[Ray]

BJL,
We both know that close-up viewing of a large print can reveal fuzziness and/ or pixelation, yet photographs on the wall somehow invite us to get close to 'appreciate'  the detail.

Paintings that people hang on their walls, whether originals or prints, tend to be much larger than the everage photograph. We expect visible brush strokes and almost unrecognisable forms when we go close to a painting, especially an impressionistic painting. However, such close-up detail is not uniform and 'machine like'. There's great variety in those brush strokes.

When we approach a large photograph, say a billboard or advertisement  which looks pin-sharp from across the road, we see a uniform, evenly spaced quantity of precisely formed circular dots. It's obvious that the image has been produced by a machine. We are suddenly confronted by the impersonality and machine-like qualities of the process, and that's a bit off-putting.

As I see it, large format cameras simply allow us to make larger prints that can withstand that close scrutiny and still maintain a close representation of reality (or the photographer's concept). After all, when we get close to, and peer at a 'real' object, it doesn't suddenly become pixelated or appear to be composed of dots.

[BlasR]

Hello to all. I have a question about a digital back in the rain.
Michael say about some one in the past got back to he's car because of the rain.
what about the p25 in the rain, will be ok to stand outside rainning and take photos with the back?
or the back is not for rainning days?


Thank You

BlasR

[eleanorbrown]

Hi Blas, Recently i photographed in moderate rain using a Contax with a P25 back.  no problems. eleanor brown

Hello to all. I have a question about a digital back in the rain.
Michael say about some one in the past got back to he's car because of the rain.
what about the p25 in the rain, will be ok to stand outside rainning and take photos with the back?
or the back is not for rainning days?
Thank You

BlasR
[a href=\"index.php?act=findpost&pid=51084\"][{POST_SNAPBACK}][/a]

[michael]

Light to moderate rain and snow is not an issue for any camera gear. Photographers are far more paranoid about this than they need to be. If you can take it, it's likely that your camera can as well.

If I'm going shooting in the rain I just put a small hand towel in my back pocket to wipe off excess rain or snow, and a microfiber cloth to clean the UV filter (which I use when it's raining).

Michael

[BlasR]

Thank you...I ask because p25 have three opens space in the back, and canon don't . I was little concern about it.

BlasR

[michael]

I've been bugging Phase One for plugs for these sockets, but they have failed to respond.

What I do is simply put some black electrical tape over the holes when I'm shooting in bad weather.

Michael

[BlasR]

I plug the back with some part of my ipod   ..the  plug the firewire or USB it perfect for the back,
but I guess I will be using the black tape for the side...Thanks for the tips...
maybe you like to do your self  with the plug the come with the apple mouse to cover the USB cable
I used the one from my ipod but is the same


BlasR

[BJL]

Ray,

   I am not quite sure how your comment relates to mine above, but here are some observations of mine about the way the people view really large sharp prints, like 4'x5' prints (that's feet) made from 4"x5" format film, and how it relates to resolution needs.

Overwhelmingly, people do "view bigger" than with smaller print sizes, meaning close enough that the image has a larger apparent size --- but mostly to a limit of viewing distance not much less than the short dimension of the image, or about half of the traditional "normal" viewing distance of roughly image diagonal length. A minimum of 3' or so from those 4' by 5' prints seems almost universal, certainly not the ten inches to one foot envisioned by talk of the need for 300ppi printing regardless of print size.

I rather suspect that between diffraction limitations with smaller apertures and DOF limits with larger apertures, it would be very rare for sharpness to hold up under substantially closer viewing than "half normal", except for a few panoramas of large, distant glories. Format and pixel count have no effect on the balance of diffraction spot size with circle of confusion sizes; the same balance comes at equal aperture diameter and so at different aperture ratios with different formats.

[Ray]

Format and pixel count have no effect on the balance of diffraction spot size with circle of confusion sizes; the same balance comes at equal aperture diameter and so at different aperture ratios with different formats.
[a href=\"index.php?act=findpost&pid=51308\"][{POST_SNAPBACK}][/a]

BJL,
Been travelling in Nepal, Tailand and Cambodia for the past few weeks so a technical discussion on Airy discs from an internet cafe will help keep the grey matter active.

"The same balance comes at equal aperture diameter..."?? Okay! But supposing we ignore DoF for a moment and consider the balance between diffraction and all the other aberrations. At the plane of focus, the larger format, given sufficient pixel count, will always produce greater resolution, assuming we are using state-of-the-art lenses.

To put it another way, f64 with an 8x12" camera produces a similar DoF as f8 with 35mm, but the 35mm lens at f8 is about as sharp as you can get. There's a balance between diffration and the other lens aberrations. Very few 35mm lenses are sharper at smaller apertures than f8 but large format lenses continure to get sharper as the aperture opens up from f64. Of course, the DoF is reduced as it is with 35mm at apertures bigger (f stops smaller) than f8. However, if you want higher resolution, at least you can get it with the larger format although a tilt mechanism will be necessary to maintain good DoF.

[BJL]

BJL,
Been travelling in Nepal, Tailand and Cambodia for the past few weeks ...
[a href=\"index.php?act=findpost&pid=54040\"][{POST_SNAPBACK}][/a]

Sounds like a good start to life with the 5D and 24-105 f/4! We both need to start posting some pictures, don't we?

"The same balance comes at equal aperture diameter..."?? Okay! But supposing we ignore DoF for a moment and consider the balance between diffraction and all the other aberrations.
[a href=\"index.php?act=findpost&pid=54040\"][{POST_SNAPBACK}][/a]

I agree that at low f-stops, aberrations are an important factor in determining how much "spatial detail" a camera can record. I think it is best to use the technical phrase "angular resolution of the subject", to avoid confusions when measures like lp/mm or pixel spacing are used in comparison between formats. I will shorten that to just "resolution".

Ignoring DOF makes no sense to me: the circles of confusion at every point not in the exact focal plane limit resolution just as much as diffraction spots and the smearing by aberrations. And view camera motions do not really increase DOF: they only convert the region of adequately sharp focus from being a thin slab perpendicular to the lens axis to being a thin wedge at some oblique angle to the axis. It is no panacea for inadequate DOF, only dealing with the situation where all the elements of which one desires a sharp image lie roughly in an inclined plane. (And to avoid this discussion becoming excessively theoretical, I note that in practice it is extremely uncommon for DLSR users to use view camera motions like tilt, and I doubt that tilt lenses will ever have much effect on mainstream choices of interchangeable lens camera systems.)

Without further study of  all I can see is the likelihood that optimal control of aberration limits one to apertures no larger than f/4 or even f/5.6, and thus diffraction limited resolution of about 150-200lp/mm at 50% MTF (formulas at Norm Koren's site). This is turn is about the resolution possible with the dominant Bayer CFA style sensors with pixel spacing of about 2 microns, or about 3 microns for X3 (Foveon style) sensors. Darn, no more than about 6000x8000 pixels for Four Thirds before I have to move to digital large format like 36x24mm or 48x36mm!

[Ray]

Ignoring DOF makes no sense to me: the circles of confusion at every point not in the exact focal plane limit resolution just as much as diffraction spots and the smearing by aberrations. And view camera motions do not really increase DOF: they only convert the region of adequately sharp focus from being a thin slab perpendicular to the lens axis to being a thin wedge at some oblique angle to the axis. It is no panacea for inadequate DOF, only dealing with the situation where all the elements of which one desires a sharp image lie roughly in an inclined plane. (And to avoid this discussion becoming excessively theoretical, I note that in practice it is extremely uncommon for DLSR users to use view camera motions like tilt, and I doubt that tilt lenses will ever have much effect on mainstream choices of interchangeable lens camera systems.)


[a href=\"index.php?act=findpost&pid=54153\"][{POST_SNAPBACK}][/a]

BJL,
Sorry for the long delay but you understand I'm travelling.

I can't speak with authority on large format usage but I'm very interested in the principles relating to any resolution advantage. You seem to be belittling any advantage whilst I'm under the delusion that there may be a real and substantail advantage, so bear with me while I try to work through the muddle.

Checking once again the Photodo results I see that the sterling performer is the Canon 200/1.8 with a rating of 90 at f4, followed by a Contax Planar T 50/1.7 with a rating of 88 at f8, Contax G Planar 45/2 with a rating of 88 at F4 and 87 at f8, and Canon 50/1.4 with a rating of 86 at f8.

I tend to think that the diffraction limited resolution of good 35mm lenses is closer to f11 than f8, which equates to f88 in an 8x12" format regarding DoF. In other words, if it were possible to produce a 16MP 8x12" sensor, then the resolution at f88 would be identical to a 1Ds2 at f11, but slightly greater at f64 than the 1Ds2 at f8.

Given appropriate technology, the DR of the 8x12" at f88 would be vastly superior.

The real point I'm making (assuming diffraction spot size is proportional to f stop) is that the large format at f32 produces double the resolution compared to f64, whereas one of the most expensive lenses that Canon produce (discontinued?) has only a marginal increase in resolution at f4 (compared with f8) but the same DoF as 8x12' at f32.

Furthermore, halve the f stop of the large format again to f16 and you've doubled again the resolution, whilst achieving the shallow DoF of 35mm F2. I know of no 35mm lens that increases in resolution by opening up to f2.

My conclusion is, if a DoF equivalent to 35mm F2 is required, then an 8x12" sensor with sufficient pixel count will deliver more than 4x the resolution. That's 4x horizontal and 4x vertical (assuming large format lenses can be truly diffraction limited at f16). Increasing the DoF through tilt, which I agree doesn't always produce the intended result, is a bonus.

If there's a flaw in my reasoning, please set me right   .

[Jonathan Wienke]

Given appropriate technology, the DR of the 8x12" at f88 would be vastly superior.

True. But it would come at the cost of the 8x12's exposure time needing to be slightly more than 89x as long as the 35mm camera, which most people would regard as somewhat of a tradeoff. Larger formats are certainly capable of greater resolution and DR than smaller ones. But all else equal, they require much longer exposure times to capitalize on these advantages, often to the point where the advantages are no longer relevant. Which is why people don't use 8x10 view cameras to cover sporting events. There is no free lunch.

[Ray]

True. But it would come at the cost of the 8x12's exposure time needing to be slightly more than 89x as long as the 35mm camera, which most people would regard as somewhat of a tradeoff. Larger formats are certainly capable of greater resolution and DR than smaller ones. But all else equal, they require much longer exposure times to capitalize on these advantages, often to the point where the advantages are no longer relevant. Which is why people don't use 8x10 view cameras to cover sporting events. There is no free lunch.
[a href=\"index.php?act=findpost&pid=54736\"][{POST_SNAPBACK}][/a]

Hi! once again Jonathan,

Of course you're right there is a trade off but I'm not suggesting any free lunches. As always it's the right tool for the job. However, if we could conceive of an 8x10 digital camera, the difference between it and 35mm full frame is about the same order of magnitude as the difference between 35mm and the smallest digicam or mobile phone camera. The larger pixels compensate for the higher ISO's needed for faster shutter speeds to get equivalent DoFs. For example, the 5D is as noisless at ISO 800 as the avarage digicam at ISO 100. The lower noise and higher dynamic range of the larger format camera offsets the benefit of a smaller f stop number and corresponding faster shutter speed of the small digicam.

Given the appropriate noise reduction technology, if you wanted to use an 8x10 digital camera for a sporting event you would probably be using ISO 6400 at f16. Same DoF and shutter speed as f2 and ISO 100 with 35mm and theoretically just as low noise.

But to reinforce my point, 4x the resolution unless 35mm lenses can become truly diffraction limited at f2.

Perhaps the technological choice is this: Which is more feasible? (a) production of a diffraction limited lens at f2 for a 200MB 35mm sensor, or production of a 200MB 8x10" sensor? (I'm just guessing at the 200MB figure. I haven't got my calculator out.)

[Ray]

I suppose the central point I'm making here, which I've made before, is that the advantage of being able to use a faster shutter speed with the smaller format (because small f stops provide the same DoF as larger f stops with a larger format camera) is actually illusory.

It's always more sensible to use the cheaper, lighter and smaller camera if it can provide the required image quality. If the shot requires 1/50th at f2.8 and ISO 100 with a small digicam (say in poor lighting), I might initially be fooled into thinking 'wow! this little camera can do something my 35mm can't. At f2.8 DoF would be much too shallow with my heavy, bulky, expensive 35mm DSLR.'

But the fact is, I can get the same shot with my 35mm DSLR. I just have to use different settings. Instead of F2.8 and ISO 100, I use f8 and ISO 800, or instead of 1/200th, f2.8 and ISO 400, I use 1/200th, f8 and ISO 3200. Experience tells me the 35mm DSLR will still provide better image quality in the above examples. There's nothing the smaller format camera can do that the larger format can't, except be cheaper, smaller and lighter.