The GetDPI Photography Forum
Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!
Why did you choose your back?
- Thread starter RayM
- Start date
Guy Mancuso
Administrator, Instructor
LOL Not yet , maybe next year. Mamiya AFDIII and the new Phase One are one in the same . If you buy a new Phase back you et a new Phase One body and 80D lens. Right now i am shooting with a loaner AFD which i hate. So just waiting on my final piece to be delivered. Should be any day now. Along with about 6 other members that just bought new Phase backs.
Guy Mancuso
Administrator, Instructor
Maybe the 60mpx back announced at Photokinia. LOL I do wonder what is coming next though.
Guy,
my eMotion75 @ 33 MP is already more than I need for 99% of my work.
A big file is handy,because I can crop.
Also the file can be easily doubled in size with no loss,with the right software.
Cheers,
Willem.
N.B I doubt if there ever will be a 60MB back and if so at what cost?
my eMotion75 @ 33 MP is already more than I need for 99% of my work.
A big file is handy,because I can crop.
Also the file can be easily doubled in size with no loss,with the right software.
Cheers,
Willem.
N.B I doubt if there ever will be a 60MB back and if so at what cost?
Guy Mancuso
Administrator, Instructor
I agree 60 mpx is going over the top for me. No way i would buy one . I think the max at 39mpx today is way more than enough. Rather see other improvements to the backs than more Pixels.
David Klepacki
New member
Steve, I want to publicly admit that I have made a TERRIBLE choice of words in my earlier note, and that never once have I heard you (or Dave McRitchie) "complain", about anything. Rather, the more accurate wording should have indicated that both you and Dave "share my concerns" in this backorder situation. Again, I want to thank you for the professional support and service you have provided to me and others in the MF community over the years.
David
Agree completely and might make an interesting topic for another post, i.e. what improvements would you like to see in your back. I'm guessing that a bigger, brighter viewfinder ala Nikon D3 would be up near the top of many lists. Even out on the beach I can see that screen, hit a button to zoom and know if I've nailed it within two seconds or so.
60 MPix in a 4.5 X 6 format would result in about the same lp/mm resolution capability of the current M8 sensor. This would create a lens limited platform. Unfortunately, the physics of it all would reduce the signal to noise ratio as well just due to photon shot noise, so some of the MF-ness would go away.
-bob
-bob
David Klepacki
New member
Bob, this is not technically true. The Sinar 54H back is capable of producing 89 MPix in 645 format. I use this back with both Hasselblad V and Contax 645 cameras, and I assure you that it is not lens limited. In fact, this back produces superior image quality with perhaps one of the highest signal to noise ratios on the market today, since it achieves this resolution with large 9 micron pixels and multiple shots. There is full RGB color information at every RGB pixel location, so there is no need for software interpolation to estimate the missing 33% of the image data from single shot backs. The result is incredible image quality with no artifacts.
Of course, this multi-shot back is really only for studio use. I would say the difference in image quality between the mult-shot 89 MPix images and standard single-shot 39MPix images is at least the same jump in quality as seen between 35mm DSLR cameras and medium format, but maybe even somewhat better because of the total absence of moire and false colors with the 89MPix images.
David
David Klepacki
New member
Guy, the next step up in image quality is with the multi-shot backs. There is even a significant leap in quality going from a single-shot 39MPix back to a four-shot 39MPix back. Again, the reason is that the single-shot backs only capture one color per pixel and so must estimate the remaining 33% of the image data that it does not capture. For the most part, the software estimation of this missing information does a decent job, but it is no substitute for actually capturing all of the genuine RGB information with the back itself.
Guy Mancuso
Administrator, Instructor
I know and that is something about the multi shot backs . Great for the studio no question
Graham Mitchell
New member
The per-pixel noise might drop to DSLR levels but remember that the enlargement is much lower, so the noise is less apparent on these large files.
I think people forget that when they look at the noise of a 33/39MP back at 100% and compare with a DSLR. MF can get away with a lot more noise. If they look comparable on a per-pixel basis, the MF image will appear to have much less noise at final output size.
David Klepacki
New member
Just correcting my own mistake....software is needed to estimate 66% of the missing image information for all single-shot backs...amazing when you think about it.
Steve Hendrix
Well-known member
David - that is ok. I appreciate you posting this publicly, but it's really difficult to write on forums, I think. It's so easy to have a feeling you're trying to convey, and just be typing away, and it comes out all wrong. I've certainly done it enough times.
Thanks,
Steve Hendrix
www.ppratlanta.com/digital.php
David,
If I am wrong, it won't be th first time.
for reference please see http://www.luminous-landscape.com/tutorials/resolution.shtml and http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html#pixelsize
My comment was relative to a hypothetical bayer sensor 60 MP 4.5X6 back and the loss of signal to noise ratio (loss of MF-ness) as well as the MTF of available MF lenses (lens limited). This limitation is somewhat subjective since it depends on the contrast ratio that is considered acceptable in the resolution of a line pair. Many folks use a 25%:1 ratio as do I, some use a 9%:1 as that is considered to be the limit of discernment of the human eye.
With any digital array-sensor system, the signal to noise ratio of each shot is dictated by pixel size. Signal to noise ratio is related to the square root of the number of shots if the individual shots are aggregated. Optical aggregation would be the result of simply using the larger file and printing it at the same size as a single shot image would produce. One of the major reasons for multi-shot I think is the elimination of sampling artifacts such as Moire which is caused by the near identical periodicity of the on-sensor resolved image with the pixel pitch. The sub-pixel shift and shoot creates a partial-pixel shifted overlapped set of images, but not really more resolution than the lens is capable of providing.
The number of shots taken do not change the diffraction limitation story which as solely based on circle of confusion, and for digital, pixel size. i have assumed in my thinking an Airy disk diameter to pixel size ratio of 1:2. These assumptions also include perfect focus, which is not a three dimensional thing, it exists in only one plane. If focus is not perfect, or if the lens is stopped down below its diffraction limit, or if real-world issues such as camera vibration or even power supply noise are considered the sensor somewhat less lens limited since these factors start reducing the delivered resolution from the Nyquist ideal. We are getting very close to lens limited and several manufacturers have been forced to bring out better lenses just because of these digital factors.
Please if you will, point out the errors in my reasoning.
thanks
-bob
Granted to some degree, but not if the reason one buys the larger back is to print larger and maintain a better than 6lp.mm final resolution.
I am expecting to make prints a whole lot bigger with MF.
-bob
T
thsinar
Guest
True what David says: amazing quality out of a 22 MPx sensor. Do use the best possible lenses like HR lenses and you shall be blown away when doing a 16-shot.
BTW: it is not 33% of the colour/luminance information which is missing and has to be interpolated, but 2 out of 3 basic colour out of RGB are not registered for each pixels, thus 66% has to be interpolated.
Thierry
BTW: it is not 33% of the colour/luminance information which is missing and has to be interpolated, but 2 out of 3 basic colour out of RGB are not registered for each pixels, thus 66% has to be interpolated.
Thierry
David Klepacki
New member
Thierry, you must have missed my post above. I already corrected myself.
Also, I prefer the Hasselblad lenses over the HR ones since the close focusing abilities are superior (e.g., the floating elements of the 40 CFE IF). The HR lenses are corrected only for infinity, have too large a close focusing distance for me, and their image circles are not large enough to accommodate any significant movements. In fact, if anyone eventually makes a full frame square sensor, the HR lenses will not be able to cover the 55x55 image circle.
David
T
thsinar
Guest
Yes, I have read after posting, sorry about my "rush".
Yes, you are right concerning the HR lenses and close focusing ability. But on the other hand they have been built and optimized to be used on a view camera and to resolve the current sensors' resolution, and in fact their image circle is wide enough in most cases and needs. You will be surprised to see how big this image circle is, in fact, from 70mm to 80mm at infinity and stopped down 2 f-stops (5th column from left).
The resulting shift ways are quite important (6th column from left): please have a look at the "Lens Data" table I have attached here. These shift possibilities given in this table (5th right column) are VERY conservatives values given by us, and are typically much more without any loss at full open aperture. As an example, the 28mm HR can easily be shifted 10mm FULL OPEN, and still sharp in the corners, and this at infinity.
The resolution of these lenses (3rd column from right) are given as a MINIMUM and at the edge of the image circle: in the center it is much more.
Best regards,
Thierry
Yes, you are right concerning the HR lenses and close focusing ability. But on the other hand they have been built and optimized to be used on a view camera and to resolve the current sensors' resolution, and in fact their image circle is wide enough in most cases and needs. You will be surprised to see how big this image circle is, in fact, from 70mm to 80mm at infinity and stopped down 2 f-stops (5th column from left).
The resulting shift ways are quite important (6th column from left): please have a look at the "Lens Data" table I have attached here. These shift possibilities given in this table (5th right column) are VERY conservatives values given by us, and are typically much more without any loss at full open aperture. As an example, the 28mm HR can easily be shifted 10mm FULL OPEN, and still sharp in the corners, and this at infinity.
The resolution of these lenses (3rd column from right) are given as a MINIMUM and at the edge of the image circle: in the center it is much more.
Best regards,
Thierry
David Klepacki
New member
Bob, the sensor is "Bayer-limited", and is the primary reason why most "decent" lenses do not yet limit the image capture (although it is possible for some lenses that are of truly lesser quality).
The fact is that you are capturing only 25% of the red and blue channel content of any image, and only 50% of the green channel content. Much of the online analyses are flawed since they assume a uniform monochromatic sensor as their first assumption, and everything goes wrong from there. A single image capture using a Bayer sensor will require you to "guess" (albeit with sophisticated mathematics) the missing 75% of the red information, the missing 75% of the blue information, and the missing 50% of the green information.....that's quite a lot of missing information.
Maybe you can see this with an extreme example. If you were to photograph a hypothetical object with extremely fine detail (say content with 100 lp/mm), but that only consisted of a very large range of red intensities and no other wavelengths, you would find the effective resolution of the sensor to be much lower than the capability of the lens. In this case, it is because the sensor resolution has been effectively reduced to the number of pixels that can actually capture red, since the other pixels would be rendered useless in providing any of the missing red pixels on those other sites.
Even if your object is mostly green (the most advantageous in terms of digital capture resolution), it is still only half of the green content that is contained in the image. And, the sampling frequency for these wavelengths are still not uniform at pixel width spacing, but rather consist of interspersed arrays of data resulting in a pixel spacing of two pixel-widths, which lowers the Nyquist frequency for both the horizontal and vertical dimensions in this wavelength. It is a common mistake to interpolate these samples and then claim a higher Nyquist frequency for the sensor.
So, bottom line is that most "decent" lenses (eg, let's say ones that can resolve at least 60 lp/mm or more at 50% MTF) are still outresolving, not limiting, the MF sensors.
Last edited: