Leica M 9, 35mm digital versus35mm AgX
Peter Drucker, the guru of management thinkers, noted that management is not a progressive science because the management problems of today are the same as of yesteryear. The same dilemma’s and the same choices keep popping up. This stationary status is being clouded by the enormous amount of management theories and fads and even hypes. In fact the same problems are being addressed by different and often wrong methods and approaches.
Photography is in the same state: here too we find a very stationary situation: most of the major photographic themes were fixed already in the period 1880 to 1900, when landscapes, portraits, reportage, nudes and war pictures were defined as artistic categories. And the big photographic challenges are still the same: correct exposure, interesting angle of view, proper use of light, proper fixture of the moment. This handful of everlasting problems is being covered by an immense amount of cameras, cameratypes, technical expertise and equipment functionalities. Film-loading cameras and AgX technology use basic thinking to get to the bottom of the problem. Digital capture and digital image files have brought new technology and will create unpredictable consequences.
One of the most overrated topics at this moment in the digital arena is the issue of noise. It is easy to see and somewhat less easier to measure in a predictable way, but it is only a problem because digital sensors can be tweaked to very high signal-to-noise ratios. ISO values of 5000 and more may sound impressive and because one camera has it, all want it. But I do not see what of the basic photographic problems is being solved by shooting at ISO 5000. In AgX technology the useful speed was limited to ISO 400 and some used ISO 1200, but that was already an extreme situation. If you limit the ISO values on the digital cameras to 400 or 800 there is no noise to speak of and you can get fine pictures in most situations.
For the comparison between film and sensor with the same size I used the M9 at ISO160 and the M7 with Ilford Delta100 at ISO80 and Spur HRX-3 and also Spur Orthopan UR with Nanospeed UR at ISO20.
The ISO20 may be dismissed as ridiculously slow when compared to the speeds of 12000 and more that are being promoted as must-have functionality, but on a cloudy November day I could take handheld pictures with the M7 and the Summilux-M 1.4/50mm asph. wide open at shutter speeds of 1/125 to 1/500. Good enough for me and delivering very pleasant out-of focus scenes with clearly defined sharpness planes.
For the comparison test I used testcharts and a tripod and as lens I selected the Elmar-M 3.8/24mm asph. When the aperture is set to f/5.6 the resolving power of the lens exceeds the 100 lp/mm boundary with ease and with high contrast. On the MTF equipment the contrast value of 50% was found at 80 lp/mm. The performance of this lens does not limit the capabilities of the cameras. To eliminate focus errors I used the magnifier 1.4 and set the distance at 1.2 meter. At this distance you can visually detect any focus error.
The Leica M9 could resolve with good clarity about 60 lp/mm. This is close to the Nyquist limit of theoretical 73 lp/mm. These values are the same for the M8 as the pixel pitch is identical. The effect of the thicker filter on resolution is hardly detectible. I have remarked in previous reports that effective resolution is about 15% lower than the Nyquist limit. This is borne out in this case with the 60 lp/mm on the image file.
The Ilford Delta 100 is in speed comparable with the basic speed of the M9 and is therefore a good comparison. This film in Spur HRX-3 (very fine grain, good sharpness) delivered a excellent 70 lp/mm and with some reservations a useable 80 lp/mm. For most intents and purposes we can say that the M9 technology in 135 sensor size supports a performance that is now as good as that of one of the best medium speed films with the same size and identical settings (magnification and lens).
The Spur Orthopan UR is derived from the Agfa/Gevaert Copex HDP13 and identical to Adox CMS20. These references indicate outstandingly fine grain, but a restricted tonal range. But the Zone System has 6 stops useful rage and this film is above 5 stops. But is a resolution champion and here we see that digital has some distance to travel. The Orthopan resolved 125 lp/m with excellent definition and clarity and had a useful 140 lp/mm. The limit was found at 160 lp/mm but that value is not reproducible on paper. This film in the M7 then has twice the resolution of the M9. For handheld photography the differences are smaller of course, but when you want or need to exploit the limits of the optical/mechanical equipment the very low speed film is the only choice and there is no fear of noise. Even the S2 will be defeated by this film/lens combo. It is true that the deployment space is limited: tripod, careful technique, good lens, low tolerance equipment, but these are the prime characteristics of the Leica camera.
Digital technology is the first choice when speed, ease of use, and the need for hundreds pictures in one shooting session are in demand. I can see a pairing of digital and analog, especially in the case of the Leica CRF with the M9 as the main camera for large shooting volumes, ease of handling and processing and the M7/MP (or any of the other M cameras from M3 to M6) for careful work where finest details at large print sizes and the look of classical pictures is required. The sometimes harsh sharpness of the digital processing technique can be countered with the subtle and cultivated look of a fine balanced black and white baryta print.
In the B part of this article I will present the results of both technologies when negatives are printed on paper and the image files are printed with the Epson 3800.
Leica M 9: the comparison between the prints of M9 image files and AgX negatives.
First the specs: On the M9 and the M7 I attached the Elmar-M 3.8/24mm ASPH. The testtarget was photographed at the same distance and aperture (5.6). The M9 was set to ISO160. The pictures were made in ambient light with very low contrast to create the worst possible case. The results represent the minimum quality one can expect. In high contrast lighting the performance would be increased. The films were developed to normal contrast.
The M9 files were processed with Lightroom 3 beta and NIK output sharpener optimized for a viewing distance of 60cm. The files were printed with the Epson 3800 on glossy Epson best quality photo paper. The resolution was set to 240 dpi to produce the size of the testtarget that can be compared with the silver halide prints.
The negatives were printed with the Leitz V35 and the Schneider 2.8/40mm lens set at f/11. Enlargement was 16 times, the highest the V35 can handle. Paper was Ilford Multigrade IV and developer was Amaloco 6000.
All prints were scanned with the Epson V700 with 1200 dpi, the M9 prints in 48bit color and the AgX prints in 16 bit greyscale and output as TIFF files. The resulting files were 90 Mb and 60 Mb.
Given the large number of processing steps in both workflows it is inevitable that quality reductions are unavoidable. The size of the JPEG images you see is around 300K! It is a very small selection of the original negatives and image files and the section you see is enlarged to show the details. On the prints you see the 1.4 pattern with unaided visual inspection. To see the smaller patterns (only available in the Orthopan prints) you need a magnifier or a very sharp eye.
The M9 print.
The test target has a clean black and white pattern, but the M9 algorithms produced colored artifacts. To show these I had to print the file as a color print. The color shading is visible in the top left pattern that is clearly resolved. The worst artifacts are produced at the resolutions beyond the Nyquist limit. The smallest visible pattern is the 1.1, in this case corresponding to 55 lp/mm. The Nyquist limit is 73 lp/mm and after applying the practical rule that effective resolution is 15% less, we get 62 lp/mm as best (calculated guess) resolution. The actual value is quite close.
The colored artifacts are the consequence of the very thin cover glass and the lack of the anti-alias filter. The positive result is of course the very high level of definition.
The Ilford Delta100 print.
The print shows somewhat less contrast as the film needed to be printed at grade 4.5 with a minimum of exposure. One of the main challenges of printing high resolution negatives is the problem of clogging. Grain clumps tend to expand a bit at the edge when the exposure is generous. The classical rule is to use thin layer emulsions and a surface type developer and in the printing stage just enough exposure to ensure good blacks, but preferably a bit less. Given the very low lighting contrast when taking the pictures it is difficult to create deep blacks and clean whites when following the high definition rules outlined here.
The film resolves 62 lp/mm (1.25 target) and the 70 lp/mm (1.4 target) is just discernable. You might even wish to see detail in the 1.6 target (80 lp/mm) but that is really stretching the imagination. Basically you can say that the M9 sensor and post processing does now deliver the same level of quality we can expect from a top-class ISO100 film. The Delta print has somewhat less clean delineations of fine structures (the grain is beginning to become visible), but lacks the artifact distortions of the M9 images.
The Orthopan UR print.
This film/developer combo has undoubtedly the best practical performance. The limit is here simply the inability of the enlarger to offer enlargements factors of more than 16 times. On the negative the 3.2 target is just visible, corresponding to a resolution of 160 lp/mm, almost three times more than the maximum resolution of the M9. But this is cheating a bit, as we must compare the results in identical situations. The 1.8 target (90 lp/mm) is just visible in this print and the 1.4 target is clearly visible in the print when looked at with the naked eye. The star-pattern above right in the image clearly indicates the difference in performance between the three images. The Orthopan is inherently a very contrasty film with a steep characteristic curve. The resulting picture needed a gradation of 4.8 to get decent results.
Conclusion
The arguments why the photographic community has so rapidly and fully adopted and adapted to the equipment and workflow of the digital capture are very well known and in most cases true as well. The claim that digital capture vastly surpasses the performance of silver halide material has a strong psychological basis and a not so well founded factual basis.
The results presented in this article show that a digital camera with the calibre of an M9 (which belongs performance-wise in the sparsely populated top-class) can now favorably be compared with the best ISO 100 emulsions and can hold this quality into the ISO400 legue. The quality limits of the M9 sensor and in-camera-software can be clearly detected in these examples. Current microfilms (of which the Orthopan is an excellent representative) still command the premium role in high definition recording. Not everyone needs this level of detail or wants to exploit it, but there is no dispute which one does offer the best image quality and does demand the highest optical quality. The Spur company informed me that the current developer for Orthopan will be replaced by a new one that is claimed to deliver a classical characteristic curve and even better definition. We will see.
The results shown here will not be feasible when handheld shooting under dynamic conditions is required. And in those situations the M9 can play some of its stronger cards. The film-loading Leica cameras can offer emotionality and classic esthetics, not the better imagery. In more stationary or considered situations or when doing studio flash photography the slow speed films have distinctive advantages and even a fifty year old M3 can deliver state-of-the-art quality when fitted with modern Leica lenses.
This comparison does not want to reiterate the digital-analog debate (that would be flogging a dead horse), but it does want to ask for some relativism when claiming that digital is best of class. Film has clear advantages (in a limited domain to be sure) and it would be wrong to dismiss film as an obsolete medium. As soon as a future Leica M offers 35 Mp in a 35mm sized sensor, the gap may be closing, but for now the slow speed silver halide emulsions have the best performance and a distinguished look.
I am happy that I never have abandoned silver halide technology when adding solid state capture technique to my range of photographic tools.
Leica M 9, Leica testcharts additional info.
There seems to be some confusion and some questions about the origin of my test images. Let me explain. All my pictures with M8 and M9 and all other cameras I use, are always captured as RAW files, sometimes complemented by JPG files for easy viewing.
For the M9 resolution tests I set the M9 to uncompressed DNG and made the pictures. The files have a size of 36.4 Mb. These files are processed by the RAW programs of choice, in this case it was Lightroom 3.0 beta, as Adobe stated that this version has an improved raw engine. The resulting pictures were saved as TIFF files (typical size 54.3 Mb) and reduced to size for web publishing with GraphicConverter and converted to JPG to keep the size manageable. Some readers of my tests assumed that these reproduced JPG images were the original files and speculated that the color artefacts in the images have been caused by the original JPG algorithm. I can state that the color artefacts are visible in the original DNG files as processed by LR.
I used LR because it is recommended by Leica and because it is the most used RAW program in the market today. One could compare LR with Kodak D76 in AgX terms. It is very good, widely used and everyone understands its characteristics.
There are alternatives of course. especially in the dedicated RAW converter domain. At least a hundred such programs are available for free or a handful of dollars. Most of these programs use the raw engine “ dcraw” developed by David Coffin (dc). THis program is also used by Silverfast DR for example and also by Imatest. On the Mac we have RAW developer that uses dcraw but adds some sharpness algorithms. This program may be seen as a proxy for most dcraw-based programs as it is one of the best implementations.
I also tried Capture One, now in version 5 and Picture Window Pro, in version 5 too.
All these programs exhibited the color artefacts after the DNG file was interpreted. To avoid misunderstandings the precise protocol is this: DNG file is opened in the program and where necessary the sharpness is adjusted or when there are several algorithms available, every one is tried to see what happens. The best result is selected. The file is then exported to a TIFF file and the image is checked again to see if the original information has not been altered. Then the TIFF is opened with GraphicConverter, the small section of the file is selected and exported as TIFF again. Checks are made to ensure nothing has changed. These files are the ones you see here. The size is much bigger that the original JPG files, but there is now no possible contamination of the information by the conversion to JPG.
I can state that there is no difference between the DNG-TIFF-TIFF selection flow and the original DNG-TIFF-JPG selection flow.
There is the question what is the origin of the color artefacts: the in-camera processing to create the DNG files might be the cause or it is the post processing software. If it is the case that the post processing programs create the artefacts, then most programs produce the same results. It is typical for Leica DNG files however, because these programs when processing Canon or Olympus or Nikon files do not show the same pattern. The artefacts can be found in the very fine resolution area beyond the limiting resolution, the so-called spurious resolution. In this area the RGB pattern goes wild and that may be a cause for the color pattern.
I have found one program that interprets the Leica DNG files without color artefacts. It is a rather strange program called SharpRawPro and this one uses a neural network algorithm to convert the raw files. The base engine is again dcraw, but embedded in a neural network system. This is the only program I know that produces clean images, but here the reproduction of very fine detail is slightly less good than the best of the comparable programs. This result complicates the question: is this program smarter than the others or are the artefacts added to the DNG file by the program. More analysis is required, but one issue at least we can discard: it is not the JPG algorithm that is the villain, because that one is not part of the equation.
First the specs: On the M9 and the M7 I attached the Elmar-M 3.8/24mm ASPH. The testtarget was photographed at the same distance and aperture (5.6). The M9 was set to ISO160. The pictures were made in ambient light with very low contrast to create the worst possible case. The results represent the minimum quality one can expect. In high contrast lighting the performance would be increased. The films were developed to normal contrast.
The M9 files were processed with Lightroom 3 beta and NIK output sharpener optimized for a viewing distance of 60cm. The files were printed with the Epson 3800 on glossy Epson best quality photo paper. The resolution was set to 240 dpi to produce the size of the testtarget that can be compared with the silver halide prints.
The negatives were printed with the Leitz V35 and the Schneider 2.8/40mm lens set at f/11. Enlargement was 16 times, the highest the V35 can handle. Paper was Ilford Multigrade IV and developer was Amaloco 6000.
All prints were scanned with the Epson V700 with 1200 dpi, the M9 prints in 48bit color and the AgX prints in 16 bit greyscale and output as TIFF files. The resulting files were 90 Mb and 60 Mb.
Given the large number of processing steps in both workflows it is inevitable that quality reductions are unavoidable. The size of the JPEG images you see is around 300K! It is a very small selection of the original negatives and image files and the section you see is enlarged to show the details. On the prints you see the 1.4 pattern with unaided visual inspection. To see the smaller patterns (only available in the Orthopan prints) you need a magnifier or a very sharp eye.
The M9 print.
The test target has a clean black and white pattern, but the M9 algorithms produced colored artifacts. To show these I had to print the file as a color print. The color shading is visible in the top left pattern that is clearly resolved. The worst artifacts are produced at the resolutions beyond the Nyquist limit. The smallest visible pattern is the 1.1, in this case corresponding to 55 lp/mm. The Nyquist limit is 73 lp/mm and after applying the practical rule that effective resolution is 15% less, we get 62 lp/mm as best (calculated guess) resolution. The actual value is quite close.
The colored artifacts are the consequence of the very thin cover glass and the lack of the anti-alias filter. The positive result is of course the very high level of definition.
The Ilford Delta100 print.
The print shows somewhat less contrast as the film needed to be printed at grade 4.5 with a minimum of exposure. One of the main challenges of printing high resolution negatives is the problem of clogging. Grain clumps tend to expand a bit at the edge when the exposure is generous. The classical rule is to use thin layer emulsions and a surface type developer and in the printing stage just enough exposure to ensure good blacks, but preferably a bit less. Given the very low lighting contrast when taking the pictures it is difficult to create deep blacks and clean whites when following the high definition rules outlined here.
The film resolves 62 lp/mm (1.25 target) and the 70 lp/mm (1.4 target) is just discernable. You might even wish to see detail in the 1.6 target (80 lp/mm) but that is really stretching the imagination. Basically you can say that the M9 sensor and post processing does now deliver the same level of quality we can expect from a top-class ISO100 film. The Delta print has somewhat less clean delineations of fine structures (the grain is beginning to become visible), but lacks the artifact distortions of the M9 images.
The Orthopan UR print.
This film/developer combo has undoubtedly the best practical performance. The limit is here simply the inability of the enlarger to offer enlargements factors of more than 16 times. On the negative the 3.2 target is just visible, corresponding to a resolution of 160 lp/mm, almost three times more than the maximum resolution of the M9. But this is cheating a bit, as we must compare the results in identical situations. The 1.8 target (90 lp/mm) is just visible in this print and the 1.4 target is clearly visible in the print when looked at with the naked eye. The star-pattern above right in the image clearly indicates the difference in performance between the three images. The Orthopan is inherently a very contrasty film with a steep characteristic curve. The resulting picture needed a gradation of 4.8 to get decent results.
Conclusion
The arguments why the photographic community has so rapidly and fully adopted and adapted to the equipment and workflow of the digital capture are very well known and in most cases true as well. The claim that digital capture vastly surpasses the performance of silver halide material has a strong psychological basis and a not so well founded factual basis.
The results presented in this article show that a digital camera with the calibre of an M9 (which belongs performance-wise in the sparsely populated top-class) can now favorably be compared with the best ISO 100 emulsions and can hold this quality into the ISO400 legue. The quality limits of the M9 sensor and in-camera-software can be clearly detected in these examples. Current microfilms (of which the Orthopan is an excellent representative) still command the premium role in high definition recording. Not everyone needs this level of detail or wants to exploit it, but there is no dispute which one does offer the best image quality and does demand the highest optical quality. The Spur company informed me that the current developer for Orthopan will be replaced by a new one that is claimed to deliver a classical characteristic curve and even better definition. We will see.
The results shown here will not be feasible when handheld shooting under dynamic conditions is required. And in those situations the M9 can play some of its stronger cards. The film-loading Leica cameras can offer emotionality and classic esthetics, not the better imagery. In more stationary or considered situations or when doing studio flash photography the slow speed films have distinctive advantages and even a fifty year old M3 can deliver state-of-the-art quality when fitted with modern Leica lenses.
This comparison does not want to reiterate the digital-analog debate (that would be flogging a dead horse), but it does want to ask for some relativism when claiming that digital is best of class. Film has clear advantages (in a limited domain to be sure) and it would be wrong to dismiss film as an obsolete medium. As soon as a future Leica M offers 35 Mp in a 35mm sized sensor, the gap may be closing, but for now the slow speed silver halide emulsions have the best performance and a distinguished look.
I am happy that I never have abandoned silver halide technology when adding solid state capture technique to my range of photographic tools.
Leica M 9, Leica testcharts additional info.
There seems to be some confusion and some questions about the origin of my test images. Let me explain. All my pictures with M8 and M9 and all other cameras I use, are always captured as RAW files, sometimes complemented by JPG files for easy viewing.
For the M9 resolution tests I set the M9 to uncompressed DNG and made the pictures. The files have a size of 36.4 Mb. These files are processed by the RAW programs of choice, in this case it was Lightroom 3.0 beta, as Adobe stated that this version has an improved raw engine. The resulting pictures were saved as TIFF files (typical size 54.3 Mb) and reduced to size for web publishing with GraphicConverter and converted to JPG to keep the size manageable. Some readers of my tests assumed that these reproduced JPG images were the original files and speculated that the color artefacts in the images have been caused by the original JPG algorithm. I can state that the color artefacts are visible in the original DNG files as processed by LR.
I used LR because it is recommended by Leica and because it is the most used RAW program in the market today. One could compare LR with Kodak D76 in AgX terms. It is very good, widely used and everyone understands its characteristics.
There are alternatives of course. especially in the dedicated RAW converter domain. At least a hundred such programs are available for free or a handful of dollars. Most of these programs use the raw engine “ dcraw” developed by David Coffin (dc). THis program is also used by Silverfast DR for example and also by Imatest. On the Mac we have RAW developer that uses dcraw but adds some sharpness algorithms. This program may be seen as a proxy for most dcraw-based programs as it is one of the best implementations.
I also tried Capture One, now in version 5 and Picture Window Pro, in version 5 too.
All these programs exhibited the color artefacts after the DNG file was interpreted. To avoid misunderstandings the precise protocol is this: DNG file is opened in the program and where necessary the sharpness is adjusted or when there are several algorithms available, every one is tried to see what happens. The best result is selected. The file is then exported to a TIFF file and the image is checked again to see if the original information has not been altered. Then the TIFF is opened with GraphicConverter, the small section of the file is selected and exported as TIFF again. Checks are made to ensure nothing has changed. These files are the ones you see here. The size is much bigger that the original JPG files, but there is now no possible contamination of the information by the conversion to JPG.
I can state that there is no difference between the DNG-TIFF-TIFF selection flow and the original DNG-TIFF-JPG selection flow.
There is the question what is the origin of the color artefacts: the in-camera processing to create the DNG files might be the cause or it is the post processing software. If it is the case that the post processing programs create the artefacts, then most programs produce the same results. It is typical for Leica DNG files however, because these programs when processing Canon or Olympus or Nikon files do not show the same pattern. The artefacts can be found in the very fine resolution area beyond the limiting resolution, the so-called spurious resolution. In this area the RGB pattern goes wild and that may be a cause for the color pattern.
I have found one program that interprets the Leica DNG files without color artefacts. It is a rather strange program called SharpRawPro and this one uses a neural network algorithm to convert the raw files. The base engine is again dcraw, but embedded in a neural network system. This is the only program I know that produces clean images, but here the reproduction of very fine detail is slightly less good than the best of the comparable programs. This result complicates the question: is this program smarter than the others or are the artefacts added to the DNG file by the program. More analysis is required, but one issue at least we can discard: it is not the JPG algorithm that is the villain, because that one is not part of the equation.