Alan7140

Taking this right back to the beginning perhaps explains the problem best.
 
I got my X-Pro1 in late June 2012, probably from the second or third batch into Australia. Adobe came out with their ACR support for RAF soon afterwards, but my first few weeks were spent shooting OOC jpeg while familiarising myself with the camera. Once I started to process raw files with ACR I noticed that things weren't the best, but compared to SilkyPix (I use Windows, they were the only two RAF processors available at the time) there was not much difference, although SP seemed a bit softer and ACR more contrasty.
 
Then Corel got on the RAF train (September 2012) with their process in PaintShop Pro X5, and while worse overall, it probably best reveals why both early SilkyPix and ACR acted like they did. Below are two 100% sections  of a photo I took at the time deliberately to highlight with vegetation and high contrast edge rendition I'd been noticing with ACR (and to a lesser degree with SilkyPix), and which many later called the "zipper effect".
 
The first is ACR, the second is Corel PSP X5. (Again, these are the first 2012 versions of RAF processing, not the current versions):
 

 

 
The Corel version perhaps shows the problem that Adobe etc with Fuji files, and the original dcraw showed a similar result to the PSP X5 version - spurious coloured pixels appearing along high contrast edges (note the lines between grille and white paint on the car and around the number plate, and the choppy edges of the geometric grille pattern against white plain detail; in fact anywhere there is a high contrast edge. If you compare the top edge of the door mirror you can just see the slight waviness in the edge of the Adobe version corresponding to where the rows of spurious pixels occur in the X5 version, again probably confirming default filter application designed to smooth edges.
 
My guess was (and still is) that both SilkyPix and Adobe tackled this with a default addition of noise reduction (more heavily so in Adobe's case), and compensated for the NR's softening of detail and loss of colour by adding a hefty dose of saturation and sharpening. The spreading of colour in the two sticks embedded in the ground further attests to the likelihood of NR having been applied - the green/yellow colour of the grass spreading has nearly killed the brown in the highlights.
 
As mentioned, I tried dcraw in command line form when Mac forums started mentioning it as being the basis for RPP & Iridient, and it gave a similar result to Corel's demosaic, but Dave Coffin then responded quickly to criticism and changed his demosaic algorithm to give the clean, sharp rendition that is with us today as the basis of programs such as Photo Ninja, Iridient, RPP, LightZone, Photivo, Helicon Filter, etc etc. Corel took a lot longer to respond, and still hasn't quite got there with Aftershot Pro.
 
Adobe (and probably SilkyPix) on the other hand, seem to have stuck with their original algorithm and simply refined the degree of default NR, saturation and sharpening applied to reduce the outline and spreading effect of the original. This would probably also explain why Adobe ACR/LR/DNG files in particular can react so nervously to post sharpening, sharpening apparently already having been applied by default during the demosaic.
 
Until someone from any of the companies mentioned actually comes out and publicly states that this is not the case (and they haven't in five years), then I'm inclined to stick with the above explanation.
 
As such ACR/LR/DNG is basically flawed in my book, and while many might think their results are good enough, my personal quality standards (and those of many others) won't permit me to use ACR for any commercial job. One never knows how the final image may be used, at what size and with what enhancements, and as such it is the professional's duty to provide the best possible image to the client, not one that may - or may not - be "good enough". Iridient's X-Transformer is definitely a viable demosaic alternative to those who feel welded to Adobe and the ease of ACR/LR, just be sure to properly tune the base settings (to their credit Iridient do listen and have already changed the defaults at least once).

The 2,8/50 Touit Makro is worth every cent over cheaper alternatives.
 
True 1:1 magnification with no need for extension tubes or close-up dioptres, and Zeiss image quality that is to die for. The AF does hunt a bit at times, but who uses AF with macro photography, anyway?
 
It remains one of my favourite lenses for my X-cameras.

Taking this right back to the beginning perhaps explains the problem best.
 
I got my X-Pro1 in late June 2012, probably from the second or third batch into Australia. Adobe came out with their ACR support for RAF soon afterwards, but my first few weeks were spent shooting OOC jpeg while familiarising myself with the camera. Once I started to process raw files with ACR I noticed that things weren't the best, but compared to SilkyPix (I use Windows, they were the only two RAF processors available at the time) there was not much difference, although SP seemed a bit softer and ACR more contrasty.
 
Then Corel got on the RAF train (September 2012) with their process in PaintShop Pro X5, and while worse overall, it probably best reveals why both early SilkyPix and ACR acted like they did. Below are two 100% sections  of a photo I took at the time deliberately to highlight with vegetation and high contrast edge rendition I'd been noticing with ACR (and to a lesser degree with SilkyPix), and which many later called the "zipper effect".
 
The first is ACR, the second is Corel PSP X5. (Again, these are the first 2012 versions of RAF processing, not the current versions):
 

 

 
The Corel version perhaps shows the problem that Adobe etc with Fuji files, and the original dcraw showed a similar result to the PSP X5 version - spurious coloured pixels appearing along high contrast edges (note the lines between grille and white paint on the car and around the number plate, and the choppy edges of the geometric grille pattern against white plain detail; in fact anywhere there is a high contrast edge. If you compare the top edge of the door mirror you can just see the slight waviness in the edge of the Adobe version corresponding to where the rows of spurious pixels occur in the X5 version, again probably confirming default filter application designed to smooth edges.
 
My guess was (and still is) that both SilkyPix and Adobe tackled this with a default addition of noise reduction (more heavily so in Adobe's case), and compensated for the NR's softening of detail and loss of colour by adding a hefty dose of saturation and sharpening. The spreading of colour in the two sticks embedded in the ground further attests to the likelihood of NR having been applied - the green/yellow colour of the grass spreading has nearly killed the brown in the highlights.
 
As mentioned, I tried dcraw in command line form when Mac forums started mentioning it as being the basis for RPP & Iridient, and it gave a similar result to Corel's demosaic, but Dave Coffin then responded quickly to criticism and changed his demosaic algorithm to give the clean, sharp rendition that is with us today as the basis of programs such as Photo Ninja, Iridient, RPP, LightZone, Photivo, Helicon Filter, etc etc. Corel took a lot longer to respond, and still hasn't quite got there with Aftershot Pro.
 
Adobe (and probably SilkyPix) on the other hand, seem to have stuck with their original algorithm and simply refined the degree of default NR, saturation and sharpening applied to reduce the outline and spreading effect of the original. This would probably also explain why Adobe ACR/LR/DNG files in particular can react so nervously to post sharpening, sharpening apparently already having been applied by default during the demosaic.
 
Until someone from any of the companies mentioned actually comes out and publicly states that this is not the case (and they haven't in five years), then I'm inclined to stick with the above explanation.
 
As such ACR/LR/DNG is basically flawed in my book, and while many might think their results are good enough, my personal quality standards (and those of many others) won't permit me to use ACR for any commercial job. One never knows how the final image may be used, at what size and with what enhancements, and as such it is the professional's duty to provide the best possible image to the client, not one that may - or may not - be "good enough". Iridient's X-Transformer is definitely a viable demosaic alternative to those who feel welded to Adobe and the ease of ACR/LR, just be sure to properly tune the base settings (to their credit Iridient do listen and have already changed the defaults at least once).

Taking this right back to the beginning perhaps explains the problem best.
 
I got my X-Pro1 in late June 2012, probably from the second or third batch into Australia. Adobe came out with their ACR support for RAF soon afterwards, but my first few weeks were spent shooting OOC jpeg while familiarising myself with the camera. Once I started to process raw files with ACR I noticed that things weren't the best, but compared to SilkyPix (I use Windows, they were the only two RAF processors available at the time) there was not much difference, although SP seemed a bit softer and ACR more contrasty.
 
Then Corel got on the RAF train (September 2012) with their process in PaintShop Pro X5, and while worse overall, it probably best reveals why both early SilkyPix and ACR acted like they did. Below are two 100% sections  of a photo I took at the time deliberately to highlight with vegetation and high contrast edge rendition I'd been noticing with ACR (and to a lesser degree with SilkyPix), and which many later called the "zipper effect".
 
The first is ACR, the second is Corel PSP X5. (Again, these are the first 2012 versions of RAF processing, not the current versions):
 

 

 
The Corel version perhaps shows the problem that Adobe etc with Fuji files, and the original dcraw showed a similar result to the PSP X5 version - spurious coloured pixels appearing along high contrast edges (note the lines between grille and white paint on the car and around the number plate, and the choppy edges of the geometric grille pattern against white plain detail; in fact anywhere there is a high contrast edge. If you compare the top edge of the door mirror you can just see the slight waviness in the edge of the Adobe version corresponding to where the rows of spurious pixels occur in the X5 version, again probably confirming default filter application designed to smooth edges.
 
My guess was (and still is) that both SilkyPix and Adobe tackled this with a default addition of noise reduction (more heavily so in Adobe's case), and compensated for the NR's softening of detail and loss of colour by adding a hefty dose of saturation and sharpening. The spreading of colour in the two sticks embedded in the ground further attests to the likelihood of NR having been applied - the green/yellow colour of the grass spreading has nearly killed the brown in the highlights.
 
As mentioned, I tried dcraw in command line form when Mac forums started mentioning it as being the basis for RPP & Iridient, and it gave a similar result to Corel's demosaic, but Dave Coffin then responded quickly to criticism and changed his demosaic algorithm to give the clean, sharp rendition that is with us today as the basis of programs such as Photo Ninja, Iridient, RPP, LightZone, Photivo, Helicon Filter, etc etc. Corel took a lot longer to respond, and still hasn't quite got there with Aftershot Pro.
 
Adobe (and probably SilkyPix) on the other hand, seem to have stuck with their original algorithm and simply refined the degree of default NR, saturation and sharpening applied to reduce the outline and spreading effect of the original. This would probably also explain why Adobe ACR/LR/DNG files in particular can react so nervously to post sharpening, sharpening apparently already having been applied by default during the demosaic.
 
Until someone from any of the companies mentioned actually comes out and publicly states that this is not the case (and they haven't in five years), then I'm inclined to stick with the above explanation.
 
As such ACR/LR/DNG is basically flawed in my book, and while many might think their results are good enough, my personal quality standards (and those of many others) won't permit me to use ACR for any commercial job. One never knows how the final image may be used, at what size and with what enhancements, and as such it is the professional's duty to provide the best possible image to the client, not one that may - or may not - be "good enough". Iridient's X-Transformer is definitely a viable demosaic alternative to those who feel welded to Adobe and the ease of ACR/LR, just be sure to properly tune the base settings (to their credit Iridient do listen and have already changed the defaults at least once).

Taking this right back to the beginning perhaps explains the problem best.
 
I got my X-Pro1 in late June 2012, probably from the second or third batch into Australia. Adobe came out with their ACR support for RAF soon afterwards, but my first few weeks were spent shooting OOC jpeg while familiarising myself with the camera. Once I started to process raw files with ACR I noticed that things weren't the best, but compared to SilkyPix (I use Windows, they were the only two RAF processors available at the time) there was not much difference, although SP seemed a bit softer and ACR more contrasty.
 
Then Corel got on the RAF train (September 2012) with their process in PaintShop Pro X5, and while worse overall, it probably best reveals why both early SilkyPix and ACR acted like they did. Below are two 100% sections  of a photo I took at the time deliberately to highlight with vegetation and high contrast edge rendition I'd been noticing with ACR (and to a lesser degree with SilkyPix), and which many later called the "zipper effect".
 
The first is ACR, the second is Corel PSP X5. (Again, these are the first 2012 versions of RAF processing, not the current versions):
 

 

 
The Corel version perhaps shows the problem that Adobe etc with Fuji files, and the original dcraw showed a similar result to the PSP X5 version - spurious coloured pixels appearing along high contrast edges (note the lines between grille and white paint on the car and around the number plate, and the choppy edges of the geometric grille pattern against white plain detail; in fact anywhere there is a high contrast edge. If you compare the top edge of the door mirror you can just see the slight waviness in the edge of the Adobe version corresponding to where the rows of spurious pixels occur in the X5 version, again probably confirming default filter application designed to smooth edges.
 
My guess was (and still is) that both SilkyPix and Adobe tackled this with a default addition of noise reduction (more heavily so in Adobe's case), and compensated for the NR's softening of detail and loss of colour by adding a hefty dose of saturation and sharpening. The spreading of colour in the two sticks embedded in the ground further attests to the likelihood of NR having been applied - the green/yellow colour of the grass spreading has nearly killed the brown in the highlights.
 
As mentioned, I tried dcraw in command line form when Mac forums started mentioning it as being the basis for RPP & Iridient, and it gave a similar result to Corel's demosaic, but Dave Coffin then responded quickly to criticism and changed his demosaic algorithm to give the clean, sharp rendition that is with us today as the basis of programs such as Photo Ninja, Iridient, RPP, LightZone, Photivo, Helicon Filter, etc etc. Corel took a lot longer to respond, and still hasn't quite got there with Aftershot Pro.
 
Adobe (and probably SilkyPix) on the other hand, seem to have stuck with their original algorithm and simply refined the degree of default NR, saturation and sharpening applied to reduce the outline and spreading effect of the original. This would probably also explain why Adobe ACR/LR/DNG files in particular can react so nervously to post sharpening, sharpening apparently already having been applied by default during the demosaic.
 
Until someone from any of the companies mentioned actually comes out and publicly states that this is not the case (and they haven't in five years), then I'm inclined to stick with the above explanation.
 
As such ACR/LR/DNG is basically flawed in my book, and while many might think their results are good enough, my personal quality standards (and those of many others) won't permit me to use ACR for any commercial job. One never knows how the final image may be used, at what size and with what enhancements, and as such it is the professional's duty to provide the best possible image to the client, not one that may - or may not - be "good enough". Iridient's X-Transformer is definitely a viable demosaic alternative to those who feel welded to Adobe and the ease of ACR/LR, just be sure to properly tune the base settings (to their credit Iridient do listen and have already changed the defaults at least once).

Taking this right back to the beginning perhaps explains the problem best.
 
I got my X-Pro1 in late June 2012, probably from the second or third batch into Australia. Adobe came out with their ACR support for RAF soon afterwards, but my first few weeks were spent shooting OOC jpeg while familiarising myself with the camera. Once I started to process raw files with ACR I noticed that things weren't the best, but compared to SilkyPix (I use Windows, they were the only two RAF processors available at the time) there was not much difference, although SP seemed a bit softer and ACR more contrasty.
 
Then Corel got on the RAF train (September 2012) with their process in PaintShop Pro X5, and while worse overall, it probably best reveals why both early SilkyPix and ACR acted like they did. Below are two 100% sections  of a photo I took at the time deliberately to highlight with vegetation and high contrast edge rendition I'd been noticing with ACR (and to a lesser degree with SilkyPix), and which many later called the "zipper effect".
 
The first is ACR, the second is Corel PSP X5. (Again, these are the first 2012 versions of RAF processing, not the current versions):
 

 

 
The Corel version perhaps shows the problem that Adobe etc with Fuji files, and the original dcraw showed a similar result to the PSP X5 version - spurious coloured pixels appearing along high contrast edges (note the lines between grille and white paint on the car and around the number plate, and the choppy edges of the geometric grille pattern against white plain detail; in fact anywhere there is a high contrast edge. If you compare the top edge of the door mirror you can just see the slight waviness in the edge of the Adobe version corresponding to where the rows of spurious pixels occur in the X5 version, again probably confirming default filter application designed to smooth edges.
 
My guess was (and still is) that both SilkyPix and Adobe tackled this with a default addition of noise reduction (more heavily so in Adobe's case), and compensated for the NR's softening of detail and loss of colour by adding a hefty dose of saturation and sharpening. The spreading of colour in the two sticks embedded in the ground further attests to the likelihood of NR having been applied - the green/yellow colour of the grass spreading has nearly killed the brown in the highlights.
 
As mentioned, I tried dcraw in command line form when Mac forums started mentioning it as being the basis for RPP & Iridient, and it gave a similar result to Corel's demosaic, but Dave Coffin then responded quickly to criticism and changed his demosaic algorithm to give the clean, sharp rendition that is with us today as the basis of programs such as Photo Ninja, Iridient, RPP, LightZone, Photivo, Helicon Filter, etc etc. Corel took a lot longer to respond, and still hasn't quite got there with Aftershot Pro.
 
Adobe (and probably SilkyPix) on the other hand, seem to have stuck with their original algorithm and simply refined the degree of default NR, saturation and sharpening applied to reduce the outline and spreading effect of the original. This would probably also explain why Adobe ACR/LR/DNG files in particular can react so nervously to post sharpening, sharpening apparently already having been applied by default during the demosaic.
 
Until someone from any of the companies mentioned actually comes out and publicly states that this is not the case (and they haven't in five years), then I'm inclined to stick with the above explanation.
 
As such ACR/LR/DNG is basically flawed in my book, and while many might think their results are good enough, my personal quality standards (and those of many others) won't permit me to use ACR for any commercial job. One never knows how the final image may be used, at what size and with what enhancements, and as such it is the professional's duty to provide the best possible image to the client, not one that may - or may not - be "good enough". Iridient's X-Transformer is definitely a viable demosaic alternative to those who feel welded to Adobe and the ease of ACR/LR, just be sure to properly tune the base settings (to their credit Iridient do listen and have already changed the defaults at least once).

Yes, you can.
 
Yes, you can.
Sigma Foveon DP Merrill.
No AA filter, zero moire.

Theoretical vs practical going on here - I guess they're basing their assumption on the fact that there are a known number of pixels at a known size on a sensor, whereas film has a potentially vastly greater numbers of much smaller unexposed silver halide grains randomly spread in an emulsion, therefore allowing them to completely and randomly cover every part of the sensitive material to a considerable depth, and not just the single layer of pixels (or three layers in the case of Foveon) as with digital.
 
Of course this changes during the exposure and processing procedure - digital loses nothing in number, order or arrangement of its pixels, whereas film increasingly exposes fewer and fewer halides in the shadow areas, and clumps more and more together in the highlight areas. The resulting film "grain" chops the image up, particularly in those areas, and if you are photographing highly complex fine and contrasty detail it will probably run into trouble well before digital does. The exposed halides are reduced to metallic silver grains which are "developed" to a size that gives the image visible density and can thus be printed. They are no longer so tiny as their silver halide origins, the unexposed and therefore undeveloped portion of which are dissolved away in the fixing process.
 
That said, well exposed, properly processed fine-grain film of subject matter that does not have too higher dynamic range and is largely comprised around tones of 18% grey will more than likely be able to stand much higher enlargement, if only for the fact that its grain is random and therefore does not form regular patterns such as in digital where "jaggies" effectively put a ceiling on enlargement capability as soon as they become apparent at normal viewing distance. The eye more readily accommodates the random grain of film, both because regular patterns like a sensor array are rare in the natural world, and there also is a much longer history of photographs produced with film than with sensors, so familiarity plays a role (although that is diminishing with time). Digital noise has, for most of its earlier development been jarring to look at.
 
Like comparing lenses of different focal lengths between different digital camera formats, comparing film with digital is every bit as pointless. Each does its own unique thing in use, and it's up to the user to choose that which best suits the purpose at hand.
 
If, for instance, you were to take a well exposed and processes 8x10" collodion plate photograph and compare it with a digital file from any sized commercial sensor and compare a 30x40 print from each, the digital file will fail miserably. In practical terms, collodion plates became redundant in the late 19th century for some exceedingly good reasons. Carrying a full darkroom around with you and having to coat the glass plate with fresh emulsion and expose and process it it while it was still wet did limit its practicality somewhat.