Problems. Part One

The scanning back has presented a set of problems that I thoroughly underestimated and to add to this my conceptual resolve is under significant strain also. I'll start with the technical issues first. All my testing has lead to one thing becoming very obvious. With the increased effective sensor size their has been a dramatic loss of depth of field. This is something I really over looked even though I have experienced it. This .pdf outlines the concepts and mathematics involved clearly. So even with fewer exposures required to cover the image circle of the lens there is an increase in the number of images required to create a fully focused image-by-image stacking. The common adage of stopping the lens down to clasp back some depth is incredibly mis leading. Whiles this methodology works to an extent it also at some point gives way to lens diffraction that’s begins to soften the entire image. And in the case of my 210mm lens even at a reasonable aperture of f32, focused at 10ft I only have 6" of depth in focus. At 20ft the depth of field rises to around 2ft. The implications for image stacking are hideous. To cover an image like the Basement is going to take at least 5 scans and that’s pushing it. I have attempted to increase the perception of depth of field via some delicate sharpening but for any perceptual gain it is soon lost in the layer stacking process that just blurs the poor transition from to out. This test below show all the new problems. Made using 4 layers per camera movement at around f16 with each scan taking about 8 minutes. The transitions between each layer that are out of focus and have no data thats in focus come out blurred just like with the dslr. Another issue that has occurred now, is that at the extremes of the camera movement there is some serious lens distortion. I was aware of this with the dslr composites but as the BetterLight back can be moved beyond the image circle the effect becomes incredibly obvious. The use of the dslr had limited the sensor ever getting to these distorted areas. This brings me to the field of view. Now that the 90mm lens is in commission I have been testing out what it can do. Having a wider field of view (100 deg) compared to the 210mm (72 deg) means it benefits from a compressing effect allowing for a potentially greater perceptual depth of field. This comes at the cost of some distortion and a positively miniature projected image circle of 216mm. The smaller imaging circle turns into fast light loss and crazy distortion. So it’s not really a solution for large composites....
But their is hope. Having tested my lenses solidly for the last few days I have been able to calculate the highest aperture that I can use before the diffraction cannot be mitigated with sharpening. The rather mundane result below has depth of field from as far as the hilltops to the boot of the silver car in the foreground. All these issues with depth of field are very landscape people type issues. Of course the methods im using to create these images are dated and by no means innovative. The real development and exciting stuff is being researched and explored in labs. Methods for gaining more depth of field are a popular area for Computation Photography. Two of the most accessible methods are Coded Aperture and Lightfield. Both require a change in the hardware of a camera to work but all use software to resolve their final output.