Okay, so a quick reminder of what we're looking at. These shots were all taken at low ISO's (noise is not an issue) with lots of light through an F/16 stop down ring on an APS-C crop sensored DSLR. Below I have presented a couple of the photos I took along with 100% crops taken from key areas of each image. They should give me a good impression of the overall performance of the lens.
Firstly though I would like to make a general analysis of the results. From an overall image quality perspective, the photos are 'usable'. They have acceptable image quality to a vast majority of camera users. For example this level of quality would be good for holiday users, and for general day to day snaps. The sharpness is slightly lacking and benefits strongly from a boosted in post processing. However, detail seems to be adequate with readable signs and smaller details across most of the frame when the image is sharpened. Colour is okay and contrast is a little on the low side. Forming an early conclusion I would say that overall the images produced by this lens in this configuration are good. From a peronal perspective by comparing the results to my expectations, given the incredible simplicity of the design I would say the results are excellent! They show huge promise and great potential for future prototyping.
That said, lets take a look in more detail at some of the sample images:
Over all the image is nice and bright. Contrast and detail look good initially around centre frame, with strong blurring towards the corners. At F/16, I don't think this is depth of field, I think it is spherical abberation creeping into the frame. Another initial thought is that the lens might have been off axis slightly, angled towards the right of the frame. The buildings on the right seem sharper than on the left. This is a mount rather than a lens problem, showing that the push pull focusing has caused the optimal focus plane to become unparalleled to the image plane.
Dead centre frame, the IQ is good. details down to the gold text is retained, even with it's lack of contrast against the bright stone. Finer details such as colour breaks between the bricks are washed out, but still visible if you really look hard. Further more you can see the sharpening pattern applied by Adobe Lightroom struggling slightly to pull back detail, causing a slight artificial look. This tells me that while the result is okay, the resolution of this lens isn't as high as it really needs to be for digital sensors. This needs to be improved.
Max top frame shows spherical abberations creeping into shot. Haloing, coma, and SA make the image look like it's being stretched. This is the result of the off axis rays not converging on the image plane. A good deal of detail is retained but when looking at it 100% it's obviously not good. Colour rendition on the other hand is okay, with the green tinted copper showing up well next to the blue sky.
Mid far right frame. Detail here, and the lack of strong coma is surprising, and leads me to think the lens was unparalleled to the sensor. Detail is good, contrast is a little low, with some of the similar tones washing into each other.
I took this photos specifically to demonstrate detail rendition. At first glance, again, I think the lens was not parralel, as the SA in the bottom right hand corner is much stronger than the left. A serious rethink of the mount might be in order.
Dead centre frame detail is good, not great. Lightroom is visibly finding it difficult to pull back detail. Texture in the mortar is not fine but is visible, which is better than nothing.
CA or comma from off axis rays (not sure which) contributes to this obvious edge degradation. This is happening all around the outside of the image but is very noticeable around strong contrasting blocks of tone, as shown here between the shadowed stone and the bright sky. It would have been interesting to have this composition in centre frame to determine if it was CA or coma causing the problem, my eyes aren't good enough yet to tell the difference.
Far left corner. Visible stretching caused by coma, as to be expected. Slight bit of CA I think is causing the bluish tinge around details on the stone.
I love this photos because this shot is exactly as the scene looked when I took the photo. The building was rendered very well. This is probably a result of low detail however, with big flat planes of tone being very generous to the lens. Detail on the signs at the bottom is readable, which is good. Again, Lightroom is struggling, but the contrast is there and means the words can be clearly understood, even though it is far from centre frame (the sharpest area with least errors). CA is obvious along the top and bottom of the signs as they do not infact, glow blue.
What I am doing here is called pixel peeping, and there is mixed opinions in the photographic industry about whether people should do it or not. Many say the quality of a photo is in the composition and other qualities, and looking this close is effectively just hunting for errors to be picky. As a designer, it's my job to do this, I have to be critical in order to learn and improve my designs. The images from this lens are usable to the vast majority of people, and promising for a first prototype, but on pixel level image quality there is a lot to be desired. I have seen older film lenses with similar resolving power to these photos, but they were used wide open and this lens had to be stopped right down to F/16 to be this good. My future prototypes have to address a huge amount of errors, specifically SA. The next lens I design needs to be usable at an aperture that doesn't require blinding sunlight. For future prototypes the aim will be to attain the level of detail seen here wide open or around F/4 max. This means heavy image correction will be needed, and that means only one thing: multi element lens designs.
