[50.1 PMMA @ F1]
What a mess. There's no getting around the fact that the image is unusable. I was personally hoping that F/1 would be a little better than this, but I always knew it was never going to be good enough for clear photos. To analyse this, lets list the good and bad points (because in-fact, it isn't all as bad as it looks):
Good points:
- While it may be hazy, the lens does render an image. Lets not forget that this is the first lens I've ever made, and requirement number 1 was for the optic to at least focus at the correct focal length (50mm ± Focus breath). While not crisp, you can see that the image is at least in focus, with contrasting lines between the bottle and the background, and the ukulele and the background appearing strong and defined.
- These test photos were shot using a Canon 5D Mark II with a full 36x24mm sensor. This requires a large image circle, meaning that the second plus point for this lens is that the image circle rendered by the optic is big enough to cover any 35mm based DSLR sensor area. Also, while it is difficult to tell and maybe not even appropriate to asses at this point, there seems to be little to no vignetting (darkening of the corners caused by reduction of light intensity near the edge of the image circle).
Bad points:
- There is one huge error present over the entire image which unfortunately masks both the potential good points of a lens (sharpness, contrast, colour rendition) and the bad points (sharpness again, CA, distortion), making this first test shot not as informative as I would like. The error is called Spherical Abberation, and in the case of this lens wide open, it is so bad that it ruins the image.
To explain what SA is, I have borrowed Wikipedias diagram above. At the top is a 'perfect lens'. In an ideal world, this is what we would have, a lens that renders all entering rays to a single convergence. This would result in a pin sharp image. However, the design of this optic which I am testing has spherical front and back surfaces, based on a fixed radii. Every material has a refractive index, and this effects the angle in which the light will refract when it passes through a surface of the material. In this case, due to the spherical surface, light refracting through the middle of the lens will not converge at the same point as light that has refracted nearer the edge of the lens. Therefore, as you can see in the bottom ray trace, the light rays are somewhat scattered and don't all converge on the image plane (the vertical line on the right hand side).
This is exactly what we can see happening in our test image. The closer the entrance rays are to the optical axis of the lens (dead centre), the closer they converge, which is why we can see an image and not just a complete blur. However, the rays coming in through the outer portions of the lens converge way out infront of the image plane. This means that by the time those rays actually make it to the sensor, they have passed convergence and have carried on in their own direction away from each other, contributing to the intense haze and fuzzyness we see across the entire image plane.
Just to prove this isn't all theory here is a ray trace of the exact lens design I made in the optical ray tracer software package called ZeMax, showing how the rays coming in from the edges become scattered:
ZeMax allows us to do something else quite helpful; it allows us to generate image simulations giving us an impression of what a photo would look like if taken though the lens. Using the same configuration as above, this is what was generated. Don't the errors look familiar?:
On the whole we can already draw one positive conclusion from the testing. The simulations of the design, and the photos taken with the lens in real life, both show exactly the same positive and negative points, meaning that the lens does exactly what it is supposed to do. The limitations show in image quality are not dependent on my skill or the material used, but the effectiveness of the Lens Makers formula which I designed the lens with. Put simply: to make a fast wide aperture lens, the lens makers equation is not good enough, you need to look into corrected formulas like asphericals.
