Let's say scientists want to record what happens to me when I am in pain. They record me getting poked, and a highly sensitive machine models exactly what happened. Assume that they know what every particle did during the time I am in pain. Their full physical account would not include the sensation of pain that I felt.
Of course it would. They may not be able to identify that that's what they are looking at, but nonetheless it would be there.
I have a picture of my dog on the hard drive of my PC. If I look at the hard drive at the molecular level, there are just iron, oxygen and silicon atoms - and nothing that looks remotely like a dog.
If I use my knowledge of hard drive design, I can extract the binary ones and zeros that constitute the picture file; and they too look nothing like a dog.
Yet the picture of my dog is real, and physically exists on the hard drive.
Pain is a software phenomenon. It cannot be directly detected by examining the hardware; but that doesn't imply that it exists independently of the hardware.
With adequate knowledge not only of the physical state of the hard drive, but also of what a hard drive does, and how it does it, I can reproduce the picture of my dog from the information on the hard drive. Not only that, but that 'adequate knowledge' can itself be derived from physical examination of the hard drive - the system contains enough information to reconstruct not just the picture, but also the protocols and encoding required to reconstruct the picture, and the basic instruction set that allows reconstruction of those protocols and encoding techniques. The system can 'pull itself up by the bootstraps', and generate all of the needed software purely from the hardware present; Understanding this process at the fine detail level of ones and zeros (or even at the level of iron, silicon and oxygen atoms) is VERY HARD INDEED, but it is demonstrably possible - it happens all the time.
The same is true of the brain; understanding exactly how to get from a bunch of neurons exchanging nerve impulses to the sensation of pain is VERY HARD INDEED, but you don't get to claim that it is IMPOSSIBLE unless you can PROVE it - particularly as we observe it happening all the time.
It is the structure of a computer that translates ones and zeros on a hard drive into clear and detailed pictures on a screen or printed on a piece of paper. It is the structure of a brain that translates nervous impulses into qualia. To do so requires multiple levels of recursion in both cases, as the system has to 'pull itself up by the bootstraps' to render the hugely complex end result using a trivially simple starting point - this is where the word 'boot' comes from to mean 'start up a computer'.
Your argument is not evidence for a non-physical consciousness; it is an argument for Ryan's inability to grasp all the steps needed to get from the structure of the brain to qualia. But your inability to grasp all the steps, or how a series of small steps can build up to a hugely complex set of emergent properties that are not discernible when looking at individual particles, is not evidence for anything other than your ignorance.
You can't understand a forest just by looking at a tree. You can't understand a tree just by looking at a leaf. You can't understand a leaf just by looking at a chloroplast. You can't understand a chloroplast just by looking at a magnesium ion. None of this means that a forest is non-physical; nor that a forest cannot exist without some magical woo that falls outside the Standard Model.