My $.02
Transports: Real Time error prone reading of physical media by optical means (That's Shiny Platters to you) vs "Computer Based" HD / SSD sourced bits.
Any system that cannot reliably output the sourced input bits is broken by definition. It either works perfectly or it fails.
How would you feel if your bank only accepted some of the bits of your paycheck?
The HDCD light test is a good test for bit-perfectness.
A decent transport should deliver the right bits at the right time. This is where we get into jitter. Lots has been said on this, but recovering an invariant clock from a push based protocol with no error correction / retransmit control seems somewhat backwards compared to a local clock based pull protocol - but thats hard to do with unbuffered real-time reading
Does any of this make a difference.
Bit correctness: Yes - its either right or wrong.
Jitter: Possibly - If you can't correctly differentiate transports to a statistically significant degree - then by definition its imperceptible to you. Save your money and be happy.
If you can, then you have your answer.
If you are undecided, and want confirmation of your biases, we''re here to help.
If you want an objective measure of output differences may I suggest Audio DiffMaker
Other people have suggested perception is reality, and there are many, many things that alter perception. You might enjoy some of them.
Final words: Whatever makes you happy. Me? I'd be happy with some Kaizen, so we end up with demonstrably "better" products at lower prices over time. How hard could it be. Look at what Moore's law has done to your 'phone. I live in hope.