- Joined
- Mar 6, 2019
- Messages
- 35
- Likes
- 16
Jim,
Sorry, not a troll. I have been a lurker for five years. My screen name is as a result of my admiration for Richard Feynman whom I considered to be the best physicist of the 20th century. As an engineer of long standing I was trained on scientific principles and statistical evidence. That's one reason why I agree with you that double blind testing is the gold standard.
Again, I have no wish to offend anyone.
I'm just pointing out once again that prejudgement of the equivalence of different samples without measurement of them is not very reliable. The underlying fallacy in your analogies is the implied condition that any given Dac being tested will measure no better than a $100 Dac in each case i.e., all Dacs over $100 dollars in cost will perform equivalently. From a statistical standpoint you can't say that for certain until you've measured all Dacs or at least all Dacs with the same components and software. Look, you may be right that they are all equivalent and that spending over $100 is a giant waste of money but until you've measured them all, you can't say that for certain i.e, a confidence level of 1.0.
No one, least of all me, is trying to contest scientific principles. I adhere to them rigorously. There is no scientific principle which states that you can measure some products in a marketplace and guarantee that there is nothing better now nor will there be anything better in the future. We often used to talk about the fact that the patent office was going to be shut down in the late 1890's because everything of importance had been invented. You write about fixed designs with tolerances being the only variable. What you are implying is that there is really only one design for a Dac and the only variable is component tolerances. I've spent years working on wavelets. That data sampling approach would make a fantastic Dac by focusing on sampling areas having the greatest rate of change in a given time period and wouldn't be limited by the current design approaches. Still, I'm open to being proven wrong. New and unforeseen designs and approaches do come along that upend prior knowledge all the time.
I think we've beaten this dead horse enough or as Mark Twain once said "There is little to be learned from the second kick of the mule." I had no idea that I would engender such controversy. This is my last post on the subject.
PS I was in high school as well when Sputnik was launched. As for nuclear reactors and water coolant, what about thorium based pebble reactors- no water coolant only graphite-moderated, gas coolant.
Sorry, not a troll. I have been a lurker for five years. My screen name is as a result of my admiration for Richard Feynman whom I considered to be the best physicist of the 20th century. As an engineer of long standing I was trained on scientific principles and statistical evidence. That's one reason why I agree with you that double blind testing is the gold standard.
Again, I have no wish to offend anyone.
I'm just pointing out once again that prejudgement of the equivalence of different samples without measurement of them is not very reliable. The underlying fallacy in your analogies is the implied condition that any given Dac being tested will measure no better than a $100 Dac in each case i.e., all Dacs over $100 dollars in cost will perform equivalently. From a statistical standpoint you can't say that for certain until you've measured all Dacs or at least all Dacs with the same components and software. Look, you may be right that they are all equivalent and that spending over $100 is a giant waste of money but until you've measured them all, you can't say that for certain i.e, a confidence level of 1.0.
No one, least of all me, is trying to contest scientific principles. I adhere to them rigorously. There is no scientific principle which states that you can measure some products in a marketplace and guarantee that there is nothing better now nor will there be anything better in the future. We often used to talk about the fact that the patent office was going to be shut down in the late 1890's because everything of importance had been invented. You write about fixed designs with tolerances being the only variable. What you are implying is that there is really only one design for a Dac and the only variable is component tolerances. I've spent years working on wavelets. That data sampling approach would make a fantastic Dac by focusing on sampling areas having the greatest rate of change in a given time period and wouldn't be limited by the current design approaches. Still, I'm open to being proven wrong. New and unforeseen designs and approaches do come along that upend prior knowledge all the time.
I think we've beaten this dead horse enough or as Mark Twain once said "There is little to be learned from the second kick of the mule." I had no idea that I would engender such controversy. This is my last post on the subject.
PS I was in high school as well when Sputnik was launched. As for nuclear reactors and water coolant, what about thorium based pebble reactors- no water coolant only graphite-moderated, gas coolant.
Last edited:
