@#&%!!
When it comes to my experiments, nothing ever seems to be as easy as I'd expect. I don't want to say "easy as I'd hoped" because it's no fun if it's easy. But, there are times when I wished that when I think I've finally figured out the solution to a problem, that would, infact, be the time that I actually really have figured out the solution to a problem. For example, I came up with this really cool code that up until now worked amazingly well at noise reduction when I'm processing recordings from my rig. Then I tried this new technique that would allow me to take better data than my previous measurements but require only about 1/4 the time (3 hours instead of 12).
I made the measurements last night, (got home at 11pm instead 7am like normal) and am processing the data today. And just looking at the data, I can see it is good. I was hoping i could just sort of plug it into my super code, and voila, out would pop these amazing results. I came into the lab this morning imagining how great it would look.
But when I use my code to compute actual values I get a huge mess. It doesn't make any sense. If I take a sort of defunct, older version of a code I used to process other data with, and sort of parse up the measurements and take out the good portions then stick them all together, I can get a good picture of the measurement. And, as suspected, the measurement is perfectly fine.
So why isn't my totally butt-kicking adaptive moving power spectrum swept sine analyzer code, that makes cool snake-like pictures while it's processing the data, coming up with anything better than a scatter-plot????
And this concludes my Saturday.
I made the measurements last night, (got home at 11pm instead 7am like normal) and am processing the data today. And just looking at the data, I can see it is good. I was hoping i could just sort of plug it into my super code, and voila, out would pop these amazing results. I came into the lab this morning imagining how great it would look.
But when I use my code to compute actual values I get a huge mess. It doesn't make any sense. If I take a sort of defunct, older version of a code I used to process other data with, and sort of parse up the measurements and take out the good portions then stick them all together, I can get a good picture of the measurement. And, as suspected, the measurement is perfectly fine.
So why isn't my totally butt-kicking adaptive moving power spectrum swept sine analyzer code, that makes cool snake-like pictures while it's processing the data, coming up with anything better than a scatter-plot????
And this concludes my Saturday.
posted by tmm at
9:01 PM
3 Comments:
"I wished that when I think I've finally figured out the solution to a problem, that would, infact, be the time that I actually really have figured out the solution to a problem."
You just described 90% of exploritative science i.e. research. Keep it up and you'll eventually get the 10% that works. Or is that just for biology?
Heh - Jacob, not sure if I'd have the patience for exploratory science like that. Part of what keeps me going is the fact that I think I can do it. If I only had a 10% success rate, I would probably get fed up. I dunno though... do you live for the 10%?
No, I try to live for Christ. The 10% just pays the bills.
Post a Comment
Subscribe to Post Comments [Atom]
<< Home