Comment to Dean Pesnell, Solar Dynamics Observatory about open lossless accessible data for global comparisons
Subject: Exactly how are the 4096_211193171 images made from AIA 211 193 and 171? 48 hour movies
Dean,
I am looking at these blended images, and they are helpful But I would like to be sure I know exactly what goes into them and also be able to change the composition.
https://sdo.gsfc.nasa.gov/data/
https://sdo.gsfc.nasa.gov/assets/img/browse/2023/06/28/20230628_045709_4096_211193171n.jpg
I asked Google Bard what you were doing, and it said you used SolarSoft but could give no information on coefficients or lookup tables or the exact steps and process used. Since there are also (304 211 171), (094 335 193) , (171 HMIB) blended images, I am suggesting you post an explanation of all of those right next to the images so people can immediately see how the image was created.
You are still posting lossy formats, so putting the higher resolutions is only giving more noise. It looks ok, but not good for correlations, comparisons, models and such.
I bought a new computer, and put Python on it, and am learning how to download the lossless images. But I think you could consider the scientific sharing needs on the Internet.
Right now you have (1) pretty pictures in lossy formats for people to look at, but the compression makes most all the pixel values wrong. (2) lossless format that are very difficult to find and use, and require lots of installations and background reading from very scattered documentation, some that seems to be missing.
There are JP2000 images, but that seems to be an incomplete effort. It MIGHT be lossless, it is hard to tell. I have not had time to find the lossless data and compare. You might hear the Python is popular, but there are only a few hundred thousand people out of about 5 Billion using the Internet that use it well. And when they go to learn it and apply to real problems the learning curve, uncoordinated teaching and support are very weak. And, from an “open” criteria view, it is difficult to impossible to trace where things are done, who created them and who if anyone, takes responsibility for what comes out. A lot of hackers sharing does not add up to stable global supported format and open data sharing that is suitable for 2 billion children learning for the first time, nor for increasing number of citizen scientists and more and more global research activities where data from many different sources is being compared and used. That last can only come from open lossless archives where the tools are immediately accessible to all browser users. About 5 billion now and growing.
I think the 211193171n and other blended images would make nice 48 hour videos. But processing lossy format seems a bit of a waste. And they are still pretty pictures until the meaning of the colors is posted and accessible. You do not have to put legends into the images. You can have a hover box on an “info” icon next to the image.
I wish I had more time to spend to give you specifics. But I can only say it is important to make ALL data like you are trying to share with “everyone” (about 8 Billion humans) accessible to everyone with the tools and background they have. Not force billions to replicate the graduate student path that most people ‘in the business” had to go through.
Richard Collins, The Internet Foundation
