The Simons Observatory, a group of microwave telescopes in the high desert of Chile, is starting to gather data to attempt to prove or disprove the theory of inflation.

Un-paywalled article from the NY Times.

A growing number of researchers in the field are using their expertise to fight the climate crisis.

The article spotlights several astronomers who are attempting to fight climate change, sometimes through changing careers.

NYT gift link, should be un-paywalled.

I love seeing the astro images posted here, but may I share an algorithm for making them even more beautiful?

Most astro images are created from separate red, green and blue images taken with electronic detectors (whether using classic BVR filters in an attempt to replicate what the eye might see, or some other combination in a "false color" image). There are two big problems that are common with the images created in this way (even by professionals).

The first is in the choice of stretch: how brightness on the detector maps to brightness on the displayed image. Most choose a linear or a logarithmic stretch. A linear stretch brings out fine detail at the faint end, but can leave the viewer ignorant of details at the bright end. A logarithmic stretch allows you to bring out details at the bright end, but not the faint end. Instead of these, choose an asinh (inverse hyperbolic sine) stretch, which is able to bring out both the faint and bright features. It scales linearly at the faint end and logorithmically at the bright end, giving you the best of both worlds.

The second is in the handling of saturation: how to display pixels that are too bright for the chosen stretch. Most apply the stretch separately in the red, green and blue channels. This makes the cores of bright objects appear as white in the color image, while they are surrounded by a halo that is more appropriate to the actual color of the object. The color of a pixel should instead be set by considering all of the channels together. This way, bright objects will have a uniform color, regardless of whether the stretch has been saturated in any of the channels.

See here for a direct comparison between the classic approach and this (not really) new algorithm on the old Hubble Deep Field.

If you would like to adopt this algorithm for your own work, there is a python implementation that you might find useful.

I've always considered the Australian accent to be fairly homogeneous across regions, but certainly there isn't the extreme diversity that the UK and America have. How much diversity is there, and what are the various characteristics? How long would it take you to tell a Cockroach from a Cane Toad when you meet him in the street?

... researchers noted the similarities between the game and the real-world pandemics. Both had an immediate impact on dense urban areas, which limited the effectiveness of containment procedures in stopping the spread of disease, while air travel, like fast travel, allowed infections to spread across large parts of the world with ease. Lofgren compared the in-game "first responders", many of whom contracted Corrupted Blood when they attempted to heal others, to healthcare workers that were overrun with COVID-19 patients and became infected themselves. While a direct analogue was not made to griefers [players who engage in bad faith multiplayer game tactics], meanwhile, Lofgren also acknowledged individuals who contracted the COVID-19 virus but chose not to quarantine, thus infecting others through negligence.

cross-posted from: https://lemmy.world/post/2207523

Kristen Browde has amassed over 380,000 followers on TikTok by debunking the anti-LGBTQ+ “groomer” myth, and no one is more shocked by her success on the platform than she is. “TikTok is supposed to be this app where teenagers do dance moves, but it turns out it's an incredibly effective political organizing tool,” she tells Them in an interview. “I am delighted that people are finding this useful information and are sharing it.”

Browde, an attorney and former correspondent for CBS News, posts weekly videos aggregating local media reports of children being targeted by sexual predators, sourced from keyword searches on Google and DuckDuckGo, as well as tips from followers. (She posts all her data at her website WhoIsMakingNews.com, where it's available for anyone to download.) In the six months since she created her first TikTok video, she has noticed a continuing trend: that very few of the perpetrators who are being arrested for preying on kids are drag queens or trans people, the targets du jour of the right. Instead, Browde found that the individuals who are responsible for the vast majority of reported incidents hail from the very groups pointing the finger at the LGBTQ+ community, from religious leaders to Republican politicians. “They’re the ones who are committing the crimes,” she says. “They’re the ones who are doing this.”