This would especially be good for me as I do the weather in Nebraska from a studio I built at home in SoCal. As far as I know I'm the only regularly scheduled TV news anchor to work from home in the US and possibly the world.
Do you have any problems with this remote setup like Sinclair has with “Centralcasting” local news from far away? (Graphics and visuals for the wrong station showing up, talent’s obvious lack of knowledge of local events and terminology, etc...)
What happens in the event of a California-centric disaster like an earthquake or wildfire, and you’re not available? Or if there’s a major weather event in Nebraska and you’re not there, or not able to feed into Nebraska because Vyvx/internet goes down or the satellite dish freezes over?
It's difficult to explain but what I do is designed for me. There's a lot I do because I understand both sides of the equation which means my studio isn't a turnkey solution anyone could walk into. You need to be able to understand 'why' when something doesn't work.
That being said, IMHO Sinclair did it the wrong way. The weather unit should be an autonomous production, not dependent on the regular studio being free. My all-in studio cost is around Hyundai money. It removes complexity and allows live production where needed and closer to air always.
I had a trip from SoCal to Milwaukee cancelled because a few guys who robbed a gas station decided to drive the getaway car to Sky Harbor Airport in Phoenix -- where my first flight was headed. Random shit happens.
In three years of doing this what you describe hasn't ever happened.
We use the Internet for data. I have fiber to the wall.
How did you end up building such a studio at home etc? Sounds like a lot of work to me!
(Always cool to see what kind of people hang around here on HN as well!)
Maybe the most unusual part is, none of this would work unless I could run the show by myself. My TriCaster is programmable with macros which is what I've done. I am my own director while on the air.
Oh -- with a friend I created a map making system which runs on an i5 and produces ~ 40,000 maps a day. Here are some samples I threw together a few months ago. These are 100% produced using FOSS including the map databases and fonts! https://drive.google.com/drive/folders/172O4Xl35np8RnbRi07PD...
My last computer class was as a senior in high school, 1967-68 semester. This was BEFORE computers had screens.
https://photos.app.goo.gl/wPD2M1yrGZ6ZsEFt7
We park on the street. :)
I am a professional. My job is to forecast their weather and I am diligent in that pursuit. Windows are overrated.
I mean, the rain effect itself, not the whole UI around it here.
Edit: Here's[1] the accompanying article, and here's[2] some other stuff I made (although my website is also just as old and hasn't been updated since!)
[1]http://tympanus.net/codrops/rain-water-effect-experiments/
Thank you for finding that.
Welp. It's identical. I even tried to defend the current link with "Well, some cool things can happen when you hook up a UI to a shader..." but the UI doesn't even do anything except turn the rain off and on.
There's lightning too in that other one! That's so cool.
As an aside, I’m sorry someone stole this from you and used it unattributed.
You've probably come across it though, as it does seem to gloss over some basic stuff. But it does cover the basics pretty well, just gotta keep an eye for details and seemingly unimportant one-liners that turn out to be fundamental.
It also takes some googling for deeper explanation of some functions, which often result on Stack Overflow answers by the author himself.
Dealing with the WebGL API itself is a bit... let's say tricky, though. But working with shaders for me is a joy.
This also seems promising, though I only skimmed over it: https://thebookofshaders.com/
And also, look into https://shadertoy.com/. Both for using the shaders there as example and as a playground.
[1] shader: https://www.shadertoy.com/view/MdfBRX
I wrote an article about it here[1], but I made this back when I was first learning WebGL (it was my first WebGL project, in fact), so be aware that a lot of things could be better!
[1]https://tympanus.net/codrops/2015/11/04/rain-water-effect-ex...
On osx mpb performance is stellar in Chrome sadly, while both Safari and Firefox only delivers 2-3 fps.
Weirdly on an iPhone with Safari the performance is smooth as silk.
I would really like to ditch Chrome, but it seems i get full speed fans with Firefox on sites sporadically, and while Safari is much better it's still not on par with any kinds of heavy graphics.
I don't get it. Why can my iPhone 6 run this super smoothly while my 3000 USD mpb struggles? And why can Chrome do it while Safari can't when it should have native access to GPU api's ?
Somebody has made a version using WebGL alone[1] (sadly they got the refraction a bit wrong), try it to see if it works well enough (and let me know, I'm curious myself).
Would be pretty weird if the osx native browser didn't hardware accelerate while Chrome does it on a MPB, but Apple are often weird like that.
Anyone tried it on Firefox + Windows?
[1]: can't verify this as I only have a slowich some years old android phone, on which it runs fluidish but miss renders the effect of the large rain drops...
I was confused at first because I thought the sliders were parameters of the simulation and I wasn't hearing anything.
Curious if this is the best done in WebGL to date, or if there's even better out there? Or what this does and doesn't do, and where next obvious areas for improvement would be?
I actually came here to comment on the blur on the edges of the large drops. My guess was it was a fast way to give a passable result for the more complicated refractions occurring at the edge of the drop. Is there a link at which I can follow your work? Thanks!
This vfx is actually pretty simple but quite satisfying. Having the drops refract based on look angle is what really pushes it over the edge.
The realism is so-so. Big raindrops do not accumulate smaller ones well enough when touching. And there is no splash.
You can see the app at https://rainpaper.com
Awesome app overall though!
https://play.google.com/store/apps/details?id=org.oftn.rainp...
Since seeking is not required, it could do away with keyframes entirely.
Something about the distribution of the droplets and the movement doesn't seem "random" enough. it seems pretty "regular" for lack of a better word.
The drops themselves look amazing.
Example (it's using an IP for some reason): http://142.44.247.245/sounds/rain/rainbest160.mp3
Not sure what the license is for the sound, so can't say whether it's free or not.
For playback it's using jplayer, which is open source: http://jplayer.org (I'm surprised their site doesn't list support for Opus).
I'm wondering which would be more interesting to watch. Perhaps a combination of the two.
Maybe a better middleground is to use an animation with transparency that you can instead overlay any background, like a looping gif.
Try this version for comparison: https://codepen.io/stefanweck/pen/Vbgeax
It seems like when you change the settings, it can take 15 or more seconds to take effect.
