We tend to think that newer is better and faster, but it's also wise to think that old apis had to be fast on old hardware. Hardware that is not the target of new APIs.
1,143 karma · joined September 6, 2019
In the sidelines buiding open-source: - https://audiomass.co (a full-blown audio editor in just 60kb of frontend js)
- https://github.com/bldbrne-vr/bldbrne-vr a full re-imagination of Bloodborne as a VR game
- https://pantel.is/projects/css3d/ (an old timey 3d adventure rendered entirely in CSS)
We tend to think that newer is better and faster, but it's also wise to think that old apis had to be fast on old hardware. Hardware that is not the target of new APIs.
On device (eg mobile phone) processing with battery usage that respects the user. Older hardware/models inclusion as well.
Of course the above aren't cases were the stream itself is 100+fps, but more of broader general benefits. For a 100+ fps stream.. well there are many things that go fast, imagine you wanted a robot that tracks or catches a fly before it takes off. Flies have a reaction time of 5ms (200fps), that's why it's hard for us to catch! Expand and apply the same concept to other things (that are fast, or happen very quickly) now...
Interesting pic. I think it tries to show the hdr, dynamic rage of shadows + cleanliness even though the person is in motion.
Of course this won't fix the fact that you have a lambda behind api gateway that does some heic->jpg conversion and can't be hit outside the DMZ, or some esoteric SQS queue that you can't mimmic locally - but it should get you almost there.
“If you want to build a ship, don’t drum up the men to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea.” - Antoine de Saint-Exupéry
The fear I have going back to the office is people just touching stuff, door handles, glassware in the kitchen, etc. Being on top of every tiny violation is a new concept and a huge lift for any small/random company. So who knows, maybe this could be a good opportunity for wework.
It's billboarding! (and the css lightsources are duplicated in the webgl world, so that the worm is properly shaded)
This is meant to be an experiment of rendering a full 3D scene in html/css. Of course there is a ton of (vanilla) JS to handle pretty much everything - but the main point is to explore what good ol' divs can do in a 3D context.
It might seem a bit off, because I am not a good artist/world builder, but it contains various features such as lights and shadows calculation (rendered on divs though background images + blend modes -- try clicking on the candle). Billboarding and sprite effects, geometry culling etc, as well as scalene triangle computation (which allows us to convert actual 3d models to html elements - https://pantelis.s3-us-west-2.amazonaws.com/0/css3d/readme/t...)
-----
If this interests you, I am giving a presentation on how it is built and lessons learnt this coming Wednesday (2nd of September) at CascadiaJS 2020 ( https://2020.cascadiajs.com/ ), among a line up of awesome speakers!
So... webgl is still the way! This is meant to be an experiment to show what css/html is capable of!
You can see it is very basic 90deg angle geometry, instead of actual pointclouds. Still pretty cool! https://imgur.com/a/EgC0RbN
Perhaps I should add support for the control key, but let other combinations were conflicts might arise. Getting good UX is a such fascinating topic.
Some of the most horrible web experiences, occur because websites try to take control of native functions. For example "smooth" scrolling, or highjacking the back button, or trying to abuse the clipboard! Confusing the user, breaking navigation, making the page less accessible and in some cases breaking after X months when browser or OS behavior changes.
What I have now is unfortunately far from "good" and there is improvement and experimentation waiting to happen, but the straight forward approach is not much better either...
Being able to label things with markers and different colors though would be fantastic! It's in the plans, once a bit better audio handling of slices (eg cross-fading, mix-dragging etc) is added!
There are decades of research, feedback loops and work put into tools such as FruityLoops, Cubase, ProTools etc, and the thought of reimplementing parts of it for web is such a daunting task.
And performance is fine, but it can be so but soooo much better! Like just by adding sprinkling some wasm or asmjs in the "onaudioprocess" loops, to speed up the buffer traversing loops and escape the dreaded garbage collector!
- First of all, I am amazed at the suggestions and love this is getting. I have been dog fooding it by using it to quickly edit foley audio from a tascam hand-microphone device, in order to make some cheap sound effects for a game project I am working on. The point is, I thought I was aware of all bugs, and all areas of improvement, and I am humbled to have my mind opened and see how valuable outside perspectives are! It's so easy get tunnel vision and think you know best I guess.
- Secondly... as I said I wrote this in June 2018, and just... kept it... I guess I was afraid of sharing it to the world, perhaps the audio people would get mad at me for making mistakes with the audio api (like the fade in/fade out being linear). Perhaps the javascript people would make fun of me for just using Vanilla JS.
But if this is impressive in 2020, imagine how impressive it would have been back in 2018! So I guess my point is. Share your work! Do not be afraid to put it out there!
- If anyone is interested on how it is built, and how the interface complexities are managed, even though it is just plain old school JS that has the reputation of being notoriously difficult to maintain, I would be happy to make a write-up shortly, or perhaps give a talk on it.
- Third... (hopefully that is ok). If you like AudioMass, and like the way it is built and it performs, perhaps you might enjoy working together with me. We are doing cutting edge computer vision, and well.. some CRUD stuff too! My company is hiring (info in my profile). But please be advised that due to covid-19 things may take longer or may not be fully up to date.
PS. As for license, I will probably choose something like "wtfpl.net". if it can help you learn something, or build something, go ahead! If you noticed, the page doesn't have any tracking (I realy don't know how many visitors came (I also disabled nginx logs)). And of course no ads at all. I 'm just trying to build cool and useful stuff!