Are people still using AVS, Milkdrop etc.?

[jheriko]

Just curious as I am planning to start work on new visualisation software (I can't disclose too many details, sorry) soon and just wondering if this stuff is still as popular as it used to be.

Surprised this place still exists.

Also, check out my blog and its vaguely AVS related stuff:

http://jheriko-rtw.blogspot.com/2009/05/whittakers-method-and-brief-history-of.html

[Yathosho]

i think the download numbers on winamp.com, the submission rate on deviantart and the amount of comments in this forum speak a clear language. yes, there are people using those plugins, but there aren't many of them. i guess the real question is, who still uses winamp these days?

anyway, that visualization software. is that the kind that never sees the days of light or the one with the source code you will "lose" after the first release? seriously, every couple of years someone pops up (usually people that haven't been around for ages) and promise to 1) revive the avs scene or 2) that they're working on an alternative visualization plugin. it's always a bunch of bullshit, nothing ever gets released, there aren't even any following annoucements. the only serious effort i recall was fragmer's avsx, which in the end was sacked.

plus, nobody needs a plugin for just one player, least of all winamp. nullsoft seems only interested in the development of winamp for android, which has nothing in common with winamp except for the name. if you want to build a visualization software these days you first got to look around what's available and evaluate if it's necessary to start something new or if it wouldn't do better if you put your efforts into an existing project (there are PLENTY). okay, you made your decision, you want something new. why limit yourself to a platform? think big, bigger than a plugin for software x and bigger than for just one operating system. that jscript version unconed showed off recently made me think about how web technologies can be used to build music visualization. it maybe too soon, too slow, but it's the future. all major browser manufacturers are putting most their effort in improving jscript performance, but then there are alternatives such as nativeclient... etc etc etc

back to the issue. i can accept a serious call for support, tell us what you want to do, tell us where you need help, at least have a project page on github/google code/sourceforge. if you develop on your own, shut your mouth until you got something to show! (sorry if that sounds rude, i don't mean to insult you personally)

there are people that could join your cause, but as the fridge disaster showed that doesn't mean anything. if i were you, i'd talk to people like fragmer, grandchild and micro.d! fragmer got quite far and is quite knowledgeable when it comes to CUDA, grandchild showed some interested in developing something himself, and i'm sure micro.d can contribute in some form as he's quite active in the vvvv scene and as a person who's bringing visualizations to the club as a "vj".

[Xetick]

I can't speak for the rest (avs, milkdrop) but I can backup my own Plane9 visualizer with some real numbers. Since the release that included the stats gathering, about 1 year ago, the winamp visualizer has been running for 51 years totally. I would guess avs and specially milkdrop has magnitudes more usage than that.

But from what I can see so does it look like user generated content even for those famous programs are released less and less. So if your working on a new visualizer where users can create their own content, keep in mind that it's difficult to get people to contribute to it these days.

[Raz]

Cool, I'd play with some new vis stuff if it showed up. I did a few nifty experiments with html canvas and the experimental firefox audio buffer API recently and felt a bit of that old excitement.

[Yathosho]

oh btw -> http://forums.winamp.com/showthread.php?t=335388

[Xetick]

Cool, I'd play with some new vis stuff if it showed up. I did a few nifty experiments with html canvas and the experimental firefox audio buffer API recently and felt a bit of that old excitement.

If you haven't tried it you could always give Plane9 a spin on a rainy day. It comes with quite a few scenes from the start and all can be shown and edited in the included editor.

[jheriko]

i think the download numbers on winamp.com, the submission rate on deviantart and the amount of comments in this forum speak a clear language. yes, there are people using those plugins, but there aren't many of them. i guess the real question is, who still uses winamp these days?

no one i think, but lots of people have smart devices these days...

anyway, that visualization software. is that the kind that never sees the days of light or the one with the source code you will "lose" after the first release? seriously, every couple of years someone pops up (usually people that haven't been around for ages) and promise to 1) revive the avs scene or 2) that they're working on an alternative visualization plugin. it's always a bunch of bullshit, nothing ever gets released, there aren't even any following annoucements. the only serious effort i recall was fragmer's avsx, which in the end was sacked.

i assume you refer to the gvm source code i lost?

fridge stagnated because it wasn't going where i wanted and nobody really contributed except pak-9 who i think tried reviving it recently.

this is the kind i do at work, where if a prototype is successful, it might happen. no promises

plus, nobody needs a plugin for just one player, least of all winamp. nullsoft seems only interested in the development of winamp for android, which has nothing in common with winamp except for the name. if you want to build a visualization software these days you first got to look around what's available and evaluate if it's necessary to start something new or if it wouldn't do better if you put your efforts into an existing project (there are PLENTY). okay, you made your decision, you want something new. why limit yourself to a platform? think big, bigger than a plugin for software x and bigger than for just one operating system.

thanks, but i do sort of get the big picture. the cross platform techniques of my game engine (which works and is mostly publicly available) have taught me everything i need to know about how to not be restricted to one platform or specific os etc.

there are people that could join your cause, but as the fridge disaster showed that doesn't mean anything. if i were you, i'd talk to people like fragmer, grandchild and micro.d! fragmer got quite far and is quite knowledgeable when it comes to CUDA, grandchild showed some interested in developing something himself, and i'm sure micro.d can contribute in some form as he's quite active in the vvvv scene and as a person who's bringing visualizations to the club as a "vj".

well, like i said, it would be through work, and might not happen at all. i was just curious if anyone had a good handle on how popular this stuff is, because the available software is mediocre at best.

p.s. you might not have seen in the news how little spare time i had to devote to any kind of home programming etc. over the last year. that really didn't help p9s attempt on fridge. which btw, is on google code, but is an overengineered pile of complicated ui made to satisfy fringe users instead of ship a product.

[Grandchild]

I am using AVS. In fact we are, Hurricane and I. See http://vimeo.com/effekthasch for our VJing efforts.
micro.D is using AVS in a very similar way, regularly VJing in clubs.
micro.D even made a neat AVS Mixer in VVVV, which I rewrote and extended and which we are using for our Effekthasch-gigs. It can be found at https://github.com/grandchild/AVS-Mixer. On that website you can also find the detailed description of our setup, if anyone is interested.

As for rewriting AVS - I, like so many others, started a project, and I have a lot of ideas that I want to incorporate. I am constantly - although very slowly - working on that. A snapshot can be found also on my github page. Technical details: A GUI in Qt and a backend using SDL, planned to run on Win/Lin/Mac... As with all the others, I have no clue if I'll ever finish that, but right now I have every intention to do so.

So long,
Jakob/Grandchild

[Yathosho]

i'm currently trying to get a hold of fragmer, because i think you guys should definitely exchange your experiences, even if you want to keep developing your own projects. from what i remember (and this was at least 2 years ago), fragmer was working on a cuda-based avs-clone. eventually he ran into some problems, and i don't remember if these were only related to keeping backwards compatibility. however, the problem was marrying gpu-based tasks with cpu-based task. apparantly, certain things still performed better on the cpu.

as i said before, to avoid going through common mistakes and to exchange some good practices, you guys should communicate if you want to bring your project anywhere.

i think it's good to think outside of winamp and even outside of windows. which is why recruiting people (and looking for advice) on this site only will not be enough. i'm not sure if vjforums is the optimal platform, but it's certainly one step ahead from here.

[Yathosho]

btw, here's a video of avsx in action

http://vimeo.com/9784539 (show me)

[Tuggummi]

Things prohibiting a lot of the new people to truly enjoy AVS and consequently a lot of AVS-Artist to make new stuff is that the software is outdate, it doesn't even work properly with new winamp versions (you have to downgrade your winamp to get best results) and all that just leads to a niche community still stuck in the past sort of said.

And before you go off flaming, Im one of those people who still use AVS as it should be used :P

[Raz]

The way to get a visualiser going again is to make it first of all accessible, then you need an accessible community platform, then you need flexibility. In that order. AVS grew up being the accessible visualiser and it died being an obscure hidden thing buried down here in winamp with a bunch of cryptic inaccessible math-heavy guides that newbies got roughly directed to. The software being outdated is a side effect of nobody giving a crap anymore, you could pack it full of flexible complicated fast and efficient new features and it still wouldn't get used because it's not accessible or visible.

[Grandchild]

Raz, what do you mean by accessibility? Do you mean a simple, easy-to-use interface? Open development? ...?
Also I don't know about the math-heavy guides - those were bound to happen from the start, because a lot of the stuff in AVS could only be done "on foot" - like 3D and scripted animations and that kind of thing, where you would expect some helper functions or utilities from a software nowadays, to help you with tedious standard tasks - like 3D rotation. New people being "roughly directed" to those guides, that's another question. But Pak-9's guide was math-heavy to most people perhaps, but it was not inaccessible - in fact I found it to be rather well-written.

But yeah, I am really interested in what you would like a "new AVS" to be like, to be "accessible".


@Yathosho
If you ever do get a hold of Fragmer, that would be great - would like to hear how that avsx stuff went...

[fragmer]

Hello; Yathosho got hold of me. "avsx" was a little proof-of-concept project, which used OpenGL and revolved around EXT_framebuffer_object.

Presets were made up of "nodes", which could be connected into a directed graph. There were three types of nodes: producers, filters, and consumers.

• Producer nodes only had output(s), and included: previous frame's buffer, saved user-defined buffers (similar to Misc->Buffer Save), bitmaps, render modules.
• Filter nodes had both inputs and outputs, and included transforms (done by mapping textures onto a polygon rendering it back to itself), convolution (shader), blenders (merged two inputs into one output using a function), and color filters.
• Consumer nodes only had inputs, and included "Display" and buffer-saving.

Each input accepted exactly one link/connection to an output. Each output could be connected to zero or more inputs. Original plan was to have a WYSIWYG editor, like this: http://fragmer.net/temp/uiconcept.png

Call stack was constructed at runtime by iterating over the graph, figuring out a call order that satisfied all input/output dependencies, then allocating and assigning buffers. Circular node connections were not allowed. A stack function pointers was created. I wasn't well-versed with shaders at the time, so it was mostly immediate-mode OpenGL, not very efficient.

EDIT: If I were to redo avsx today, I'd go with OpenCL/OpenGL combination code. This will allow keeping all buffer data on the GPU, and still doing general-purpose computations. I am familiar with the basics of this setup, but doing advanced things (like recreating Winamp's scripting language for use with OpenCL) is way beyond me.

[Raz]

But yeah, I am really interested in what you would like a "new AVS" to be like, to be "accessible".
.

What I mean by accessible is how easy it is to find out about and see, and how easy it is to start using it. Right now with AVS, to even SEE presets you either have to look at a recording on youtube, or you have to go and find and download a huge program, Winamp. Then you have to go into the plugin settings, switch the default visualiser to AVS from milkdrop, you have to know that the visualiser exists also and how to get to it. Then you can open it up and see a few presets that cycle automatically. Etc etc etc. It's a nightmare of unexplained hidden features and it's a miracle anybody found it to begin with.

edited: Relevant point, I don't even look at AVS anymore, and not because I don't want to, it's just that whenever it strikes me, I think about the bother of installing it all and going around my existing music library direct to the files and migrating it into winamp, and forget about using spotify or one of those. Then I just don't bother, can't be fussed with all the hassle.

By accesible I mean you can click a thing and see visualisers to music that people are already playing to start with. Not through an arbitrary player. New presets would be browsable and viewable from the visualiser itself, use new or popular presets buttons, see more from this artist buttons, etc. A usable interface for playlists. And that's for viewing. For building it needs to encourage people to tweak what already exists. Off the top of my head, have a "remix this preset" butto and have remixes show up like video responses on youtube. Spend more time developing tools and options with a usable GUI to customise and tweak so that it doesn't need as much coding to make anything new. Presets would be tied to a central location integrated with the program with a proper rating and comments and remixing system. From there you can tie into social media using the facebook api and etc etc etc.

I'm just pulling ideas out of my ass here, if I sat down and put some thought into it we could really get a plan going. I'm not the man to code a visualiser though, the hardcore coding was never my strong suit, much as I tried.