[Rovastar]

As someone who primarily uses XBCM to play music (on the Mac), I'm saddened by the distinct lack of visualization options. There are a few items in the menu, but honestly the only "real" choice is ProjectM. What options we do have are either simplistic our outdated.

Now, I know this is the internet, but I'm not just here to complain. I have a positive suggestion to contribute... honest!

Part of the beauty of open source projects like XBMC is that aspects can be opened up and handed off to others for expansion. Take fan art for example. Thanks to the efforts of fanart.tv, we gain access to an actively maintained (living) pool of content to improve our UI experience.

Why not do something similar with visualizations?

The core XBMC development team should not be burdened with the task of creating/porting/maintaining visualizations, but rather opening up the API so others can. Much like how we now install addons via an in-application UI, the same should exist for visualizations.

And rather than trying to create a new community of visualization developers, it may be a good idea to try and leverage existing communities, by making it easy for their visualizations to be imported. In this case, I am speaking specifically of the "demo scene" community.


http://www.demoscene.tv/
http://pouet.net/
http://scene.org/
http://en.wikipedia.org/wiki/Demoscene


Here you have an active community of artists and programmers that create complex 2d and 3d visualizations (many of which are designed to interact with music). A partnership here could really breathe some new life into the world of XBMC visualizations.


Thoughts?

/Chris

As someone that has been writing visualizations on and off for the past decade I thought I would give my insight here.

Firstly I would say across the board on all players/formats there are fewer visualizations out there. Very few people make them now-a-days and was a handful of people at its height.

I think the fashion of visualizations has past its peak. People just aren't as interested as your used to be. Also the barrier to entry is higher I feel, people want something looking as good if not better than is already available.

Also the barrier is high because the public are expecting the more fancy graphics, etc to compete with latest Call of Duty they see advertised on TV.

10 years ago it was possible to write something easily to compete with the latest games now not possible and mostly visualizations are perceived about graphical quality. Now they have teams of professional fulltime graphics artists just working on the textures you see let alone the entire engine and game.

I find this discourages many, in general, from taking up the untraditional coding route of visualizations.

Those that do something invariably get comments like:

"What options we do have are either simplistic our outdated."

As I understand it there are a few people that do make visualizations for the XBMC platform. Encourage these.

You have the open source reserve engineered MilkDrop clone ProjectM.
I think it is difficult to build on that if it as a baseline. The MilkDrop engine was a full time job paid from Winamp/AOL by a talented dev Ryan Geiss who then went on to code demoes for Nvidia and wrote the software for the tracking in the XBOX Kinect with Microsoft. Even then he just wrote the engine (that is key too an editable engine) and was not the most creative that came from the larger base at the time of the Winamp community and even then I can only remember 3-4 regular authors of merit at any one time.

ProjectM then built on that success and just ported an engine that was relativity easy to do, the difficult bit is getting the ideas in the first place.

I think it is unlikely to get the seemingly, with hindsight, perfect mix again.

For the demo scene that in itself is getting smaller and smaller year on year and they haven't ever embraced the visualization scene in the past I am not sure they will do that now.

But going forward that will look more and more dated over time despite authors finding new things to do with the engine. Also it is still fundamentally a 2D visualization despite the shaders and new other technology.

Writing 3d stuff is much more of a challenge especially if you want to have a editable engine so you can have scenes/presets made by the wider community that the most successful visualization have had in the past.

However you do have potential with your own visualization here Vortex.

I remember when that was being written by, I think MrC here, a 3D visual and based-ish on R4 visualization. However from what I understand (and that is little) there is not much community uptake for this. I imagine that that doesn't help/encourage more work in this area of creating a new visualization for the community.

3D based engines for visualizations haven't really been all that successful as there is more to learn and to make something unique will require more patience and OpenGl/DX like understanding/skills. In fact R4, was probably the most successful and then I can only think of me and 1 or 2 more that actually made any and shared it with the community and it does take a lot of documentation/work/design to create an engine and make it flexible. So much so that most of our latest/current visualization Morphyre has no editor for people. The question was should we spend x weeks making a system so 1 person might edit something when we could make y more scenes for everyone. Anyway back more onto topic.

I would either:

a) try and build on Vortex that you have here - by that I mean people try and write something for that. Look at others too like FishBMC and try and help out there.
b) make a plugin for running Winamp plugins, WMP plugins, iTunes, etc. However this require more dev work.

2 different paths and the second one is sadly not that compatible with the fully open source model.

Sorry we the long post, I haven't been here for a year or two and just thought I would share that hope I didn't bore you too much for all those that read this far.

Good luck.