|
|
#41 |
|
wellspring of milk
Major Dude |
I'm a recent (2010) graduate student from Germany and I'm working as a software developer. I know Winamp since version 2.x or so when my father's friend brought a mp3 CD from a russia trip with a Winamp installer, some albums and an encoder tool. My old 80486 needed over ten minutes for one CD track! I've made my first "programming experiences" on a 80286 with Logo turtle graphics, but you can't take that too serious since i was only 10-12 years old. Later I've learnt TurboPascal and then came the internet boom. My first PC had a 56 kbit/s modem from when they just became available. I taught myself HTML and Javascript "DHTML" in a few weeks and then i tried scripting browser and SVG Dom. About that time i had also discovered first the original Geiss visualizer, AVS and Milkdrop.
So you know Forth? My dad is an active member in the community forth-ev.de, (oh and we'd have one such GA144 evaluation board handy: http://www.greenarraychips.com/) In fact he's programming microcontrollers and embedded systems for over 20 years: fortech.de - recently also more industrial image processing and computer vision / tracking systems, mostly with C++. I studied major in Object Oriented Systems, and the model-driven development courses. I have collected quite some experience with Java, Eclipse (RCP), and enterprise application and webtechnology. I have also coded Flash and Actionscript for a while, but not so much the animation parts. And since not yet a year I'm programming together with a small team on a back office administration software in C#. We're reverse-engineering an old spa town tourism administration software from 4d.com to .NET - I'm in charge of the data access layer. I can well remember the time before Milkdrop 2.0 and the pixel shaders: See my uncanny introduction to the forum here: http://forums.winamp.com/showthread.php?t=269749
Go Home! | get your stuff done | Global Change 2009 | WebGL Playground | OpenProcessing | studio sketchpad | Twitter @ Google+ @ Facebook Last edited by Flexi; 6th July 2012 at 21:52. |
|
|
|
|
|
#42 |
|
Junior Member
Join Date: Nov 2011
Posts: 39
|
@Flexi: Wow! 144 computers on a chip. That is mind blowing for the programmer. And I like your introductory presets. Unfortunately I like them all and one could spend a lifetime in just looking at those available. It sort of could become an obsession always finding new ones, like sheep looking for new grass.
It strikes me that there are two distinct paths with MD. One is time related and the other music related. The former is simply the mesmerism of lava lamps. The latter is far more interesting to me and therefore the aim is to reduce the time element and increase the faithful? following of the music... @Nitorami:... and what you have to say unfortunately puts a limit on that. I find the responsiveness pretty good and have been playing with simple coloured shapes but am always defeated by the three channel overlap. The mid seems to respond to everything. The more one tries the more it forces each preset to be played with a specific music track that it works well with. Narrow Bandpass is required. Resignation is the key here. Don't waste time on the impossible, make do with what you have. I actually want something that I can use, not make. I want to play music and watch without tweaking constantly and missing the music and the effect. Impossible. I also seem to be becoming a rambler (and not with my legs). |
|
|
|
|
|
#43 |
|
wellspring of milk
Major Dude |
I'd still give the existing beat detection codes for the predefined bass/mid/treb bands a chance, to be reused in custom bands from the spectrum curve data. Maybe i can align them all next to each other in the particle preset when I'm in the right mood. Don't expect anything too soon.
|
|
|
|
|
|
#44 |
|
Major Dude
|
@Donone
1. Resignation is not an option ![]() 2. If your analog filters are sharper, then they are slow, which can mean something like 100ms perhaps. You may not have noted this because incandescent lamps are slow in the first place. 3. Some instruments DO cover a large part of the spectrum, snares etc. certainly do. 4. I agree that one does not want to constantly tweak the preset with the music, although definitely a decent DJ should select presets fitting to the music. 5. You can do much more exciting things with MD than simple flickering lights. I personally prefer making presets a lot over using them, but that is of course a personal attitude. 6. Technically what contributes to the apparently better frequency separation of your light bulbs is probably their nonlinear response. Light bulbs will hardly start even glowing when operated below 1/3 of their rated voltage, plus they are much slower in the low voltage range. Their brightness is proportional to something like the square of the voltage, or possible even the third power. Attached a sample of how you might simulate this in MD. Not all steps may be required. |
|
|
|
|
|
#45 |
|
Junior Member
Join Date: Nov 2011
Posts: 39
|
@Flexi: I look forward to your mood change.
@Nitorami:I must seem rude not returning but have some legal problems that take priority. Having carefully read what you say, I am sure that you are correct. 1. True! 2. Equally true. 3. True also, but, I (until now) have not been interested in every instrument or band responding because all colours would virtually be on together. On the other hand it is not possible for MD to emulate the brain and ears and pick out only that which is 'perceived' as being visually predominant. Generally one can watch coloured shapes and hear the music and know whether the two are 'in sync (opation)'. I have been (past tense) striving to achieve that. I can equally see that it is probably impossible. To follow a bass drum is OK but to respond to each 'beat' of a trumpet tune and ignore the wiggly bits in between is something else, the frequency might be the same but how to discard the wiggly bits?? 5. Yes you can. Back to respond to music or time. 6. Absolutely spot on but I had not given it that consideration. So if one were to pursue simple lights one would need attack/decay, envelope on brightness. Have not had time to download your link but will soon. Thanks I have not gone away, but perhaps I should! Be with you after some more legal. |
|
|
|
|
|
#46 |
|
Major Dude
Join Date: Nov 2003
Posts: 979
|
In the milkdrop beta (the one redi jedi programmed) we had reassignable htz frequency bands I think it was "sound (low value,high value)". That way you could make things respond however you wanted to outside the confines of bass mid and treb. We also had variables that you could assign to midi controllers and keyboard commands. These things made milkdrop much more VJable.
For me at least, the optimal way to do this is to have your app be able to play more than one audio file at once. This way if you make your own music, you can split the tracks you want stuff to respond to into different groups. So your beat is on one track, the bass on another, and so on and so forth. This is how we want to set things up in Biome, but we don't know how to get the currently playing audio data in the right format in XNA yet. Instead of trying to detect the beat though the other elements, already have that work done and just focus on response. Although, I could see how it would be less useful for music that you haven't written yourself. |
|
|
|
![]() |
|
|||||||
| Thread Tools | Search this Thread |
| Display Modes | |
|
|