milkdrop.us

[redi jedi]

well, i've ordered my new website, it gonna be milkdrop.us now we would all have to hit every item on the site like 30-50 times a day for me to go over bandwith... yay!
its not up yet, says one busness day, so it might not be up till monday.. but i'll have the beta6.1 source up when the site is.

[51Strong]

Ah great you got your site back :)

[Logy]

yeaaahh... let we make things better... step leap ahead... !!!

[redi jedi]

ok its up, beta 6.1 .dll and source are there, check the dev page. thanks to everyone for understanding, and supporting me during this time

[MechxWarrior]

Great to hear it :)

Cant wait for a fix to the multi-monitors.

[Phat]

I updated the wikkipedia artical here..
http://en.wikipedia.org/wiki/MilkDrop#External_links

Also you're guest book seems to be broken.

[redi jedi]

ya i have to actualy add the database for it through my hosts control panal

[Koopa]

redi jedi: will your betas get the fixes from MilkDrop 1.04d (multiuser compatibility, fixed high CPU usage), or is there no source code for 1.04d aviable?

[redi jedi]

1.04d???
i was under the impresion that the last release was 1.04b... i've personaly seen 1.04c, and now i'm hearing of 1.04d... if there is source i havent found it yet. if you find it i would be happy to include it

[Koopa]

I'll ask, if the sources are aviable, the changes since 1.04b:

* Fixed: [vis_milk] 100% cpu usage when paused
* Fixed: [vis_milk] missing files from distro (not releated to your development)
* Fixed: Visualizations for 24bit and 32bit songs
* Improved: Visualization data calculations (not sure, if this is MilkDrop related)
* Fixed: Multiuser compatibility

It makes no sense, if you develop 1.05 and Milkdrop 1.04d has bug fixes and improvements, wich are missing in 1.05. ;)

Keep on your good work.

[Benski]

The following changes were made...
*Grabbing the user's INI directory via a Winamp IPC call.
*Using Winamp's mersenne twister random number generator (you can grab the function pointer via an IPC call)
*Setting winampVisModule::delayMS to 10

at some point in the future, I will be removing the FFT calculation from milkdrop and start using the spectrum data directly. I have improved Winamp's spectrum viz data calculation dramatically (it's safe to say that it used to be completely broken)

you can contact me directly if you need any more details for source changes.

[redi jedi]

a copy of the source would be nice, or at least the code for the updates... i didnt know that vis plugins had direct access to frequency info... makes me wonder what the fft was for in the first place.. i'll research how to acces it and prolly add that feature in myself( i try to leave the current process intact so old presets still work, i'll prolly change my sound(a,b) function to get its info from winamp directly... sweet!

i'm assuming the delay 10ms is what fixed the 100% usage?
i would prefer to not have to redo work thats been done, so a copy of "current" source would be nice... why was only the source for 1.04b released when theres 2 more versions?

[Benski]

i didnt know that vis plugins had direct access to frequency info... makes me wonder what the fft was for in the first place..
Because Ryan felt that Winamp's FFT wasn't good enough (and he was right).
i'll research how to acces it and prolly add that feature in myself( i try to leave the current process intact so old presets still work, i'll prolly change my sound(a,b) function to get its info from winamp directly... sweet!
set winampVisModule::spectrumNch equal to how many channels you want (1 or 2) and Winamp will start giving you data via winampVisModule::spectrumData

i'm assuming the delay 10ms is what fixed the 100% usage?
Yes.

i would prefer to not have to redo work thats been done, so a copy of "current" source would be nice... why was only the source for 1.04b released when theres 2 more versions?
Because these changes were made after the source was released.

[redi jedi]

ahhh! i guess i just never heard about them because my head has been burried in ryans code for the past 1/2 year or however long its been out.

did you guys do these updates or someone else? if it was you is the source thats posted now updated? other wise where can i find it? if posible please e-mail it to
redi-jedi@milkdrop.us and i would be happy to include it in my next beta...


as for the spect data, this is now top of my list(it was pretty empty before though :0) since it was only there to fix an inequity in winamp witch is now fixed, i'll go ahead and remove the old fft since it, that should up fps at least a little...

[RonnyJ]

Not sure whether this is the right place to post this, but since both Benski/redi jedi seem to be involved in Milkdrop development, I hope it's relevant here.

There's an issue with Milkdrop and DEP (Data Execution Protection) on AMD 64 chips, which results in an illegal operation.

An example here:
http://forums.winamp.com/showthread.php?s=&threadid=237684
and
http://forums.winamp.com/showthread.php?s=&threadid=200025

I've been using Milkdrop 1.04, which has this problem, and so I've had to manually disable DEP for Winamp, after finding the solution buried in a thread posted a year ago. I've just re-enabled it to see if it's fixed with redi jedi's beta 6.1 build, but it still crashes.

[redi jedi]

sounds like a windows problem to me... i guess it could be added to the readme or something...

[mr_jrt]

From my experiments with the code I noticed Milkdrop does some kind of dynamic code generation. I suspected at the time it would cause problems with DEP, but as I only have an old P3, it didn't really bother me. I'm not sure what the correct way to solve the problem is (i.e. tell windows to turn off DEP for some data progmatically).

[bpetersen]

Originally posted by Benski
The following changes were made...
[...]
*Using Winamp's mersenne twister random number generator (you can grab the function pointer via an IPC call) [...]


Can you please tell me how this works?

Best regards,
Bjoern

[Phat]

Honestly a 512 band FFT sux. Unless you have like a bizzilan points, it has SOOO much noise it's useless. Even Sound Forge's like 6000 band FFT is still noisey in the sub bass frequencys.

Eo.S. and I, while making a music video KNOW what frequency bands we wish to respond in certain ways. STILL we cannot get milkdrop to respond to certain bands sufficiently enough to be visually accurate when in the context of ONE track.

And thats when we have the audio we want a preset to respond to directly, and can look at the spectral analysis in audio software and pick out what frequency bands certain elements of the track spike.

Also, would changing milkdrop in this manner make it so that people using older versions of winamp would have a broken copy of milkdrop?

[redi jedi]

i was thinking of checking for the version of winamp, and doing the old style if its not after the "fixed" version... still not sure how to do that though... been really busy latly, and havent had much time to devote to MD programing, or research, but i'll get to it sometime, any ideas on how specficly it should would would be helpfull

[Sherwin Maxawow]

COOL Redi....I like what I see there. Also like the presets. Hope you appreciate sincere my compliment.

[redi jedi]

thanks Sherwin! it really does mean alot to me to see you guys appreaciate what i'm doing here(although i have done anything in a while, got work and i'm trying to build 4 websites at onece for diffrent people)