Swing / SickTheramin

[MegaHurtz]

I wanted to do something with it but im not good enough to find it
Can you make it so that it pops up faster, and dissipates its bouncing energy faster.
say over 6 frames or so..

[MegaHurtz]

Found it

[MichaelBenjamin]

yes this is inside the swing module, could be done outside for better clarity.

[stw]

small update:
added a "show Midispot" feature. No you can see what you'll play. Also nice to see the ball swinging into the hole.
That drawloop thing drove me nuts.

[trogluddite]

That drawloop thing drove me nuts

Just done a small edit to set the draw loop up correctly - the comparison etc. were not needed, the Draw Loop only needs to get a ReDraw trigger from its left hand side for it to create all of the looped parts.

SickTheremin0019b.osm

(243.29 KiB) Downloaded 99 times

P.S. The MIDI 'holes' graphics could ideally do with feeding into the 'VA Cache' of the main 'arena' - that would save a few more GUI CPU cycles - but now i must go to bed!

[mwvdlee]

P.S. The MIDI 'holes' graphics could ideally do with feeding into the 'VA Cache' of the main 'arena' - that would save a few more GUI CPU cycles - but now i must go to bed!

The Arena VA Cache is background, the holes are foreground so if you wanted to cache it, it'd have to go on a foreground layer. Considering the (lack of) size and complexity of the holes though, I doubt it's going to improve performance significantly, but it should be very easy to try and compare

Anyway, I noticed a problem with the tabs. They were redrawing at 100 ticks. Not much of a problem when the tabs were closed, quite more noticable a problem when open. So I've included a check that only redraws if pixel position has changed.

SickTheremin0019c.osm

(243.68 KiB) Downloaded 96 times

[trogluddite]

SickTheremin0019d.osm

(244.14 KiB) Downloaded 148 times

The Arena VA Cache is background...
...very easy to try and compare

Added caching locally to the midi 'holes' - and was rather stunned. Made the animation far smoother - mind you, I've been testing on my little netbook with its weedy Atom CPU and built in GPU.

[MegaHurtz]

Somwhere along the line it got a bit slow. and something sounding like peakducking came in the place of the weew sound when it gets thrown. May be a nice bet to have it bounce along the edges, since these are straight it`s the
opposite direction, of the angle of attack.

[stw]

SickTheremin0019d.osm

The Arena VA Cache is background...
...very easy to try and compare

Added caching locally to the midi 'holes' - and was rather stunned. Made the animation far smoother - mind you, I've been testing on my little netbook with its weedy Atom CPU and built in GPU.

Thanks for correcting that loop redraw trog.
However i think that VA caching should be tested a bit more. I get some strange results with CPU load. With VAcache the task manager shows a decrease of 10-20%. Which seems a bit too much for me? At the same time SM (Audio-)CPU meter shows an increase from 7.5% to 8-12%.

[trogluddite]

strange results with CPU load

Hmm, I wonder how much is down to machine architecture then.
On my little Atom netbook, the 'Audio' CPU meter shows no difference between cached and uncached versions (about 25%). Task manager is at about 65% for the cached version (with nice smooth animation) and the uncached version shows about 75% - but uncached gives one CPU core maxed out, and the animation is very lumpy (I'd estimate 5-10 frames per sec, with occasional half-second freezes).
Of course that's a little unfair - I doubt many people would be running VSTs on an Atom CPU - and having one core so near to max is probably skewing the results as it is close to the 'tipping point' of 'no-can-do'.

Having said that, it is surprising how much you can gain using a bitmap cache - even redrawing tiny areas can cause a lot of GUI/Green activity when multiple layers/objects overlap, as it's not just the layer connected to the ReDraw that gets redrawn (see this thread). I have a hunch that it makes draw loops recalculate every iteration - how else would it know which iteration overlapped with any given draw area?

[stw]

minor update: Implemented trogluddites "Trigger on match" to trigger the bumpers.It's now way more accurate and audio and visuals are synced now.

[trogluddite]

Here's some more tiny tweaks...
I realised that the ball was looking just like a sunburst gradient - so replaced the draw loop (yuck) with a gradient. Also tidied the area calculations up a bit to avoid too many conversions from Area-Float-Area - and added switching so that the area covered by the 'string' is not redraw if it is turned off.

SickTheremin0019f.osm

(245.41 KiB) Downloaded 196 times

[mwvdlee]

I realised that the ball was looking just like a sunburst gradient - so replaced the draw loop (yuck) with a gradient.

That's just one of those things that are so blatently obvious that nobody even noticed it. Thanks for spotting it
I do prefer it a bit softer, as it was meant to be a dropshadow, but that's a matter of taste.

The ball no longer seems to move though

[trogluddite]

Oops - saved it with the ball 'untethered' - you just need to R-Click. (maybe unselect/select ASIO too in case saving with my driver has confused it!)

[infuzion]

Cool guys!

Do you have a time table to be done sometime...?