I'm still curious about a quick way to rebuild 4 1D arrays into a vec4, but it occurred to me that there is an OpenCL sample that prepares vec4 color structure from image in the OpenCL text extrusion example (though the extrusion aspect isn't really needed for what I was trying to do). I'm getting 40~60 fps rendering around 3000 points, with colors pulled from video. Not horrible.
A side effect I've noticed in multiple instances of using the point sprite mode on the mesh creator, is it seems to reveal some kind of "wrongness" in the spacing of point sprites. They always have less space/butt up against one another in the "middle" of the grid, and have more spacing at outer edges. Is this because of the QC projection scheme? Yesterday I tried to render it in Ortho, and it didn't make a difference. I would definitely appreciate feedback on this - I think it effects the mesh renderer even when not using Mesh Creator or point sprite, but it's harder for me to tell. In the composition, at the "render size" I'm viewing it at, I can see point sprites bunched up and "touching" in the middle, and not touching more and more from center to boundary. I made a sprite/iterator setup, and in that macro, sprites overlap (or don't overlap) with consistency. -GT On Mon, Oct 4, 2010 at 3:19 PM, George Toledo <[email protected]> wrote: > I did attach the iterator/js logic in my last post, but maybe it would help > to see something more concrete on how I aim to render this. > > The first example uses the 3rd party Image PixelS ( > http://www.magdatt.nl/software.html) in order to get R/G/B/A components of > an image. > > The second example just shows the result of "firing off" different color > values to this type of rendering engine. > > It boils down to "what is the most efficient way to prepare large float4 > structures for the Mesh Creator" for this particular instance. > > -GT > > > ---------- Forwarded message ---------- > From: George Toledo <[email protected]> > Date: Mon, Oct 4, 2010 at 2:59 PM > Subject: structures in QC programming question > To: quartzcomposer-dev list list <[email protected]> > > > What is the most efficient way/how does one do the following in QC, in a > way more efficient than my attachment? > > How do I configure a javascript or openCL patch to take in 4 structures of > floats, and output a float4 structure, with each index of the incoming > structures constituting the 4 values that are in each output float4? > > So, in essence, I want to take 4 structures that represent the R/G/B/A > values of an image, run them through a javascript program or openCL kernel, > and output a structure of float4's that would be something like: > > 0:("R,G,B,A") > 1:("R,G,B,A") > > ...and so on *and* have elements come back in correct array order. > > I'm attaching a javascript+iterator based example of this working *and > *structure > returning in a "non-mixed up" structure order. Also, in QC is it not > possible for each element to maintain proper placement inside the vec4 and > have a key with it ( I know this has been covered somewhat, but I can't find > relevant discussion)? I didn't seem to be able to use R/G/B/A labels and > keep order in my testing, even if I declared it an array. > > What would be even better is to achieve something like Image Pixel (or the > 3rd party, PixelS) that automatically delivers a float4 of RGBA value for > each pixel. That would be ideal. As is, I'm trying convert the data of > either into this structure type. > > -- > George Toledo > [email protected] > www.georgetoledo.com > > The information contained in this E-mail and any attachments may be > confidential. > If you have received this E-mail in error, please notify us immediately by > telephone or return E-mail. > You should not use or disclose the contents of this E-mail or any of the > attachments for any purpose or to any persons. > > > > > -- > George Toledo > [email protected] > www.georgetoledo.com > > The information contained in this E-mail and any attachments may be > confidential. > If you have received this E-mail in error, please notify us immediately by > telephone or return E-mail. > You should not use or disclose the contents of this E-mail or any of the > attachments for any purpose or to any persons. > > -- George Toledo [email protected] www.georgetoledo.com The information contained in this E-mail and any attachments may be confidential. If you have received this E-mail in error, please notify us immediately by telephone or return E-mail. You should not use or disclose the contents of this E-mail or any of the attachments for any purpose or to any persons.
Color Table to Point Sprite Mesh_gt.qtz
Description: Binary data
_______________________________________________ Do not post admin requests to the list. They will be ignored. Quartzcomposer-dev mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/quartzcomposer-dev/archive%40mail-archive.com This email sent to [email protected]
