Sorry for the flurry of posts: I have a semi workaround. After reorganizing it so that all patches are in a lighting environment: The sprite that doesn't move back and forth in Z still doesn't. The sprite that is directly inside of the Lighting patch *does, * which gives hope to lighting a Sprite AND a dynamically created mesh at the same time.
However, when one flicks on the shadow engine, both *entirely disappear.* Wah wah wah.... -gt On Fri, Oct 1, 2010 at 1:36 PM, George Toledo <[email protected]> wrote: > Apparently, width and height, x, y translation, and x/y/z rotation all keep > working correctly. Only z translation stops working for some mysterious > reason. > > It would be helpful to know if the sprite inside the lighting patch stays > at a "static z" for others as well, while the other that doesn't share a > macro with the mesh, moves from front to back. In addition, if lighting > works on a "drawn mesh" for others, like it works for me, it would be > helpful to know as well. > > Thanks for off-list responses that I've received from a couple people. > > -GT > > > On Fri, Oct 1, 2010 at 1:30 PM, George Toledo <[email protected]> wrote: > >> For those keeping track (this has been a fun one sided conversation): >> >> I'm attaching an example of this with working normal generation. When I >> got this working I was very excited. I was totally wrong about my queue >> setup/kernel. I was correct in my original thought that I should feed the >> entire vertices structure in order to calculate normals. >> >> So, this works, and looks like I would expect, but now there is a very odd >> side effect. I have a sprite placed in the same environment, and it no >> longer moves when powered by interpolation! This happens even without >> shadows being enabled. The sprite should be moving back and forth on the z >> axis, but it stays as though it is placed at 0. >> >> With some more testing, it becomes apparent that the Sprite can't share >> the same macro; it cannot even be placed by the Lighting environment that >> the mesh is in, it actually has to be placed outside in order for it to >> evaluate. >> >> I'm guessing this displays a new bug with evaluation? I'm happy to have >> normal generation from arbitrary vertices working, but not so happy that >> other stuff stops evaluating correctly. >> >> >> -GT >> >> >> On Fri, Oct 1, 2010 at 11:12 AM, George Toledo <[email protected]>wrote: >> >>> >>> >>> Ok, I've looked at this further, and I'm submitting something that I >>> truly believe should work. >>> >>> The composition I posted earlier was taking an entire structure of >>> vertices, and trying to convert to normals. That wouldn't work, and was >>> ill-concieved on my part. >>> >>> The first example in the folder takes each vertex, computes a normal >>> using an OpenCL kernel from the Aurora.qtz example that is "supposed to" >>> generate normals. I construct a queue of normal values, and feeds them to >>> the mesh creator, alongside the vertices. However, the kernel fails. >>> >>> Then, I'm trying a second attempt, based off of some more of the premises >>> of the Aurora.qtz sample composition, that integrate Indices into the mix - >>> by ignoring z count I may be making an error, but I am unsure. In this, >>> there is no indicated failure, but the chain fails to evaluate. I checked >>> the Aurora.qtz sample code, and notice that both kernels fail to >>> evaluate/generate normals there as well, even with the additional different >>> queueing setup than mine. >>> >>> It was somewhat difficult to find the examples of normal construction >>> since they were placed into Aurora as custom patches. They had to be >>> extracted. I believe that the first example in the folder should definitely >>> work, and I don't know why the kernel fails. I also am uncertain why the >>> whole "normals" chain fails to evaluate in Aurora.qtz. >>> >>> >>> -George Toledo >>> >>> >>> On Fri, Oct 1, 2010 at 9:25 AM, George Toledo <[email protected]>wrote: >>> >>>> I'm not trying to be strident by asking three times in a row, but how >>>> can I get a head's up about how to properly make normals, without damaging >>>> my computer in the process? >>>> >>>> To quote the QC4 release notes: >>>> >>>> "Quartz Composer now has a fully integrated geometry pipeline. This >>>> pipeline centers around the opaque QCMesh type, which contains information >>>> such as vertex locations, vertex colors, normals, texture coordinates and >>>> indices. The geometry pipeline integrates seamlessly with the rest of >>>> Quartz >>>> Composer. For example, geometry can be created and modified directly on the >>>> GPU using OpenCL." >>>> >>>> Can I please get instructions, or pointed to Apple sample code, about >>>> how to make normals for the "fully integrated geometry pipeline" that >>>> "integrates seamlessly with the rest of Quartz Composer"? >>>> >>>> When things go awry with OpenCL, it can cause crashing, locking of the >>>> system, black strobing on the display, and destruction of data or systems. >>>> If that wasn't the case, I would experiment until I found the answer >>>> myself, >>>> but that being the case, I think it's probably reasonable of me to expect >>>> an >>>> actual answer on how to do this, or get some kind of affirmation that Apple >>>> will buy me a new laptop when mine catches on fire from using QC! (tongue >>>> in >>>> cheek here...) >>>> >>>> -George Toledo >>>> >>>> >>>> On Thu, Sep 30, 2010 at 8:36 PM, George Toledo <[email protected]>wrote: >>>> >>>>> >>>>> >>>>> This is an example that *doesn't* work, to spectacular effect on my >>>>> system. The way I'm generating normals may not be correct, but I'm at a >>>>> loss >>>>> as to what to do to handle it correctly, or if it is even possible, since >>>>> the mesh creator has other bugs with the normals function. >>>>> >>>>> I've included a working version without normals attempt as well. >>>>> >>>>> -GT >>>>> >>>>> >>>>> ---------- Forwarded message ---------- >>>>> From: George Toledo <[email protected]> >>>>> Date: Thu, Sep 30, 2010 at 10:29 AM >>>>> Subject: generation of normals >>>>> To: quartzcomposer-dev list list <[email protected]> >>>>> >>>>> >>>>> What would be the appropriate way to generate normals for the mesh >>>>> creator, using the "mouse ribbon.qtz" as a reference. How is one supposed >>>>> to >>>>> appropriately handle texturing and indices as well? >>>>> >>>>> I see some examples that use a 2D grid, but these principles don't seem >>>>> to work, or at least I'm confused about how to apply them to a scenario >>>>> like >>>>> "mouse ribbon.qtz". >>>>> >>>>> -- >>>>> George Toledo >>>>> [email protected] >>>>> www.georgetoledo.com >>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> 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. >>>> >>>> >>> >>> >>> -- >>> 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. >> >> > > > -- > 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.
_______________________________________________ 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]
