Knuckleheaded Alpha Channel Question of the Week

OK...I rarely monkey around in Motion, but I have amazingly just created about 10 different animated versions of my client's logo for use in various films (cut in SD in FCP). I need to be able to library these rascals so I can just drop them into FCP when I need them.
However, and here is where the knucklehead deal enters the picture.
I am baffled as to how to export the logos out of Motion so there is no black background! I want to create a QT file of each version that I can library and then import into FCP. I have already checked the all of the Alpha channel questions dealing with this issue and the answers are like me and algebra in the 7th grade...huh?
Can somebody give me the kindergarten version of how to get rid of the black background behind my little purple animated logo.
Thanks in advance and I promise to close Motion and step away from the keyboard as soon as one of you mavens helps me solve my little non power user dilemma.

Easiest would be not to export at all, but save a different project file for each version - then just bring the Motion project file into FCP. As long as you have set the background to be transparent (Edit > Project Properties), then the project file will have transparency.
If you want to export, you need to choose a codec that supports an alpha channel, such as Animation or PNG.

Similar Messages

Trying to run code from this weeks question of the week

i've copied the code from this weeks question of the week "sending an http req from a normal class." when i compile it everything looks fine. i've installed tomcat and it works. so i guess i am confused as to where i put the servlet class file??? and do i need to change a web.xml file?
any help would be appreciated,
erik

for simplicity i just put Hello.class and HelloServlet.class in C:\tomcat\webapps\ROOT\WEB-INF\classes. i still cannot get this to work. here are my source files:
Hello.java:
import java.net.*;
import java.util.*;
import java.io.*;
public class Hello {
public static void main(String args[]) {
ObjectInputStream is;
URL url;
String uri =
"http://localhost:8080/HelloServlet";
HashMap hash = new HashMap();
try {
//calling the servlet by passing params
url = new URL(uri + "?&name=MyName&age=25");
// open input stream and read the hashmap
// returned by the servlet
is = new ObjectInputStream(url.openStream());
hash = (HashMap) is.readObject();
// print it out
System.out.println(hash);
} catch (Exception e) {
e.printStackTrace(System.err);
HelloServlet.java:
import java.util.HashMap;
import java.io.*;
import javax.servlet.*;
import javax.servlet.http.*;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
public class HelloServlet extends HttpServlet {
public void service(HttpServletRequest request,
HttpServletResponse response)
throws ServletException, IOException {
     // getting the parameters
String name = request.getParameter("name");
String age = request.getParameter("age");
     //putting them in a hashmap
HashMap hm = new HashMap();
hm.put("name", name);
hm.put("age", age);
     //returning them
try {
ObjectOutputStream p = new
     ObjectOutputStream(response.getOutputStream());
p.writeObject(hm);
p.flush();
p.close();
}catch (Exception e) {
e.printStackTrace();
can anyone see a problem. this code is supposed to be run from the command line. and i am doing that.
thanks for the help,
erik

Alpha Channel Question

Hello,
I have an animation with a white drop shadow on it that I animated in Aftereffects. I export the file from Aftereffects and no matter what I try the white drop shadow shows up as a dark shadow every time in Premiere as if it's matted with black. So I'm wondering if there is a setting I'm missing or something I need to do differently when importing to Premiere, or even when exporting from aftereffects, although I've exporting it several different way so far to no avail.
So here is what I have,
I'm using Adobe CS4 for AE and Premiere, on Windows 7
When I export from AE I'm using the color setting for unmatted (and I've tried it matted with black too just for fun, clearly that didn't work out for me:)
I've tried it as an avi and as a mov file too.
Sooo, in case your answer is for me to use Dynamic Link, I can't get that to work either. And I'm sure that is just user error since I've never used it before. I can get the aftereffects sequence to show up in my project files in Premiere, I can even get the animation to show up and play in the preview monitor, but when I put it on my Premiere sequence over the background it doesn't show up, just seems like a blank layer.
Any ideas?
Nina

Thanks for your replies! So I'm not very good at explaining myself. I did export (and by export i mean, make movie by adding it to the render queue) the file as both of you are recommending. As a movie file, with an alpha channel, all one file. The reason I was trying to do it separate next was because it wasn't displaying properly that way.
Finally, what I did as a work around, is just export the background I was using from premiere, import it into aftereffects, Put everything together in aftereffects and then render the finished file without using an alpha at all. In pic #1 so you can see how premiere cs4 won't properly display an animation with a white shadow behind it if it's imported with an alpha channel and placed on top of a background. #2 shows how it was supposed to look.. how it looked once I put everything together in Aftereffects instead. I just thought there should be a way to fix my problem without the extra step of exporting the background. But I don't think there is.
However,
THANK YOU for your replies and trys!!
(Oh yeah I should also mention that I did try, PSD sequence, Quicktime rgb+alpha, Lossless with alpha etc.. none of which seemed to work for what I was trying to do.)

Photoshop file to Shake -- alpha channel question

I've been staring at my screen for an hour wondering about this, even though I know it must be simple, so any advice is greatly appreciated:
I have a short video clip of a syringe with no needle. I want to add a VFX "needle" as a composite in Shake. I drew the needle as a .tiff file in Photoshop. I added an alpha channel in Photoshop, so I have R,G,B,A channels in my .tiff file drawing. The alpha channel has a white shape that corresponds to the needle drawing and black for the background. I was thinking that my alpha channel will mask out everything except the Photoshop "needle."
I bring the Photoshop needle file into Shake and set up my Shake script as follows: "Needle" into the left hand input of an Over node and Syringe video into the right-hand input of the Over node.
I was expecting to get just the needle "floating" on top of the syringe video clip, and I would position it to connect up with the syringe. What I got was the "needle" with its square white background floating on top of the syringe clip. My alpha channel from Photoshop did not mask out the white background.
But when I load the needle into the viewer by itself and put my cursor over it and hit the "A" key, the alpha channel is plainly there -- a white area in the shape of the needle and black everywhere else.
Also, in the Node View window, when I put the cursor over the needle and hit the "A" key, the thumbnail clearly shows the alpha channel, a white area in the shape of a needle and black everywhere else.
But when I load the Over node into the viewer, the needle is surrounded by a white background. Does it work to build your mask in Photoshop and bring it into Shake? Is there a parameter in Shake that you have to set to make the Photoshop alpha channel active?
Jim

Thanks much, Zebulun. Toggling premultiply worked. In a side experiment, I later saved my Photoshop image as a .psd file rather than .tiff and when I brought that one into Shake, it came in with its clipping-path/mask working without using premultiply in the Over node. Not sure what the difference would be in how Shake treats the two file formats, but they both achieve the same end result.

This is the noob question of the week regarding gdm

I am no linux noob per se. but i did something really really stupid just now.
started with a fresh install of arch, added e17 and gdm.
wait...did i forget something? but of course! create a non-root user. so anywho. now gdm is installed and comes up automatically on boot.
the problem: i have only a root user as of right now and gdm does not allow root logins by default. pressing ALT+CTRL+BACKSPACE quits gdm alright, but it just keeps restarting and restarting. doing killall gdm also fails to do the trick.
so pretty much... i can't log in. of course i need to create a nonroot user, but first i must figure out how to kill gdm so i have room to work.
i am thinking running a livecd (gentoo maybe?) and then chroot and edit the rc.conf so gdm does not come on automatically. then reboot into arch, create user, edit rc.conf to start gdm once again. reboot once again and login with user, but i am sure there is a less drastic measure out there somewhere i can take.
PS: i installed madwifi-ng from svn just fine, but i cant seem to find wireless-tools (at least pacman doesn't install it)... and i kinda need them, how do i go about setting them up?
and why am i using gdm when i can be using entrance? hell. you got me there. i really dont know.

munkymonk wrote:the problem: i have only a root user as of right now and gdm does not allow root logins by default.
You can set gdm from the main screen to accept a root login by going to the configure options. The exact heading it is under will depend upon the gdm style you are using (usually under the "Actions" category). It will require your root pass to access but after that just goto the appropriate tab and tick where needed. Also, you can Alt + Ctrl + F1 to exit X and get to a command prompt where you can set up your user as desired.

My dumb question of the week

I recorded audio tracks and processed them using Bias Peak. However, when I add two of the tracks to garageband and play them. The sound is lower and deeper than what I recorded them at. In affect I have a much deeper voice and it is a slower speech pattern. The audio track (s) there are two sound perfect when played back in Quicktime. Other tracks that I imported to GB work just fine.
This one has me stumped - any ideas???

From HangTime's FAQ:
Set iTunes’ Importing preference to AIFF, 44.1k, 16-bit. With the name selected in the iTunes window choose Convert to AIFF from under the Advanced menu.
When the conversion is done a new file will be created and listed in the iTunes window (Convert doesn’t actually convert, it recreates in the new format). Select this new iTunes song and chose Show Song File from under the File menu. Drag this file into your GarageBand Timeline and it will play fine.

Dumb question of the week...

how do i remove songs from my i-pod? i've looked at the manual and it doesn't make any sense to me...

There are several methods of deleting songs from the iPod and/or iTunes, and the method you choose depends on the settings you have selected for your iPod.
If you have the iPod set to automatically transfer songs, and you no longer want a copy of those songs on either the iPod or in iTunes, simply delete the songs in iTunes. Highlight the songs you no longer want then right click and select "clear", or from the edit menu select "clear". Next time you connect the iPod, the songs will be deleted from there also.
Only do this if you no longer want the songs on both the iPod and your computer.
You can also use the "only update checked songs" method in the iPod options (right click on the iPod in the source list in iTunes and select options, then select "only update checked songs"). With this preference checked, you can remove the check mark from the songs in iTunes that you no longer want on the iPod. The next time you connect the iPod, those unchecked songs will be removed from the iPod whilst still being retained in iTunes. You can check and uncheck songs as you change your mind about what songs you want to put on the iPod. Songs that have no check mark against them will be skipped over when played back in iTunes.
If you have the iPod set to manually manage your songs, and you want to delete the song from the iPod only, then select the iPod in the source list of iTunes, or click on the "music" option underneath it to show all your songs. Highlight the song(s) you no longer want on the iPod, right click and select "clear", or use the edit menu as previously mentioned.
This method only deletes the songs from the iPod, so if you no longer want the song in iTunes also, you'll need to delete it from there separately using the method above.
If you are using iTunes version 7 or later, all these settings can be found in the summary screen.
If you want to delete everything from the iPod and completely erase the hard drive and start from scratch, then you can restore the iPod to factory settings.
Only do this if you have the songs in your iTunes library, or a back up somewhere other than the iPod itself.

Stupid Question of the Week.

Ok, I hesitate to ask this, but for the sake of efficiency (not wanting to peruse the manual at this time).... here goes...
I want to change the density, and/or color balance, (or whatever) on 100 images simultaneously.... is there a method of doing this in Aperture. I don't want to click (stamp) all 100 images.
Thanks for any replies.
Dual 2.7 GHz 5.5 Gig RAM / Mac Pro 2.66 - 4 Gig RAM Mac OS X (10.4.6)

Select all the images you want to edit and use the Stamp menu item (Command-Shift-V) to stamp them all at the same time with whatever items are currently checked in the Lift and Stamp HUD (the HUD doesn't have to be visible to use this method), or use the "Stamp Selected Images" in the Lift and Stamp HUD to stamp all the selected images.
I have to say the Lift and Stamp tool is an AMAZING time saver! Setting white balance for a thousand images? Click, done. Simple.

How can the alpha channel be removed in an image with the command sips?

I solved it by converting the image to jpeg with best quality, but it's a less than optimal solution.
Is there a better way?
I want to know because I'm developing a bash script to manipulate images.
Thank you.

I'm developing a Bash script to automate PNG compression.
Converting true color images is often a good idea to reduce file size dramatically but transparency doesn't look any good then. So removing the alpha channel is usually the thing to do.
I think Apple's platform is very easy to use but as soon as I dig deeper into the system it often becomes frustrating. Why isn't there a simple way to do something so simple and basic as removing alpha channels from images from the command line? Sips should be the answer but doesn't seem to be, unless I've missed something, but the manual doesn't give any clue.

How copy an alpha channel mask to an adjustment layer?

I can copy an alpha channel mask to an image layer by Ctrl+Clicking the alpha channel then clicking the Add Layer Mask icon in Photoshop CC but this does not work for an adjustment layer. How can I copy the mask from an alpha channel to the layer mask for an adjustment layer? Thanks.

I am using Photoshop CC as I stated in my question. All updates have been installed.
I am running on Win 7 Pro 64 bit with all updates installed.
I am editing a photograph.
I have an alpha channel with a grey scale mask.
If I Ctrl+Click the alpha channel then click the Add Layer Mask button to add a layer mask to an image layer the alpha channel mask is added to the image layer.
If I delete the layer mask from an adjustment layer and follow the same steps a solid white mas is added to the adjustment layer.
Several others in the camera club I belong to have tried this with the same result so the problem is not unique to my system.
Hope this helps.

Automation: Howto swap two areas in one Alpha Channel

Dear Forum members,
another question. I have an Image with some alpha channels. In my automation plugin i need to exchange two areas in one alpha channel. At the moment i'm trying to cut one area, move the other area to the place of the first and then paste the first area at the location of the second one. Is this the optimum way of doing? Or is there something smarter / faster ?
Regards, Christof

For the moment (because i'm too lazy to extend my hidden plugin right now) i've realized it from the automation plugin without any additional channels: Cut area 1, move area 2 to area 1, paste area 1 at the position of area 2. This should also be pretty fast, because Photoshop is doing all the work for me.

Photoshop not presenting alpha channel as expected (like I'm seeing GIMP does)

Greetings,
I have a need to edit the channels independently (the values in my image mean things numerically rather than just what it looks like visually - so I need to edit the grayscale values in each of the R - G - B - and A channels seperately).
I've searched through google and other forum posts and others bring up the concept of not understanding alpha versus transparency - but I have yet to see an explanation that fully explains what I'm experiencing.
The closest was http://forums.adobe.com/message/2563436#2563436 But while I saw the exact same symptoms being described, the specific question was never really answered.
Hopefully, a concrete example with pictures will be able to communicate the issue and solicit a useful explanation and/or solution.
While, I'm specifically working with an RGBA image, below is a screenshot of a PNG that demonstrates the exact same behavior or symptom. The screenshot below represents the PNG inside GIMP and PS CS4. BTW, if you'd like to see or play with the image yourself, you can grab the same image at http://www.libpng.org/pub/png/img_png/imgcomp-440x330.png
GIMP, interestingly enough loads the PNG as expected - I see the transparency of the image treated as a true alpha channel.
In PS CS4 however, the transparency seems to be somehow embedded in the Red - Green - and Blue channels; but NO seperate alpha channel.
I can create the alpha channel through Select - Load Selection - Layer 0 Transparency - and pasting that into a newly created alpha channel - but the color channels still have the transparency mixture symptom - rather than showing the JUST the color component of that channel. (Note the channel differences of the GIMP's Red Channel versus PS's Red Channel - Said in another way, the yellow ball in the upper left of the image should have a red value of 255, so the Red Channels grayscale upper left ball should be white like it is in GIMP, rather than Gray as it is in PS).
I never thought I'd experience GIMP being superior to Photoshop - so hopefully somebody can shed light on what is going on - what configuration I may have myself in - or a process to get the channels seperated out in the manner I wish to work with them. Furthermore, maybe somebody can educate the Photoshop community what the difference between transparency and alpha[transparency] is. I haven't found quite the right explanation that makes the light go off in my head yet. As best as I have gleened "Alpha", which can be applied to any kind of channel, typically is in reference to transparency - and is applied as a document-whole transparency (as opposed to transparencies applied at a layer level - such as a layer mask).

I think I've finally determined that modern Photoshops have gone in a different direction that I feel "Alpha" should be done.
(good or bad).
Here are the steps and/or details of how I arrived at that conclusion.
If nothing else this has been very educational on how Photoshop treats alpha.
- Import your alpha'd PNG with SuperPNG format (see above on how to make sure you use the SuperPNG plugin to load the .png).
- In the SuperPNG dialog select "Alpha appears as seperate channel"
- Create a layer from the Background and Delete the original Background
(since the Background is locked and we can't edit the Background - we want to add a layer mask based on the alpha)
- Ctrl-Left Select the "Alpha 1" channel - Ctrl-C - and hit the "Add Layer Mask" button (square with the circle at the bottom of the LAYERS rollout)
- For this demo and clarity, I went ahead and deleted the alpha "Alpha 1" that came in through the SuperPNG plugin
So now I have the Layers like I kindof expected the main document to just come in as (kindof though - it is presented in a Layer Mask rather than Alpha Channel).
What is interesting - which you'll see if you toggle the layer mask on and off
(shift-click the mask thumbnail in the layer)
Note how the thumbnails in the channel views render differently during the toggle.
When the mask is off you see the individual color channels like I expect - black where there is an absence of that color.
However when you enable the layer mask - the individual color channels show in semi-transparent/checkerboard.
So this is exactly how the color channels are presented when loading in a .png (WITHOUT the SuperPNG plugin) or an .rgba
..you get the semitransparent views of the color channels - BUT the difference is - you have no way to directly see or edit the Alpha channel.
nor are the color channels represented as completely opaque versions.
What is really interesting - it is there though - if you Ctrl-Left Select the layer, Ctrl-C, and "Add Channel Mask" - you'll get the Alpha mask.
HOWEVER, the big issue is that you can't toggle or edit the Color-Channels like the Layer counterpart above;
they are ALWAYS the semitransparent/checkboard version; AND now we basically are double applying transparency
(due to the still embedded alpha/transparency AND the newly create "Alpha 1" channel)
I think this a bug (or a poor reimplementation of modern Photoshop releases; I believe older Photoshops used to break out Alpha in the "arguably correct" way - and in fact the way GIMP currently does).
Or maybe there are other explanations - like abstracting some of this complexity away from less advance users, etc?
So the SuperPNG plugin was useful in experimenting, understanding, and demonstrating what is going on.
However, if I'm trying to read in .rgba directly, I don't think I have a way of making the "Alpha appear as a seperate channel" as I think any image with "alpha" component should.
(png, tiff, rgba, etc).

Keying with an alpha channel

I have a client that needs some footage that we shot over green screen delivered to them with the green keyed out and an alpha channel already in place. The end deliverable that they are looking for is the talent over an "invisible" background that they can simply drag and drop into FCP without having to key it themselves.
So my question is, how do I key out the green and an add alpha channel to preserve the "invisibility" of what was once green?
I hope that question makes some sense!
Thanks!

The most straightforward solution is to export your keyed sequences
to Animation, PNG, or JPEG 2000 using millions of colors +.
Make sure the sequence is NOT RENDERED before you export.
(In other words, if you want to render and preview to make sure it's good,
do that, but delete your render file before exporting.) The only problem
is that the resulting files are huge and will really bog down your client's
edit.
Another solution is to export two files per clip, with the second file
being a black-and-white version of the alpha channel.
To create the second clip, add the Channel Swap filter with
red, green, and blue all coming from alpha.
Then your client places the two files over each other, with
the "alpha" clip below the regular version, and the Composite
Mode of the upper clip set to Travel Matte - Luma.
In fact, if your client already has the clips, you can just
send them the alpha versions.
The big advantage of the second solution is that you can
stay in the same codec, which should make your client's
editing go a lot faster.

More CS 5 - ProRes 4444 Alpha channel nonsense

Hi:
Quicktimes I render out of After Effects CS 5 using the ProRes 4444 codec are tagged as having an alpha channel even though my settings in the Output Module are RGB and Trillions of Colors. NOT RGB + Alpha and Trillions of Colors+. It doesn't matter whether my project is 8 bit or 16 bit.
When I bring such movies back into AE it shows them as having an alpha channel (Trillions of Colors+). If I do a Get Info in QT player it shows Millions of Colors+, and FCP shows them as having an alpha, too.
I just upgraded to Snow Leopard 10.6.8 (figured it was safe after two years) and in the process switched full time to CS 5 (I only used it in the past to open other's projects). I never had these issues with CS4. I have not installed any AJA codecs into this new OS (clean install over wiped HD).
This appears to be the opposite of the problem most have been having, which is AE won't generate an alpha in ProRes 4444 without moving codecs and other nonsense.
The reason this is a problem is because when I bring these Quicktimes with faux alphas into FCP for editing I can't export a reference movie. FCP rewrites the whole timeline.
Haven't these CS 5/ProRes 4444 codec/Alpha channel issues been going on for over a year? Do I have to buy CS 5.5 to get this to work correctly?
Thanks.
Shawn Marshall
Marshall Arts Motion Graphics

PR422HQ is way overkill for 99% of the video and film producers in the world. It's there for a specific elite of the entertainment biz, those who work in 4k on 64bit systems, I guess, that' ain't me! It's one of those "if you have to ask, you can't use it" sort of things. Anyone who uses the advanced features of the ProRes family knows why. That's the theory. In my practical experience over on the FCP forum at Apple, 99% of those using ProResHQ believe they are improving their original footage.
Apple ProRes 4444
The Apple ProRes 4444 codec offers the utmost possible quality for 4:4:4 sources and for workflows involving alpha channels. It includes the following features:
Full-resolution, mastering-quality 4:4:4:4 RGBA color (an online-quality codec for editing and finishing 4:4:4 material, such as that originating from Sony HDCAM SR or digital cinema cameras such as RED ONE, Thomson Viper FilmStream, and Panavision Genesis cameras). The R, G, and B channels are lightly compressed, with an emphasis on being perceptually indistinguishable from the original material.
Lossless alpha channel with real-time playback
High-quality solution for storing and exchanging motion graphics and composites
For 4:4:4 sources, a data rate that is roughly 50 percent higher than the data rate of Apple ProRes 422 (HQ)
Direct encoding of, and decoding to, RGB pixel formats
Support for any resolution, including SD, HD, 2K, 4K, and other resolutions
A Gamma Correction setting in the codec’s advanced compression settings pane, which allows you to disable the 1.8 to 2.2 gamma adjustment that can occur if RGB material at 2.2 gamma is misinterpreted as 1.8. This setting is also available with the Apple ProRes 422 codec.
Apple ProRes 422 (HQ)
The Apple ProRes 422 (HQ) codec offers the utmost possible quality for 4:2:2 or 4:2:0 sources (without an alpha channel) and provides the following:
Target data rate of approximately 220 Mbps (1920 x 1080 at 60i)
Higher quality than Apple ProRes 422
Apple ProRes 422
The Apple ProRes 422 codec provides the following:
Target data rate of approximately 145 Mbps (1920 x 1080 at 60i)
Higher quality than Apple ProRes 422 (LT)
Apple ProRes 422 (LT)
The Apple ProRes 422 (LT) codec provides the following:
Roughly 70 percent of the data rate of Apple ProRes 422 (thus, smaller file sizes than Apple ProRes 422)
Higher quality than Apple ProRes 422 (Proxy)
Apple ProRes 422 (Proxy)
The Apple ProRes 422 (Proxy) codec is intended for use in offline workflows and provides the following:
Roughly 30 percent of the data rate of Apple ProRes 422
High-quality offline editing at the original frame size, frame rate, and aspect ratio
High-quality edit proxy for Final Cut Server
The Apple ProRes family of codecs provides these advantages:
Quality indistinguishable from that of the most pristine sources: Maintains superb quality even after multiple encoding/decoding generations.
Mastering-quality 4:4:4:4 RGBA: Provides a lossless alpha channel with real-time playback (Apple ProRes 4444 only). Mastering-quality 4:4:4 Y′CBCRcolor and 4:2:2 Y′CBCR color are also available.
The quality of uncompressed HD at data and storage rates lower than those of uncompressed SD: Provides real-time editing performance comparable to or better than that of any other HD codecs in Final Cut Pro.
Apple ProRes encoding at any frame size—SD, HD, 2K, 4K, or other: Apple ProRes codecs can also be encoded into nonstandard frame sizes, but nonstandard frame sizes are not supported for real-time playback in Final Cut Pro.
Variable bit rate (VBR) encoding: “Smart” encoding analyzes the image. Efficiency is increased because excess bits are not wasted on simple frames.
10-bit sample depth: Preserves subtle gradients of 10-bit sources (sunsets, graphics, and the like) with no visible banding artifacts. When you import a file using an Apple ProRes codec, you don’t have to first determine whether the file is an 8-bit or 10-bit file. Apple ProRes codecs always preserve the bit depth of your original source files.
I-frame–only (intraframe) encoding: Ensures consistent quality in every frame, with no artifacts from complex motion, and speeds up editing.
Fast encoding and decoding: Delivers high-quality, real-time playback and faster rendering times.
Equipment affordability: Because of low bit rates, you can edit more streams with more real-time effects on slower drives, or have more users accessing the same media over shared storage devices.
Workflow options for any video format that does not have native Final Cut Pro support: The Apple ProRes format provides an effective workflow for projects involving multiple acquisition formats when you want to standardize on a single codec.
Better rendering for native editing: Can be used to render long-GOP MPEG-2 formats (such as HDV and XDCAM HD) to speed up editing and avoid MPEG-2 reencoding artifacts before output.

Alpha channel problem with Pixel Bender blendShaders

I'm using Pixel Blender to try and create a blend shader for a Flex 4.0 app.
It looks like this:
<languageVersion : 1.0;>
kernel PixelReverse
<   namespace : "com.abc.def.filters";
    vendor : "EdAlive";
    version : 1;
>
    input image4 foreground;
    input image4 background;
    output pixel4 dst;
    void
    evaluatePixel()
        pixel4 fgPixel = sampleNearest(foreground, outCoord());
        pixel4 bgPixel = sampleNearest(background, outCoord());
        if((bgPixel.r == 0.0) && (bgPixel.g == 0.0) && (bgPixel.b == 0.0) && (bgPixel.a == 0.0)){
            bgPixel.r = bgPixel.g = bgPixel.b = 1.0;
        dst.r = 1.0 - abs(fgPixel.r - bgPixel.r);
        dst.g = 1.0 - abs(fgPixel.g - bgPixel.g);
        dst.b = 1.0 - abs(fgPixel.b - bgPixel.b);
        dst.a = fgPixel.a;
It looks correct in the Pixel Bender preview window.
When I use it in a Flex project with proper 32 bit png images, the filter seems to work correctly, except it sets any partial alpha value on the images to fully opaque (like a 1 bit alpha channel).
Because the image preview looks correct in Pixel Bender, I'm assuming it's a problem with the bitmap data that Flex is passing to the blend shader filter.
I tried changing the filter to simply output a copy of the foreground firstly, and the background second, to test.
When returning a copy of the foreground, the imags weren't visible on stage, either because the pixels were transparent or because it was really returning the background pixels, camouflaging the images.
When returning a copy of the background, the images were just black rectangles.
The images are being place on stage programmatically inside an mx Module, and having their depth property set.
They also have draggable behaviour, where the current drag object is layered above all else during the drag.
I'm actually using a subclass of mx.controls.Image, but am simply setting the source property with preloaded bitmap data, so I'm am fairly sure the subclass is not a factor.
I hope this is enough information for someone to be able to provide some help.
Thanks

Please make sure you have installed the latest graphics drivers, from http://www.nvidia.com/Download/index.aspx?lang=en-us, for your graphics card.

Knuckleheaded Alpha Channel Question of the Week

Similar Messages

Maybe you are looking for