Short Scenes: Pygmalion Computer Graphics Project

#541

I agree. I’m saying that too. It would be better.

#542

Tughan: OK. I’ll try it. Tughan if you would be kind enough to post just one more test so we can make this decision and then you can finish your work. You can do the posting on the FTP site if you don’t have space to host it. Give me some PNG files with the background pixels set to transparent, I will make sure we can get what we need.

My problem is not with the files coming in during November and December. I’m afraid some rigging will not be done until January - February and then we might not have time to do things this way.

#543

No problem my friend. Here is a png render with an alpha channel

PNG rocks man! :buttrock:
LoL:)

#544

There is a line of green pixels all the way around the model. I assume this happened because you started with a Green Background frame and that I will not see it in your final rendered frames.

What is the process that decides what is model and what is background? How close can the result consistently be?

#545

wouldn’t all the compositing be easier if you had someone on the team who was facile with a program like Combustion? No, not me, lol. just a suggestion. If you had someone with combustion they could do the composites, a few fx and also do color correction so the entire thing has a consistent look.

#546

Actually this thin green line is caused by the anti-alliasing. It tries to blend edges of the character to background. Well, I don’t know if there is a way to turn off blending foreground objects with background. But when I disable anti-alliasing, that slimy thin green line is gone!
Check it out…

I did some compositing work for my shortfilms. But I didn’t composite over a live action before. It would be cool to get some advise, tips&tricks from a compositor before proceeding.

#547

Make sure your background is set to black, not green. You’re going to get that thin green line/color in the motion blur as well. What I usually do, in Lightwave, is render out my frames as a Targa sequence (32 bit {-24bit for color-8 bits for the alpha}). I then bring in my animation into After Effects, select targa sequence and set my matte to black. Then I can composite my characters over backgrounds etc. I think this method would be the most beneficial as far as speed and time. Keep in mind, these frames will be uncompressed so they will take up some serious HD space. By the way, I’m assuming we are working at 720x486 resolution…?

#548

Here’s a composite test of Liza. I assigned her geometry to a renderable layer in maya and only rendered that layer as a targa. I then imported the file into after effects with a premultiplied alpha channel set to black and added the background. I’m sure most 3d packages have similar options. This seems like the easiest and most efficient way to do the rendering/compositing.

[[IMG]http://img120.imageshack.us/img120/8194/lcomp7mb.jpg[/IMG]](http://imageshack.us)
#549

I also agree with rendering in .tga or .png with an alpha channel.
Color maskin is the film way, and alpha chan is the cg way.
They green line could be removed while compositing too. Even with aa
set up. Just need to tweak it, and the prosess is different in different progs.
In After Effects it’s simple. I think… :wink:

Anyway. Another update. Almost finished modelling now.
I got some questions though:

  1. How should the entrance be? The reference pic doesn’t show this well. Do you
    have any other reference pics of similar houses with a typical entrance?
  2. Should I model the fence at the ground?
  3. What about the backside? I’ve left it out for now, but with a small amount of effort
    I can get some windows n stuff on the backside too.

edit: typo #1

#550

Here’s the rest…

edit: typo #2 :stuck_out_tongue:

#551

Very good modelling Pivot. You did an awesome job. I like it very much. Even without textures. I bet it’ll look so cool in final render. Good work. :thumbsup:

#552

I have to concur, Pivot, that model is looking great. Can’t wait to see it with textures.

#553

opreska and others: The problem is the way the Project is staffed. I never give someone an assignment before their materials are ready. For example, I don’t assign a rig before the model is posted because I don’t know when the work will be finished. There is nothing worse than having a motivated volunteer but not having any work for them to do. That translates into I never recruit someone before their materials are ready.

At the end of the Project, the people that will run out of time are the compositors. If the people I need to recruit must have access to a copy of a program with a 4 digit price tag, and they know how to use it, and they have the talent required to do a good job, that limits who I can ask. They are likely to be professionals who don’t normally give away their skill. It sometimes takes a long time to find a volunteer and this would make it harder. There are many more people who can use a video editing program than Combustion or Fusion or Whatever. (I did download a copy of Jahshaka this morning and I will look at it this evening to see if we can use it.) I thought that if we prepared color keyed movie files plus a video background and gave it to someone with a VEP they would do the right thing and it would look good. They know how to work with layers also.

I consider that everyone in the Project has the same desire to finish that I do. I thought that this would get us there with less risk. I am willing to try Targa or PNG files and wait for the right person to come along to finish for us. I know the right people are out there, I don’t know if they will volunteer. I could use your help. If you know someone with the right skill set please ask them if they are available to work on the Project. It will sound better if it comes from you instead of from me.

Please make a comment on this issue. We need to make our decision soon so that Tughan can finish his work. I had the first pre production meeting for the drama part of this Project on Saturday. There were 10 people here, Set designer, Light Designer, Scenic Designer, Stage Manager, etc. I showed them your work. They were excited! They heard me talk about the animation before but this was the first time they had seen anything. I’m sure the audience will feel the same way. We must finish on time. What is the best way to do it?

#554

I still think the Targa (or PNG) method is the best way to go. Almost any editing package worth its weight will be able to read the alpha channel. If we have to key out backgrounds, then that will just add to the time it’ll take to finish. So, the Targa/PNG with alpha’s would just be more efficent.

#555

I went to the Library this morning a found a copy of Recent English Domestic Architecture dated 1912! Evidentally, the English did not photograph their front door unless it was ornate. Here is a sample.
FrontDoor

Amy told us that the iron fence is very typical in London.
Similar Fence
Here is a blow up of your reference photo
Blowup
You don’t need to do anything for the back. It will not be seen.
You are doing a Great job!

opreska: Yes, the final resolution is DVD.

It is interesting that two Project members mentioned in the last two days they have a copy of After Effects. Maybe it is more common than I thought. The problem is I can’t test that theory until all models are in and rigged. We still have 5 weeks to go before we will be in that position.

#556

That is true if the editing package makes it easy to read in a sequence of frame files. After Effects does this, all you have to do is indicate the directory and set the frame rate. If the video editor makes you specify each file name, that would not be efficient. They are set up more for movie/video file input rather than key frames or all frames.

#557

Well got put behind schedule this week. between cars, kids, work , school and Thanksgiving it was hard to get anything done. Here are two pics. Just have to make the shirt and tie.

The Texture is just a quick one will map it better and texture the clothing to look right.

I don’t have much time to finish. ( My Own Fault )

John

#558

So far the vote is 5 to 1 so Tughan can make his frames now. I’m sure we will be alright. I’m not worried about the files that will come in during December. We will be able to get those composited. I’m worried about the files that will come in during February. It may be hard to get a compositor to commit to a short deadline. The best thing we can do is get the work in as soon as possible to keep the time lines reasonable for the people who come after us.

I don’t think that Jahshaka is ready for us to use yet. There is no manual or help files and I only saw low res output formats. I did see that PNG files are acceptable inputs but just as still layers. There is no automatic way to accept a sequence of files.

jcasano: Glad to see another WIP. As soon as you are finished we will pass it on for rigging.

#559

Here’s an update of the limo. I really didn’t get much done due to Thanksgiving weekend.
CC welcome…

#560

opreska: Jeves is looking cool. Remember the photograph I posted for you here? 1912 Rolls RoyceThe body seems longer in the photo.

To make the Limo look different than the taxi we might want to stretch your body a bit. Is it too late to do that?