Scene one test, now with sound!

I've been playing around with some ideas for the first scene soundtrack and the first set of credits. Here's what I've got right now. Let me know what you think!

First scene test #1

... And the long awaited first scene test!

Again, I don't claim in any way that animation is my strong point!

And, here's the blend file!
https://docs.google.com/open?id=0B3AzOhkYNoUUOHFzV3pVek05MDQ

In the meantime...

While it's rendering I'll put some more details on here.

 The file for the first scene is made up of four blender scenes. I have one for the world, one for the satellite, one for the asteroid, and one for the star field. If you're wondering why I can't combine them, and reduce complexity, there are several reasons. One reason is that it uses two render engines and then composites them. Another reason is that it makes it a lot easier to apply the right effects to the right parts in the compositor. For example, I use a glow effect on the world, which I think looks good and really improves the scene, but that glow effect looks really weird on the star background, so I had to find a way to selectively apply compositing effects.

Animation of the satellite and asteroid is really hard (at least for me, that's one of my weaker areas) because they are in separate scenes, using different render engines. This could be solved by converting the asteroid to Cycles, but the reason I haven't is that it doesn't look as good:

That's with the same lighting as the satellite. Doesn't look nice at all, does it?

Anyway, it can be done, but for now I'll let it alone. Video coming soon!

The taming of the Frankenstein.

Well, I think I fixed it. It's rendered 300 frames or so (about half) without crashing, so I may have solved the problem. I restarted my computer, and this time resisted the urge to open my web browser. I opened Blender, and disabled Save Buffer, Instances, Free Unused, and Local Coordinates. (I think, if I remembered right). It also seems to be rendering somewhat faster too. I'm seeing render times of about one minute on my main computer.
 The main problem I see when I look at it is that my animation skills are abysmal.
Josh, would you be willing to try animating it, and then sending the .blend file back for me to render?

Update: It's around frame 550 now! I may be able to get a video out tonight!

I made a Frankenstein!

AARGH!

Well, I'll post a short test of the first scene soon if I can get it to render. It's not working right now, to the point of totally crashing my main workstation, and when I say totally, I mean the mouse stops working, then the keyboard lights start blinking, then the whole computer reboots by itself. I was going to farm this out so we could cut down on rendertime, but I'm not going to inflict anyone with a blend file that crashes every five frames!

Other than that, things are going pretty well. The satellite looks pretty decent, as do the world and asteroid, so if I can get them to render we should have some pretty good looking results. Here's a still render, I don't seem to have as much trouble with those:

-Going into troubleshooting mode... Will update if I get any large uninterrupted render segments.

...And here it is!

Here is the .blend file for scene one: 

http://www.blendswap.com/blends/scenes/asteroid-satellite-and-world/

Don't forget to unpack the textures!

Within a few hours of me posting it, it was on the front page of blendswap! Here's proof:

(O.K., so it was in the recently downloaded, but still... :D)

I will upload it to Google Docs if anyone wants, otherwise I'll just host it at blendswap. Reading through blendswap's terms, I realized that they will have limited usefulness for us, as they will not accept WIP (work in progress) blends, and I expect we'll have quite a few of those! Also, I need to learn how to make my own textures, because it's just so much easier not having to keep track of every single texture used and the license. I can only make so many though, the earth ones obviously have to come from NASA ;) Anyway, enjoy, and please post if you do anything interesting!

Title

This could come in quite handy:
http://www.blenderguru.com/videos/how-to-create-a-professional-logo-animation/

Very professional looking. I've got some fun ideas for our logo, (VA FX), but they'll have to wait, because I'm about to release the long-awaited asteroid, earth and satellite .blend file! Here's a render:

Notice any problems? It's rendered in 4K, and I tried something different with the link, so you should be able to download it and view it in full resolution. (Hint, the problem has to do with the compositing, and it's not the lack of it), I still have to put the rest of the nodes in before I release it though.

Follow focus update

Some more parts for the follow focus came today, and things are starting to fall into place, so here's what I've got right now:

Here are all the parts, just laid out.

Here the lens is on the frame (rod system) in approximately the position it will be when mounted on the camera.

 Here are two phoctos of the gear used. The pitch is 0.08" (MXL)

Detail of the handle mount, press fit. (Also surprisingly strong!)

Close up of the drive belt on the lens focus ring

More of the same

This is how the whole thing will work, I will just slide that white mounting plate in, and mount the stepper motor on it (when it gets here), and mount the gear on the motor. The calipers are just to show size and to hold the gear in place.

Here it is from a different angle.

 That's where things stand right now, and so far it's working nicely. The new stepper motor should come soon. I ordered one from MPJA.com, only to be told that that model was not in stock, and the listing was a mistake. :( Oh well, they offered a different model that I think will work, for the same price, so I'll give that a try, and see how it works. The main thing is that I want it to run on 12 VDC, and have enough torque. The motor runs on 12 V, adequacy of torque awaiting testing.

 I can't really use it with the stepper motor I have right now because (1) I haven't figured out a good way to get enough torque out of it without blowing the driver transistors, and (2) it has a gear permanently affixed to its shaft, (of the wrong size) so I would have to cut it off to get it to work with the gear and drive belt.

 Just for reference, here are the parts used for the follow focus:

Gear: https://sdp-si.com/eStore/PartDetail.asp?Opener=Group&PartID=58730&GroupID=217

(Or similar, exact specs not important unless you are trying to get a ratio right)

Belt: https://sdp-si.com/eStore/PartDetail.asp?Opener=Group&PartID=51264&GroupID=213

(Or similar, length may be changed to suit application)

I didn't calculate the reduction ratio for the focus, because for now I don't need to, and it would add a lot of complications. I'm just going to get the whole thing working before I worry about that. After I have it working I can calibrate a dial, and tweak the ratios in the software. If at some point I try to make a commercial version, I may start thinking about things like that.

Hopefully I'll have the 3rd version of the .blend file ready to be posted soon, but that's all for now, until next time! 

Satellite, asteroid, and earth

I now have the satellite, earth, and asteroid all rendering together! There are still a few glitches though, see the image below.

There are still several issues with this render. The .blend file is getting so complicated I'm going to start .blend file version 3 soon to try to simplify it. Here are the current problems I see with this render.
1. There is a outline around the asteroid.
2. The glow effect that looks good on the satellite up close doesn't scale down well. It looked O.K. when the satellite took up most of the screen, but now it just makes it look blurry.
3. I would like to have some sort of depth of field.

Anyway, I'm going to start working on blend file version 3, and see how that goes. It's really coming together nicely so far.

First scene update!

I've been working on the first scene some. I've learned quite a bit, and made some good progress. I have previously posted renders of the satellite and the earth, but never combined, as one was in Blender Internal render engine, and the other was rendered in Cycles. It turns out that compositing them is actually pretty easy. Here are some early renders, I'm pretty excited about this:

And, just for fun, here it is again with the alpha channel:

(I don't know how the blog will display the alpha channel. I uploaded it as a .png, which supports alpha channel as well as the standard RGB. In the Linux image viewer it has a nice alpha checkered background)

I'll try to post a link to the blend file soon, which brings up a question; are we willing to use Blendswap for this movie, or should we stick to posting or e-mailing links to Google Docs? Please reply in the comments below.

Here's a screenshot of the node setup, as you can see, nothing too complicated, right?

Wrong! (and only one of the node groups is expanded!)

I just figured out how to use node groups, and it can drastically cut down on clutter in the node chain.

Asteroid to be added soon!

Blender 2.64

Blender 2.64 is out! (Actually it's been out for a few days as of this post) Anyway, I thought I would share a couple tips I've found helpful or interesting.

• When rendering in Cycles, turn the viewport start resolution up to 1024 or so. (This is under the Performance panel) With a good graphics card you will see a drastic improvement when moving objects or camera in the live rendered viewport.

• Make sure the tiles (again, under the Performance panel) are set to something like 8 X 8. I got a speed improvement on my satellite model of 5X faster. (vs 1 X 1 ) 

• In user preferences, go to File, and under that, there is a checkbox that says next to it 'show thumbnails'. Make sure this is checked.

• Go through the addons section in user preferences and enable all 'add mesh' and all 'add curve'

•  Try the new 'filter glossy' next to 'no caustics'. It can be pretty handy.

• Make sure you have the GPU compute set to CUDA, not open CL if possible. (this isn't new, just something to note.)

• Check out the VSE, (Video Sequence Editor) It's much improved!

I hope those are useful, enjoy the new Blender version!

Minor details

It's kind of interesting to think about the little details, and get some specific ideas about the final product. 

(I know, I should be working on the script, but I've been making progress on that too)

Here are a number of random thoughts;

We should make a total of 12 scenes. I was originally thinking we would have 16 scenes, but I can't think of that many things to fill those spaces, so 12 would probably be better. The reason for 12 or 16 is that in the disk menus you can only fit so much, and I think you can fit 3 or 4 video tiles / icons across, and not more than 2 high. That means that you could fit 6 or 8 scenes on one menu screen, and I think 6 would work better anyway.

Color grading: We need to do this right. Once we decide what clips we are going to use, we need to take them, and exactly match all of the footage. (Except possibly the night scenes) No mood setting yet, we just make all the clips match (white balance, basic contrast, etc.) (probably rendering them each individually before putting them back into the video sequence editor) Next, we do more advanced color grading, day to night, and color mood setting. On the final output we need to do a final grade. The purpose of this is to make sure it only contain legal colors. It may slightly detract from the desired look of the film, but I want this to only contain legal colors. It probably won't matter so much since this will mostly be played on computers or DVD players, but it's good practice.

You might want to check this out: http://www.youtube.com/watch?v=PG_k2wz9mcU&hd=1

We are shooting for a general or PG audience. The only thing we need to worry about is violence. Here are the violent scenes I can think of: 1. The first gunfight. This will be set at night, only one shot will hit, and it will be dark, so not a lot of detail will be shown. 2. The car chase. the following car gets hit, swerves, and gets hit again, blowing up. Again, not a lot of detail will be shown. The first shot will hit a tire, the second shot will hit a gas tank. 3. The car stop. This is when we think that the car has a tracking device on it and one of us gets out to look for it and gets shot at. (They miss) 4. At the CIA when they are surrounded by CIA guards. We could put some action here, but there shouldn't be much.

Check out scene 7 in the script! I've worked a little on it, and I like the way it's going.

Embedded realtime blend file visualization

I just found a really cool website. It's sketchfab.com, and one of the things they let you do is .blend file visualization in real time with embedded code! It's not especially fancy, but it gets the basic idea across, and could come in handy. Here's an example:

Tracking test

Here is a tracking test that Josh and I made. It's not perfect, but we both learned a lot, and I think tracking will be important for the movie, so it's good to get some practice in.

Improved masking and keying

The mask isn't perfect, but here's a much better version: (keying too)

I'm still not spending much time on this to perfect the result but it's coming along nicely. Somewhere in the video you can see I nicked the person with the mask. It's really easy to use, but it does take a while.

Keying in Blender -now working!

I finally figured out how to get the keying to work in blender! Here's the screenshot:

It's the same footage as before, just so I have something consistent to work with.
As you can see, I just plugged the mask into the keying node using an invert node and it worked. I had tried plugging the mask into the keying node, but hadn't gotten anything before, so I don't know why it didn't work before, but I have it working now!

I'll see if I can fix the mask in the next post. I'm not going to try to track the background.

Blender mask editor

I've downloaded Blender 2.64 and am busy testing it. It is extremely unstable, to the point of crashing every other minute when I'm using the keying node in the compositor. On the other hand, it works great while it works!

The mask editor is a huge help for keying. It means that the greenscreen does not have to fill the background, or we can get rid of unwanted foreground objects. I've been working with the same test footage I had before, and have almost figured things out. Here's a screenshot of my node setup: (Click for larger)

It's not simple, I think I'm doing something wrong, and I get some weird artifacts (see red arrow) and an outline of the mask. I'm getting there though, the rod and flags in the foreground are gone!

Here's a good page about the mask editor: http://wiki.blender.org/index.php/User:Nazg-gul/MaskEditor

Blender 2.64 Testbuild!

I'm really excited about Blender 2.64. I've been trying to do chroma keying in Blender with no success because I can't figure out how to blur (feather) the mask. The new keying node should be a huge improvement. The despill node was great, but this should take the keying to professional levels.

Not only did they drastically improve the keying, planar tracking was introduced! 

This will also be a great help, and I'm really looking forward to trying it out, it's downloading as I post.

Less flag glow

I fixed the glowing flag, now all that's left for major adjustments are the solar panels

And this time the stamp is right!

Satellite update (again)

I've mostly been working on the satellite recently, and I think it's about finished. Here's what I've got, and while it will take a little tweaking, I think it should do pretty well.

It is also now clearly identified too. Showing that hammer and sickle in context with an American flag shortly after could be a good contrast.

By the way, the 100 passes note in the corner is wrong, it should be 500. I wish Cycles would let you check a box for number of passes, and then set it automatically.

Glowing satellites...

I decided to spice up the satellite a little, and here are some of the ways. The lens for the camera is also now finished, and isn't just a blank hole. I haven't decided on a color, so feel free to give input.

Cycles Satellite

Here's a quick test video render of the Cycles satellite:



Drastic overuse of the glare node. I need to fix that, but it's still a work in progress.

Cycles asteroid video test

It's not as good as the blender internal yet, and I forgot the cloud background, but it's looking usable.

Cycles Asteroid

The satellite must be cycles (I'm not doing metal in Blender Internal render engine!), so it only makes sense to do the asteroid it collides with in Cycles too:

The asteroid looks reasonable, not quite as good. I don't remember if I used the image texture or just a brown color, but the particle system is all white (it should be brown), so I need to get that fixed.

This will make rendering / compositing the first scene a lot easier once I get the world in Cycles too.

Less menacing asteroid

If we wanted something a little more bright:

Scene one early test

This was an earlier scene test video that I don't think got posted, so here it is:

I like version 3 is better, but this is one of the early tests we did. This one is also about 1/3 the length I think it should be.

Chroma keying tools

I have been working with Blender's chroma key tools, and just discovered the color spill tool. It's really great! Green bounce is always a problem when working with a greenscreen, and this tool fixes it really well. Here are two comparison images:

This is the straight keyed image.

This one has the color spill fix node.

Big difference! The main thing I love about it is that it can fix the spill without affecting the color balance.

This key obviously has some other issues, but I'm working on those too.The color spill node is a lifesaver though, and you should try it out first chance you get! Here's a screenshot of the node chain I made so you can duplicate it exactly. Nothing too complicated, just a test render.

By the way, in case you were wondering, I didn't have any greenscreen footage handy, so I downloaded the clip I'm working with from http://www.hollywoodcamerawork.us/greenscreenplates.html

It's the matchmove test close, I picked that one because it had lots of obvious green bounce.

stepper motor position control

The stepper motor now has a position control! It wasn't easy either, and it has a ways to go yet, but here's an update:

As you can see, it needs some more work yet, but it is really coming together nicely now. A few of the problems that I will try to fix are:

Jitter. It's not too bad anymore, but it's too much. I may have to use a digital potentiometer.

• Current draw. The motor is rated for 6 Amps @ 3 volts. I'm not sure that rating is right, but that was what it was specified as. The problem with that is that the transistors can't handle more than about an amp, so they are getting really hot. The other problem with the current draw is that the whole thing is running off of USB, so while this is great for testing, because it's current limited, to get the full torque out of this motor I'll need a good 3V supply and some higher current transistors. (I'm also not sure if the transistors are saturating all the way, so it might be that I just need lower voltage transistors)

• What to use for a proper enclosure.

• How to mount it to the rail system ( this is the easy one!)

Anyway, that's where I am currently. Here's the code I'm using:

---------------------------------------------------------------------------------------------------------------------
/*
 * MotorKnob
 *
 * A stepper motor follows the turns of a potentiometer
 * (or other sensor) on analog input 0.
 *
 * http://www.arduino.cc/en/Reference/Stepper
 * This example code is in the public domain.
 */

#include

// change this to the number of steps on your motor
#define STEPS 200

// create an instance of the stepper class, specifying
// the number of steps of the motor and the pins it's
// attached to
Stepper stepper(STEPS, 8, 9, 10, 11);

// the previous reading from the analog input
int previous = 0;

void setup()
{
  // set the speed of the motor to 15 RPMs
  stepper.setSpeed(15);
}

void loop()
{
  delay(100);
  // get the sensor value
  int val = analogRead(0);

  // move a number of steps equal to the change in the
  // sensor reading
  stepper.step(val - previous);

  // remember the previous value of the sensor
  previous = val;
}

---------------------------------------------------------------------------------------------------------------------

(I know that's probably not relevant to anyone else here, but I want a backup of the code online, and this is a good place to put it)

Stepper motor success at long last!

Finally!
 I now have my Arduino driving the stepper motor!
This is the motor:

Here is the driver circuit:

And here's the whole test setup:

(As you may know, I'm not especially known for neatness, as long as it works I'm happy!)

Here's the oscilloscope I'm using:

And finally, proof that it really does work:

(apologies for the poor quality)

That's it! I'll keep working on it, and hopefully add a position control soon!

DSLR Viewfinder

Have you ever tried to use a DLSR's LCD in bright sunlight?  It's a nightmare!  And since a good part of project 2.6 will be shot outdoors, I decided to buy a LCD viewfinder.  Now on B&H, viewfinders run anywhere from $60 to $500!  Viewfinders are very helpful, but are they $500 helpful?  So I then looked on Amazon, and found one made by EZ Foto for $22.  I ordered with some doubt of the quality, but was pleasantly surprised to find it rugged and sturdy.  It is attached magnetically to the LCD frame.  Here's what it looks like:

It's a dream come true in terms of precise focusing and getting white balance just right.  And it looks like you're looking at a 42 inch screen instead of the 3.5 I'm used to squinting at.  

Cheap, efficient gear like this will help project 2.6 get great shots on a low budget.  

Scene 2

First shot should have a camera pointed through a door either inside the house or at the front door. Josh runs through the door and past the camera, slight camera motion toward whatever side Josh runs by on, but don't turn the camera around. 

   Cut to a shot of Josh running toward the crash site (hidden) with camera facing Josh. Josh slows down and the camera (using the camera crane) slowly moves up to show the crash site, maybe slowly revolve around the scene? Slowly move back down as Josh approaches the crash scene. Josh looks around for a few seconds and kicks at a large piece when a smaller chunk breaks off. 

   POV switches to an over-the-shoulder shot of him reaching out and touching the piece of the satellite and burning himself. He runs out of the camera's view and the scene switches to a stationary shot of him running into the kitchen and trashing it looking for a bucket, which he impatiently fills before running out of the camera's view again. 

   Switch to another crane shot of Josh at the crash site were he dumps the water on the satellite, causing a huge cloud of water vaporous that mostly obscures the cameras view. As it clears Josh drops the bucket and goes to pick up the satellite piece gingerly. 

   Switch to a shot of Josh 

running down a hallway and into his room where he throws the piece on his desk and grabs his phone which he uses to call Zachary. 

   Switch to a shot of Zachary sitting down at his computer working on something (like turning all those tabs into BOOKMARKS). His phone rings and he answers it without even looking away from the computer. You only here one side of the conversation but Zachary stands up suddenly and runs out of the room.

Stepper motor arrival!

The stepper motor I ordered has finally arrived! Now I can get started on a follow focus!
 I have started work on it, and right now the Arduino microprocessor I'm using is programmed to tell the motor to take one step at a time. It looks everything is working in the test setup I did, but I haven't been able to get the motor to turn yet. The problem is that the Arduino outputs 5 volts at about 100 miliamps (if I remember right), but the stepper motor needs 6 amps at 3 volts, so I have to build a transistor switch setup to run it. It seems like it will be quite powerful, and might even work well for zoom, including some heavy lenses if at some time in the future I decide we want to do motorized zoom as well as focus.

I hope to post some photos soon, but the Bible Bee has started, and I've started my summer internship, so I'm going to be pretty busy.

--edit--
here's a link: http://www.allelectronics.com/make-a-store/item/SMT-119/4-LEAD-1.8-DEG-STEPPER-MOTOR/1.html

Scene 2 test

Here's our latest test, this time scene two. I am the actor you will be seeing for about a third of the movie, and yes my acting is really that bad! (according to my sister at least!)

Enjoy!
It won't spoil the plot of the movie if you watch it, but it's up to you, and we may post tests of scenes that will give it away...

DIY shoulder mount

I've recently made a dual shoulder camera mount.  I copied the design from Film Riot, a YouTube show for indy film-makers.  I modified the design a little due to lack of supplies, but overall, it came out pretty well.

As you can see, it looks extremely professional *Cough, cough* with it's duct tape finnish, and cardboard  mount.  Despite the duct tape and cardboard, it holds up surprisingly well.   Although I still don't fully trust it, and until further testing is done, I will hesitate to put a 1D mkIV with a 300mm L lens on it.   But hey, give it a break.  It only cost $10 to make!   For a pro rig like this one, we'd be looking at near a hundred or more.  

Since this is an indy film, we want to make as much of our equipment as we can, so we hope to have more DIY film gear coming soon to project 2.6!  

Camera rig

I've been working on a 15mm camera rig. A follow focus will be essential to this project, and I'll need something to mount in on once I make it, so here's what I've got so far:

The rods were from http://www.onlinemetals.com/ if I remember right, and so was the plastic. The photos should make the design so far clear. The plastic parts were cut from this plastic, 3/4" I think. Hey, it was the cheapest option by far and it works great! I cut the sections shown from a 12 X 12" piece, and this took 36 square inches (3X12) to make two brackets. I can make smaller brackets, but I wanted to make a larger one for the camera first, and I can always swap out brackets later. The measurements are 60 mm rod spacing, and 15 mm (approx., it's slightly off, but not much) diameter rods. The reason I built it to these measurements is that is what commercial DSLR rigs are using, so if we ever happen to acquire a 15mm rod accessory, (which is unlikely, as the prices are astronomical!) it should be compatible with commercial systems. I'm not the only one building or who has built a DSLR rig though, but I'm the only one who has posted it so far (hint hint, Daniel!) Next I need to work on making this into a shoulder rig, with a counterweight and shoulder pad, or I could just use it on a tripod as its main function will be for the follow focus.

Computer upgrades...the hard way

This is a long enough post that's not entirely related, so to keep clutter down here's a link to my blog:

http://kj4bxt.blogspot.com/2012/05/new-cpu-cooler.html

Check it out, it's not the usual!

Satellite parts?

When the satellite is in outer space I can make it look like anything (within my limited modeling / texturing abilities), but on the ground we have to have actual physical parts. I've started keeping my eyes open for possible satellite parts. They don't have to bear any real resemblance to the satellite, because it will have been scorched beyond recognition during its reentry of earth's atmosphere (they just have to look plausible and cool)
Here's some of the things I'm considering:

No shortage of (old) computer memory! One of these will probably be the satellite part we hack.

This is an old RF amp I had lying around. It looks interesting enough I think it could be used. ( I have two, the photo below is of the same one, just a different angle)

So, that's the latest from Project 2.6!

(Actually it isn't, the latest news is that I can't get my asteroid field to render with the satellite and world...but that will be fixed...shortly...I hope...) :D

More concept art

Here's where I am on a satellite model. It's coming along, but realism is sadly lacking. It just doesn't look right to me, try as I will... Still, it's a start unless we decide to go with a different model.

Right now the main thing I could do to improve realism would probably be to use an image texture for the solar panels.

Our most exotic location in the movie...

So, for the satellite entry scene we will need to show the satellite flying through earth's atmosphere. There are two ways to do this, and as we've already disclosed that we're using Blender and have essentially no budget you can probably guess how we're doing it! Here are some test renders.

First one, this looks pretty good, but needs some tweaking. Mixing in some fast Gaussian blur could probably improve this. It is the only one that has a starfield, because I was having too much trouble trying to keep the blur from affecting the stars, which just made it look weird. 

Second render; this is better, I also changed the lighting to get a better feel for how it would look. This is the only render that I consider totally unacceptable. Why? Because there is a wireframe looking texture in some of the clouds, really killing realism. It took a while to figure out how to get rid of it.

This is pretty decent.  I wish it had a starfield, but until I figure out an easier way to do it, it will have to wait. I'll probably leave it at this point for now, and through the test scenes. I may tweak it from time to time though, probably nothing major, other things will now take priority.

By the way, this is with Blender internal. I would like to use Cycles for more things, but I already had made this a few months ago with Blender internal render engine, and it would be a lot of work to redo it. Besides, I'm getting 20 second render times with my CPU overclocked to 3.15 GHz, and I don't think Cycles would really have any appreciable time improvement over that. Anyway, lots of fun to make. I'm hoping to start work on the satellite soon so I can post a test version of the first scene.

New video card for cycles!

Project 2.6 will now be able to render using the Cycles render!  We just got a Nvidia 560 with 2gb of DDR5 memory.  This will allow us to successfully render using cycles without crashing the program.

When idle, the 560 runs about 30 degrees Celsius, but after 1000 passes in cycles in jumps up to 75 degrees, and can render 6000 passes in just over 8 minutes.  Which is more than 10 times faster than the CPU!

We will hopefully have some more statistics and test shots later to show just what can be done with the Cycles render engine.  

Cycles!

Cycles may be a possibility for this project after all! We're having some posting / authorization issues right now, but hopefully by Saturday we should have that straightened out. Once we get that fixed we'll try to post more about Cycles and our current setup... :)

More fire

More fire!



Just playing around with a blend file... It looks pretty weird though, I may change the seed and tweak some settings to see what else I can do.

Celtx vs. Google Docs

I've tried Celtx and I've tried Google Docs. Celtx doesn't work nearly as well when multiple people are working on the same thing.  So, we'll probably leave the script in Google Docs. Celtix can be useful for making lists, and possibly some other more in depth scene breakdown, but for now, nothing major is going to change.

First Scene Location Scouting

Location scouting; Right now, I'm looking for a place to shoot the first scene. We have three houses right now that we can pick from (ours), but there's a catch. What I want to do is to have a room with some chairs / sofa and a window to the backyard. One of our houses doesn't have a backyard that I know of. Our house has a room that might work, except that the view from the room is of our deck, not the ground. It might work, but would be less than ideal. The other house I haven't seen the backyard. 

So...

Daniel, does your house have any rooms with windows not very far off the ground in the back, and if so, could you send or post some photos?

We also have two other options. One is to greenscreen the window, for the video of the impact. I don't want to do this, as that means we will have to shoot locked down, which won't look as good, or use motion tracking, which I think we're going to be doing enough of already. The other is to have the satellite fall into the front yard. This is worth considering, especially for the last scene, when Zeke has a bunch of his friends around it, telling them what happened.

So, what do you think? What should we try? We have to decide soon, because we want to shoot the trailer soon.

I think we should shoot it so that the satellite falls in the front yard. If people agree, I'll change the script.

If we did this, we would preferably shoot it at your house, Daniel. Are there any rooms that would work for this?

Camera Ladder

Camera cranes are great, but commercial ones are really expensive. I want to use one on this project, so I have to figure out how to make one. I've posted about my first ones, but I'm also looking at trying to using a ladder like this one as the arm. It has the advantage of being (slightly) collapsible, and probably lighter weight, but it's also more flexible than the triangle design of the tower that I was using, so there might be bounce on faster moving shots. Only a test run will show, so that's on the list of things to try. It's not very high on the list though, because I already have something that I think will work.

2.5K camera

The market for high resolution digital cinema cameras is really expanding. It's really interesting watching right now, when there's so much competition. Prices are also falling, which I like! Right now, I know of two different 2K RAW cameras for about $3000. That's still pretty expensive, and they don't include lenses either, but that's still less than something like Canon's 5D MK III.

Below: A photo of Black Magic Design's 2.5K camera from their website.

http://www.blackmagic-design.com/products/blackmagiccinemacamera/

It's really neat seeing what's happening right now. Another big plus of this camera is that it records to 2.5" SSDs, so recording media will be easily available, not proprietary, and it will also be easy and quick to transfer data.
Anyway, this isn't exactly related to the project, but it is of general interest. At the rate things are going, I wouldn't be surprised if we shot our next film in 2K! Maybe wishful thinking, but you never know...

Now those are some good fireballs:

http://mango.blender.org/random/smoke-simulation-tricks/

That's what we'll aim for in our movie!



It looks terrific, but I'm 70% through baking, and while it's manageable, it sure takes a long time!

720p vs 1080p

I really don't want to make this choice... But, it does have to be made. If I remember right I said a while ago that we would probably shoot this in 720p. However, I've been working with video, (from the 1D MK IV), and 1080p really looks drastically better than 720p. Some of this may be due to the fact that I'm viewing the video on a real screen, uncompressed from the camera, not a YouTube version on a smartphone. When we're done though, I want to be able to put this in our home theater and show it to friends without being ashamed of it (thought I'm sure I will be able to see / hear flaws)
The problem here is that we can shoot 24 or 30 frames per second 1080p, OR 60fps 720p. The 1080p looks much better, but it only takes 30 fps, so slow motion will be very limited. Right now, I'm thinking we will probably shoot the trailer in full 1080p to get a feel for it, and see if that will be feasible for the whole thing. It's possible that we could shoot the film in 1080p, and switch to 720p for the slow motion parts, but I don't think we'll do that.

Trailer coming soon!

(we hope)

O.K., we've decided to start work on a trailer for the movie. That will give us some idea of how things are going, and what will need to happen. It will be a good chance to test out a number of ideas and tweak our setup. We'll continue work on the script in general too, but focus will be shifted to the trailer, because we can test out how things will work with that.

A while back I had a little fun with some trailer templates from my Magix video editor. Here's one of them:

Improved camera

Now with aperture!

In case anyone's curious, it takes about 40 sec- 1 minute per frame to render with a node chain. (That's with my CPU overclocked) 

Concept Artwork

Click for full 2K! This is the current result of some of my work for this movie. This is supposed to be a lens for the satellite's camera. (the gray box at the bottom will be the camera / satellite) The boxes around the lens are supposed to be the controls for focus and zoom.

What do you think? I know it still needs a lot of work, especially texturing. What could be improved? I'm thinking of adding an iris behind the lens, but haven't yet.