Movie status: a couple of weeks ago we sat down and went through the finished segment of the movie with a fine toothed comb and made a comprehensive list of what audio effects will be needed for the movie. I made a list a couple of pages long, so this should be an interesting process. Andy and Rob want to record all new foley and effects work instead of relying on library sounds. We're due to begin that in a couple of weeks.
In other 3-D movie news, we skipped out on the Superman Returns Imax 3D Experience due to the fact that only 20 minutes of the movie were rendered in 3-D; next week, however, signals the release of Monster House in "Real-D" digital 3-D, so we'll be making a trip into Schaumberg to see that in a digital cinema. Also, Night of the Living Dead 3D will be getting a nationwide release this fall to 1500 theaters - which means it may actually play in here in Rockford!
And, Frankenstein vs the Wolfman got a mention in this month's issue of 3rd Dimension Graphics, an online Poser eZine.
Friday, July 14, 2006
Wednesday, June 7, 2006
3-D With After Effects
It's been hectic around the household lately as I've started a new position at work, one that has finally moved me from 10 years of graveyard shift 180 degrees to a normal daytime office schedule. I'm still adjusting to the change and am not getting the best night's sleep.
As such, I'm on pause with the movie after having completed the first act, and have been making demo discs as a rehersal for the final post-production phase.
When I started the project, I was planning on using a freeware program called Stereo Movie Maker to convert the left/right movies into final anaglyph and interlaced 3-D; however, I discovered last weekend that not only does Stereo Movie Maker use the old Video for Windows API, which imposes a 2GB limitation on files, but it also has a habit of converting my footage rendered at 1.2 pixel resolution (for 16:9 widescreen) to .09 pixel resolution DV. I end up with a 720x480 image that is taller than wider, and everyone is stretched out.
I have also encountered a bug in Premiere that renders black for about half of the movie when using the Huffyuv 2.1.1 codec (upgrading to 2.2.0 only made matters worse). I was forced to use a frameserver to output the left and right movies from the Premiere timeline, and then bring those files into After Effects and apply the "3-D Glasses" filter, thus compressing the movie 3x - once on the raw Poser data>editable clips, once from Premiere>conformed movie and then again from After Effects>DVD/Windows Media/etc.
However, with the help of one of the production guys at work I found an even better method: importing the Premiere project directly into After Effects, selecting all the layers and using the Pre-Compose command to flatten them into a virtual movie clip, then doing the same with the right eye version. After I bring them both into the After Effects timeline, I can apply the 3-D conversion without re-compressing the video. Now, we go from raw Poser data>Huffyuv clips>final output (DVD).
I've also discovered that the 3-D Glasses plugin for After Effects is virutally useless - no matter how much I tweak it, I cannot get the colors to the right phase of red and blue to be filtered out by the anaglyph glasses, either on the PC or on a TV screen. I wanted to use Stereo Movie Maker initially because it uses a special algorthm to produce a "half color" anaglyph which reduces retinal rivalry caused when a character wears red or blue clothing. Retinal rivalry refers to the phenomena that occurs when the red lens of the glasses makes red objects appear "white" and the blue lens sees them as "black". The half color method converts the left eye to grayscale and then siphons off the red channel, allowing red colored objects to be perceived as grey.
Without Stereo Movie Maker, I'm forced to use another After Effects plugin called Set Channels, which allows you to use the red channel from the left eye and superimpose it with the green and blue channels (creating a cyan channel) of the right. By using telling the plugin to use the green or blue channel from the left eye to extract the red from, instead of the actual red channel, it produce something closer to the half color anaglyph. The result is much easier on the eyes and works best on a PC monitor, and relatively well on a television set.
And, not satisfied with the opening title sequence yet again, I created a new version in which the main title actually shoots out at you. It's more impressive than the previous design, in which the titles started out a bit off the screen and slowly zoomed out. These new ones start right at the 2-D plane of the movie screen and zip out at you in about half a second. It's much more dramatic.
As such, I'm on pause with the movie after having completed the first act, and have been making demo discs as a rehersal for the final post-production phase.
When I started the project, I was planning on using a freeware program called Stereo Movie Maker to convert the left/right movies into final anaglyph and interlaced 3-D; however, I discovered last weekend that not only does Stereo Movie Maker use the old Video for Windows API, which imposes a 2GB limitation on files, but it also has a habit of converting my footage rendered at 1.2 pixel resolution (for 16:9 widescreen) to .09 pixel resolution DV. I end up with a 720x480 image that is taller than wider, and everyone is stretched out.
I have also encountered a bug in Premiere that renders black for about half of the movie when using the Huffyuv 2.1.1 codec (upgrading to 2.2.0 only made matters worse). I was forced to use a frameserver to output the left and right movies from the Premiere timeline, and then bring those files into After Effects and apply the "3-D Glasses" filter, thus compressing the movie 3x - once on the raw Poser data>editable clips, once from Premiere>conformed movie and then again from After Effects>DVD/Windows Media/etc.
However, with the help of one of the production guys at work I found an even better method: importing the Premiere project directly into After Effects, selecting all the layers and using the Pre-Compose command to flatten them into a virtual movie clip, then doing the same with the right eye version. After I bring them both into the After Effects timeline, I can apply the 3-D conversion without re-compressing the video. Now, we go from raw Poser data>Huffyuv clips>final output (DVD).
I've also discovered that the 3-D Glasses plugin for After Effects is virutally useless - no matter how much I tweak it, I cannot get the colors to the right phase of red and blue to be filtered out by the anaglyph glasses, either on the PC or on a TV screen. I wanted to use Stereo Movie Maker initially because it uses a special algorthm to produce a "half color" anaglyph which reduces retinal rivalry caused when a character wears red or blue clothing. Retinal rivalry refers to the phenomena that occurs when the red lens of the glasses makes red objects appear "white" and the blue lens sees them as "black". The half color method converts the left eye to grayscale and then siphons off the red channel, allowing red colored objects to be perceived as grey.
Without Stereo Movie Maker, I'm forced to use another After Effects plugin called Set Channels, which allows you to use the red channel from the left eye and superimpose it with the green and blue channels (creating a cyan channel) of the right. By using telling the plugin to use the green or blue channel from the left eye to extract the red from, instead of the actual red channel, it produce something closer to the half color anaglyph. The result is much easier on the eyes and works best on a PC monitor, and relatively well on a television set.
And, not satisfied with the opening title sequence yet again, I created a new version in which the main title actually shoots out at you. It's more impressive than the previous design, in which the titles started out a bit off the screen and slowly zoomed out. These new ones start right at the 2-D plane of the movie screen and zip out at you in about half a second. It's much more dramatic.
Wednesday, May 31, 2006
Halfway Point?
Added the Poster (credits not final) and some wallpapers to the active Poster page of the site for all of you FRANKENSTEIN VS THE WOLFMAN enthusiasts who'd like to have the boys staring back at you from your computer screen.
I've just completed that pesky sequence of shots that links up two batches of completed footage, marking the completion of about 7 1/2 minutes worth of narrative on FRANKENSTEIN VS THE WOLFMAN. The spooky thing about watching it back is just how fast the story moves. In my head, I guess I envisioned it having a slower pace. Instead, I've made FRANKENSTEIN VS THE WOLFMAN: THE ACTION MOVIE. It's pretty much BANG right out of the gate, in a race towards the finish line.
There are about a total of 12 minutes of finished sequences. These finished 8 minutes constitute about 10 pages of the script, which is 27 pages in total. I'm still going to bet the movie will come in at 20 minutes in length when it's done.
Adobe Premiere is rendering out the Left Eye view right now for distribution to my partners in crime. In the meantime, check out this site I discovered for an open source CG movie - the program used to create it and all the project files are available free. Visit Elephant's Dream.
I've just completed that pesky sequence of shots that links up two batches of completed footage, marking the completion of about 7 1/2 minutes worth of narrative on FRANKENSTEIN VS THE WOLFMAN. The spooky thing about watching it back is just how fast the story moves. In my head, I guess I envisioned it having a slower pace. Instead, I've made FRANKENSTEIN VS THE WOLFMAN: THE ACTION MOVIE. It's pretty much BANG right out of the gate, in a race towards the finish line.
There are about a total of 12 minutes of finished sequences. These finished 8 minutes constitute about 10 pages of the script, which is 27 pages in total. I'm still going to bet the movie will come in at 20 minutes in length when it's done.
Adobe Premiere is rendering out the Left Eye view right now for distribution to my partners in crime. In the meantime, check out this site I discovered for an open source CG movie - the program used to create it and all the project files are available free. Visit Elephant's Dream.
Thursday, May 18, 2006
Final Voice Recordings Completed
This past Monday we went into the new Darkhouse Productions studio (visit Darkhouseonline.com) and recorded the final three vocal performances for the movie. Darkhouse is a company co-founded by our producer, Andy Carlson, which specializes in audio recording, mixing and mastering for bands and vocal talent.
In other news, Lipsync's Mimic is becoming more of a curse than a blessing. It's voice recognition ability is fairly simple and rarely gets the lip movements to match the spoken word. For this latest scene I ended up becoming so frustrated with it that I went back to the old Raven / Raven 2 way of manually doing the lip sync using Poser's phoneme library. Time consuming, yes... but for this instance it produced more accurate results.
In other news, Lipsync's Mimic is becoming more of a curse than a blessing. It's voice recognition ability is fairly simple and rarely gets the lip movements to match the spoken word. For this latest scene I ended up becoming so frustrated with it that I went back to the old Raven / Raven 2 way of manually doing the lip sync using Poser's phoneme library. Time consuming, yes... but for this instance it produced more accurate results.
Monday, May 8, 2006
Full Speed Ahead
Despite the lack of updates to the Production Journal, animation on the movie is progressing at a good clip. There are two short sequences left to do that will act as a bridge between two lengthy sections of the film; once they are completed I will have a uninterrupted seven minutes that I can farm out to my music and sound guys for their analysis. It's exciting to see the movie's narrative start to emerge.
Co-producer/post audio supervisor/casting director Andy has scheduled a recording session for the final three voice parts this Thursday. Once I've got them, I'm anxious to begin work on a flashback sequence that I intend to model on the aesthetic of the Universal films.
More as it happens...
Co-producer/post audio supervisor/casting director Andy has scheduled a recording session for the final three voice parts this Thursday. Once I've got them, I'm anxious to begin work on a flashback sequence that I intend to model on the aesthetic of the Universal films.
More as it happens...
Wednesday, April 5, 2006
Restructuring the Workflow
Still working on the aforementioned animation sequence; its looking pretty good. There's a special effect involved now, a lit torch weilded by the Monster. I'm using Panopticum Fire in After Effects and tackling it by producing the shots in layers.
The couple of days ago I decided that the storage space issue needed to be resolved before things went too much farther. I've been producing by taking the PSD frame files from Poser and converting them into uncompressed AVIs. These resultant files are huge, and add to that the fact that I'm producing four versions of the movie (left eye, right eye, anaglyph, interlaced). I am working with an 80GB storage drive and the space available was no longer efficient.
So, I did some investigation and decided to recompress the footage with a lossless video codec - in this case Huffyuv. In the process, I took a look at the completed footage and did some more tweaking. I've never been happy with the look of the opening scene of the film, as it looks very Poser-ish. It takes place during broad daylight and there's no room for any mood lighting. Everything is very bright and sterile.
But not anymore. I went back and color corrected the hell out of not only the opening scene, but also a couple of other sequences later in the film, brining them more in line with the lighting schemes present in surrounding scenes.
Now that I have the left and right eye video files at a more manageable level, I've been re-evaluating the post-production 3D process as well. My original idea was to produce four different versions of each shot, but give them all the same name and store them in seperate folders. I would edit the left eye version in Premiere Pro, then export the finished movie directly to DVD compliant MPEG 2. Following that, I would load the anaglyph version - since the clips bear the same name, all the edits would be kept intact - and render that out, and so on.
After doing some tests with some After Effects plugins (including one made just for this purpose, called 3D Glasses), my new strategy is to have only two versions of the movie on my HD at any given time, the right and left eye versions. Once the edit is finished, I'll create full Huffyuv encoded right & left streams and combine them in After Effects. From there I can produce both anaglyph and interlaced composite MPEG 2 files for DVD distribution.
I'm also leaning towards Windows Media 9 for all video downloads and trailers that will eventually be available from this site.
The biggest downside I've seen this week is my inability to get an anaglyph version of the movie that will work well on NTSC television sets. All test versions produced work extremely well on a computer monitor, but end up producing a substantial amount of ghosting when played back on a TV. Hopefully, the anaglyph will test well on a front projection system. But right now it seems like the optimal way to see the movie in 3D on your TV is via the field sequential shutterglass version, and on your computer via the anaglyph version.
The couple of days ago I decided that the storage space issue needed to be resolved before things went too much farther. I've been producing by taking the PSD frame files from Poser and converting them into uncompressed AVIs. These resultant files are huge, and add to that the fact that I'm producing four versions of the movie (left eye, right eye, anaglyph, interlaced). I am working with an 80GB storage drive and the space available was no longer efficient.
So, I did some investigation and decided to recompress the footage with a lossless video codec - in this case Huffyuv. In the process, I took a look at the completed footage and did some more tweaking. I've never been happy with the look of the opening scene of the film, as it looks very Poser-ish. It takes place during broad daylight and there's no room for any mood lighting. Everything is very bright and sterile.
But not anymore. I went back and color corrected the hell out of not only the opening scene, but also a couple of other sequences later in the film, brining them more in line with the lighting schemes present in surrounding scenes.
Now that I have the left and right eye video files at a more manageable level, I've been re-evaluating the post-production 3D process as well. My original idea was to produce four different versions of each shot, but give them all the same name and store them in seperate folders. I would edit the left eye version in Premiere Pro, then export the finished movie directly to DVD compliant MPEG 2. Following that, I would load the anaglyph version - since the clips bear the same name, all the edits would be kept intact - and render that out, and so on.
After doing some tests with some After Effects plugins (including one made just for this purpose, called 3D Glasses), my new strategy is to have only two versions of the movie on my HD at any given time, the right and left eye versions. Once the edit is finished, I'll create full Huffyuv encoded right & left streams and combine them in After Effects. From there I can produce both anaglyph and interlaced composite MPEG 2 files for DVD distribution.
I'm also leaning towards Windows Media 9 for all video downloads and trailers that will eventually be available from this site.
The biggest downside I've seen this week is my inability to get an anaglyph version of the movie that will work well on NTSC television sets. All test versions produced work extremely well on a computer monitor, but end up producing a substantial amount of ghosting when played back on a TV. Hopefully, the anaglyph will test well on a front projection system. But right now it seems like the optimal way to see the movie in 3D on your TV is via the field sequential shutterglass version, and on your computer via the anaglyph version.
Wednesday, March 22, 2006
I'm Not Dead
Yep, lo and behold, I am still alive. I took a break for the December holidays and couldn't work myself up to come back to the painstaking process of animation. But now, at long last, I hope to make a return to the project.
I wish I could report more, but truth is I'm still working on that same chase sequence. Our heroine, Leeta, has been cornered in a sinkhole by the Wolfman and is about to meet the Frankenstein Monster and his Golden Retreiver, Yip. Action scenes, as I'm sure I've stated before, are my bete noir. They require multiple setups, each only seconds long, and take forever. I'm much more comfortable laying out lengthy dialogue sequences, where the focus is mostly on camera placement and character staging.
Andy has been working 'round the clock building a storefront recording studio, the first tentpole in the Darkhouse Productions umbrella. As such, he's been unable to get time to edit down the remaining character voice tracks. Unfortunately, two supporting characters who appear in numerous scenes are included in that batch... which forces me to work on the action sequences.
Thankfully, along with Spring comes the start of festival submission season - and although I have nothing ready for prime-time this year, I need to get cracking on this sucker in order to have something ready for next year!
Stay tuned ...
I wish I could report more, but truth is I'm still working on that same chase sequence. Our heroine, Leeta, has been cornered in a sinkhole by the Wolfman and is about to meet the Frankenstein Monster and his Golden Retreiver, Yip. Action scenes, as I'm sure I've stated before, are my bete noir. They require multiple setups, each only seconds long, and take forever. I'm much more comfortable laying out lengthy dialogue sequences, where the focus is mostly on camera placement and character staging.
Andy has been working 'round the clock building a storefront recording studio, the first tentpole in the Darkhouse Productions umbrella. As such, he's been unable to get time to edit down the remaining character voice tracks. Unfortunately, two supporting characters who appear in numerous scenes are included in that batch... which forces me to work on the action sequences.
Thankfully, along with Spring comes the start of festival submission season - and although I have nothing ready for prime-time this year, I need to get cracking on this sucker in order to have something ready for next year!
Stay tuned ...
Subscribe to:
Posts (Atom)
