Eolirin said:
Ionami said:
If you were to "record" it the way you're suggesting, you'll lose significant picture quality. Also, it is a video file. And it is scripted to play at a certain point within the demo.
The movies are not being rendered by the game's engine or anything like that, they're pre-rendered videos. The point is that you still have all of that information on screen, and it's a lot of data to process. The program that creates the videos is where they would export from, and that original video file would be massive. They could compress that video, but then they would lose the video quality. (Even with a lossless compressor, the size is still huge.)
HD footage typically works out to be about 1 minute of HD video = 1 GB of data. And that's talking about a handheld, home video camera. If you create CG images, with tons of special effects, and export out as HD video, you're file sizes are going to be massive.
Point being, they clearly did all the compressing they could to the videos, and that's about as small as they're going to get without sacrificing video quality.
Also, most movies these days still shoot on film. So an HD-DVD video was not actually shot on HD footage necessarily. It was just converted to HD for the DVD version.
It's starting to become more popular, but it's still quite recent. Michael Mann is one of the folks who have started shooting HD Hollywood status films. i.e. "Collateral" and "Miami Vice".
But... that's not true. At all. The size of uncompressed video is a factor of length and resolution, not of what's actually in the image. A pixel is a pixel is a pixel; they take up exactly the same amount of space regardless of what's in them. A completely black frame is exactly the same size as one that has several million colors. So special effects have zero impact on the size of that data. And compression algorithms aren't particularly partial to them either: depending on the method being used a particular frame will be better or worse compressed based on how many parts of it are well suited to the algorithm. The "quality" of the visuals isn't particularly relevant, only the suitability for the compression technique; lots of special effects could theoretically improve the compression rate for a given algorithm, depending on their nature.
But... it IS true.
I work with both CG AND HD footage. A pixel with nothing but solid black in it, renders MUCH faster and contains less data than a pixel with something in it. And you cannot convince me otherwise, because this is my job, and I see it happen EVERY DAY.
When we export footage and images from our CG software, it is ALWAYS much higher in size than our regular HD footage. And our regular HD footage is STILL massive. We have to compress and re-export our footage many times to get it ready for internet streaming. The final product, although it looks amazing, is still nowhere near as good quality as when we started, and the file sizes have dropped a LOT. i.e. a 40GB video has suddenly become 450 MB.
So before compression, the video is massive. After compression it is a fraction of the size. The image quality is still good. Clearly, compression has a factor in the size of the video.
The more special effects you have in a shot, the longer the render time, the more objects to render, the larger the final export. i.e. the more special effects and objects you have on screen, the more rendering that needs to happen.
EDIT:
After re-reading your post, I don't think we're talking about the same thing here.
I'm an editor, and 3D modeler. The material I'm talking about is the raw source footage that we either capture right on our HD cameras, and the 3D materials we create. A minute of HD footage that we capture, at 720p 60fps equals 1 gigabyte of data. We have a P2 Memory Card that holds 17GB, which means we can only capture 17 minutes of footage on that camera, before we have to stop and dump the footage onto a hard drive, and then format the card.
To make that footage watchable, it needs to be compressed multiple times (This goes for the 3D materials as well) with 2 different codecs. Say if we have a 15 minute video, that means the file size on that video will be at LEAST 15 GB, if not more. If I export out a 3D scene at a resolution of 6400x4800, and it's 3 minutes long, the file sizes will sometimes be over 100GB, depending on how much special effects are in that scene.
6400x4800 is an insanely high resolution, I know. But that is the size we need to work with, to make our final product look as good as it does. So, working from the source footage that can be hundreds of GB for a single video, it's not really surprising to think that an extremely detailed CG video with uncompressed audio/video might be 3-4 GB in size.