Brendan Coots
Forum Replies Created
-
I wouldn’t even try something like this in After Effects. The look you describe would most likely be done in a 3D app like Maya or 3D Studio Max. Even then, that would be an extremely challenging project to build.
-
Are you on a Mac or PC?
It sounds like an issue with hard drive space, file system type (i.e. FAT32 vs. NTFS vs. HFS) or something similar.
-
Downcoversions should look fine, although using the P2 cards with HD takes some getting used to, primarily in that you have to switch cards out fairly often and dump to drive.
If you really want a good clean (and filmic) look, shooting with the HVX200 will help simply because it’s 4:2:2 rather than DV’s 4:1:1, but also the native 16:9 aspect ratio.
-
For the record you’re thinking of the Panasonic DVX series, not HVX. DVX100B was released a year or so after the original (DVX100A), and just has a few minor updates. The primary differences are that the body is a slightly different material to better comply with environmental law, and it has a stronger baseplate. People were complaining about bottom of the DVX100A body cracking when it was attached to a tripod and receiving stress.
-
Your video card probably isn’t the issue, but your lack of RAM sure might be. 1GB is barely enough for the OS and AE to be open, much less processing anything within AE or rendering out. It’s easy to underestimate how much RAM is enough, but when in doubt get as much as you can afford. You’ll almost always notice the difference.
The other problem is simply that a G5 dual-2.0 is a bit dated. I have a recent macbook (not the pro) and in speed tests it blew the dual G5 desktops away. My only point is that the Core Duo (and Core 2 Duo) chips are way more efficient and speedy.
All of that said, I know you’re on a budget. One of the best things you could do is invest in a bunch of ram – I believe 4GB is the max that line of G5 can hold. Other World Computing is a GREAT online store for mac components, I just purchased a bunch of RAM from them for less than half of what Apple charges, and so far it’s very solid. Their site will ask which mac you have and make the appropriate memory module suggestions in terms of what type, speed etc. I would recommend the OWC brand:
https://eshop.macsales.com/shop/apple/memory/
Good Luck!
-
This depends entirely on your budget, because the sky is the limit. As mentioned by SR, your ideal camera will shoot progressive. Another key issue is the quality of the footage being shot, which can be distilled into two basic categories – format, and sampling.
With regards to format, the ideal would be a camera that offers better options than just straight DV. DV is okay for some things, but the DV format is only 25Mbps. In simple terms this just means there isn’t a whole lot of data there, resulting in less color, less vivid images and ultimately less sharpness of detail. Other formats like DVCPRO 50 offer much higher datarates and thus better detail, etc.
On the issue of sampling, it’s complicated but all digital cameras “sample” the brightness and the colors of the image separately, combining them to form a complete image. A ratio is used to describe how any given camera performs.
As an example, consumer-grade DV cameras shoot 4:1:1. This means that for every four brightness samples (the first number), the camera only records 1 sample for color (the 2nd and 3rd number both represent color). The result is that DV doesn’t yield great color results, making it tough to do color correction, greenscreen or other color-sensitive operations. Some of the higher level cameras shoot 4:2:2, which as you can guess samples color twice for every 4 brightness samples. High-end HD cameras and film sample at 4:4:4, meaning a complete image with no data left out.
So, while that’s a lot of info to consider, just focus on your needs. If you want good crisp imagery, or plan to do greenscreen work, you want a camera capable of shooting 4:2:2, and ideally offering more format choices than just DV. Personally, I would not recommend purchasing an HDV camera. The footage is difficult to work with and is 4:2:0, which is just as bad as DV in terms of color rendition. Many people report the footage plays back dog-slow in Final Cut Pro and After Effects, and isn’t friendly to work with.
Panasonic’s HVX200 is an excellent camera, but it costs around $5k. It offers 4:2:2 sampling, progressive shooting, and can shoot in DV, DVCPRO50, High-definition in the form of DVCPRO HD (not the inferior HDV) and more. It also does slow motion and a slew of other neat tricks. My studio owns one and use it regularly on professional-grade shoots.
If you need something cheaper, the little brother to the above camera, the Panasonic DVX100, is a nice DV camera with more pro features than most in its price class. We also own one of these and for DV it’s pretty great.
-
Unless you’re using some newfangled camera I’m not aware of, HDV’s color sampling is pretty poor (4:2:0, the same basic color sampling as DV, just using a different pattern). That in itself is going to present a major challenge.
Many of the suggestions here focus on outputting to another format or uncompressed to do the key in After Effects, but that won’t get you any more sampling info out of the footage (which is going to be your biggest roadblock to getting a usable key), and even worse the two extra compression stages needed for this workflow (render out of FCP, then render out of AE) will degrade your footage even more. Another thing to consider – the output from After Effects will suffer serious quality destruction unless you output to an uncompressed/lossless format. For 3 hours of uncompressed footage with alpha, you’re looking at hundreds upon hundreds of gigs (possibly even approaching 1 terrabyte) of storage needed. If it were my project I would conclude that this workflow will consumer way too much time, resources and incur a lot of potential quality loss, so it’s probably better to stay in Final Cut Pro for your entire workflow.
Because of this, I would personally just use Final Cut’s built-in keyer, which is decent enough, or go buy the DV Matte plugin for FCP, which is going to save you a LOT of headaches for only $199. There is a free demo so you can test it before buying, as well:
https://www.dvgarage.com/prod/prod.php?prod=dvmattefcp
Then you can layer your backgrounds and other elements within the FCP timeline, and make changes without the worry of hopping BACK over to AE for tweaks. If the shoot was lit properly and the quality of the footage is nice and crisp, you’ll be fine with either of these options.
Ultimately, there isn’t much you can do about the massive render time you are about to incur if the final run time is 3 hours, I would just prepare for that and maybe set it to render over the weekend or something.
-
According to Adobe project files created in the beta will open in the final release version of AE CS3.
-
Here’s a few tricks that will help. The first is matching values:
1. Apply a Levels effect to the greenscreened footage.
2. Adjust the blacks and whites to match the background. Once those match reasonably, you should also look at the color of the shadows and highlights. Most shadow areas aren’t pure black but rather a cool or warm color. Match those using the individual color channels in the Levels.
3. Create an adjustment layer above everything. Apply Levels to it, and crank the white values all the way (or most of the way) up. This will blow out the footage, but it will reveal subtle differences between your foreground and background. Identify what needs to be fixed to match the two, then temporarily hide the Adjustment layer to make fixes to the Levels on your greenscreen footage. Repeat as needed.
4. The most important thing is to match your shadows and highlights. Not just how much of the subject is in shadow, but the quality of the shadow as well. Are the shadows on the background plate milky? Deep black? Warm? Cool? Whatever it is, your greenscreen shot won’t look natural until the shadows and highlights match, then finally the mids.Another technique for matching direct lighting from background plates:
1. Examine your background to see where the primary light source is coming from, and what kind of light it is. For example, is it bright sunlight coming from the right?
2. Create an adjustment layer above your greenscreen layer. Apply levels, and crank up the whites some.
3. create a mask on the adjustment layer, using the pen tool, that is the shape of light spilling onto your actor. For example, if the light is coming from the upper right, your mask would cover the top of the actor’s head, the camera-right shoulder, and on down the camera-right arm. Feather the mask enough to make the effect look natural.
4. Go back into the Levels on this adjustment layer and reset it. Now you want to dial in settings that match the light source in the background footage. For sunlight, this usually involved upping the input whites, warming it up a touch etc.
5. For this trick to work, you will need to roughly keyframe your mask to follow the actor, but the end result is worth it if done right.
6. When you are done, you will need to precomp the greenscreen footage and the masked adjustment layer to keep the Levels effect from affecting the background as well.There are several other techniques, but these two should get you close.
-
Right, I forgot about the spill supression/grey pixels thing. I guess I’ve just been lucky and never really encountered that.