Join Larry Jordan for an in-depth discussion in this video Preparing and sizing images , part of Final Cut Studio 2: Moving on Stills .
Of all the sections in this training, this section is probably the most important. Not because it's going to show you how the software works but because it's going to explain the questions that I get day after day in my e-mail which is what size and why doesn't it look good. So we're going to start by giving you some rules to working with images. I will explain in brief the differences between video and computer images. We will determine the correct image size to use for both standard def video and high-def video and then we will wrap up with some cautionary notes to make sure your images look as good as possible.
Then we'll be done with all the preamble, important though it is, and we'll move into Photoshop and show you how to prepare your images inside Adobe Photoshop. Let's start with some rules to working with images, and here is the key point. Computer images use square pixels and video images use rectangular pixels. When you're preparing your images you must compensate for this. All video images are always 72 dpi. No, it's not that they're really 72 dpi, it's that we don't care about the dpi.
It's the total number of pixels across by the total number of pixels high. Sort of a convention. We make all of our images 72 dpi so we count pixels, not dots per inch. Images that are used in video should be bitmaps like PNGs or TIFFs or PSDs. You want to avoid JPEG where possible because there's way too much quality loss. There's artifacting and blockiness and colors tend to get kind of grungy. Also video cannot use vector images. You need to convert them to bitmap first, so I can't take an Illustrator file and pull it in or any other vector-based drawing.
It needs to be converted to a bitmap, which is why Photoshop is so essential to video because Photoshop is at its core a bitmap image editor. Also, don't zoom an image larger than 100%, even though you want to, even though you feel the strong need, don't zoom your images larger than 100%. Because bitmapped images lose quality when they are enlarged. So that means, that if you want to move around inside an image so you can sort of zoom into it, then that image needs to be larger than full-screen.
Now, this is a weird concept when all we're used to working with is a video image, but with still images, Ah! The rules change and that's what we're here to talk about. We are used to creating images and working with pictures on our computer and although it looks like it's the same picture on a computer as it does on the video screen, they are not the same. They are different because a video image is fixed in size. Whether you look at a video picture on a 13 inch field monitor or a 60 inch plasma display, the number of pixels across by the number of pixels high does not change. It's exactly the same.
All we're doing is we're making the pixels fatter. On a computer if I've got a 13 inch monitor or I've got a 30 inch monitor -- I dream of 30 inch monitors actually. It's sad. Anyway, if I put it on a 30-inch monitor I'm going to see more. But not true with a television set. Scanning by the way, most video, though not all is interlaced. That is to say the lines are woven together and there is a slight difference in time between all the odd numbered lines and all the even numbered lines. With computer and film they are progressive. All the lines are shot at the same time. There is none of this offset for time.
In video your white levels must be equal to or less than 100%. But with the computer and when you're shooting in a digital camera world, your white levels can be up to a 110%, which means that we've got to tweak our White levels to get them down into something which is broadcast safe. Most video, I mean the extreme high-end where you're spending $200,000 for a camera, that changes, but for most videos that most of us are working with we're shooting in 8 or 10 bit depth, and the computer is a 12 bit depth.
The higher the bit depth, the more accurately it can represent a color image. Black and white looks find at 8 bits, but color to look good requires a minimum of 10 and ideally 12. The problem is video only goes 8 or 10 bit. Again the higher the bit depth the greater the color fidelity. Another is gamma setting or mid-tone gray. Depending upon your computer system, your gamma setting can be anywhere from 1.8 to 2.2, and the different gamma setting means that the mid-tone gray shifts which means the exposure shifts.
This is why if you create an image on say Photoshop on the Mac and you pull it into video it's going to look light and thin and washed up. You take a still frame on video, you pull it into Photoshop, it looks dark and muddy and uninteresting. That's because Photoshop on the Mac and video images have different mid-tone grays and stuff doesn't look the same and we need to compensate for that difference. Video uses a color space in digital video called YCbCr. Apple uses the acronym YUV, which is an analog standard and I'll probably use YUV, even though I should say YCbCr, but YUV is just so much easier.
The computer uses a color space called RGB. They both give us colors. The problem is they're not the same. I have a greater range of colors in RGB than I have in YCbCr. I can create colors on my computer that I can't broadcast or cablecast or move to a DVD. Though I could post it to the web because the web is played by computers, which means that we have to work within a limited color space of YCbCr when we're working with creating images inside Photoshop that have to go onto video. Then there's color sampling. Now a discussion of color sampling has been known to cause strong people's eyes to just glaze over. They tip over, fall into a faint and they don't recover for a week.
So we won't go there. All I want to say is computer color sampling is much more accurate. Every pixel has its own color. When we work with video, pixels are grouped. Instead of having each pixel with its own color groups of pixels, a group of two, a group of four, a group of two rows have the same color value. Black and white value is always accurate, but color value? That starts to get a bit more murky, shall we say. And then finally the pixel aspect ratio. In video pixels are rectangular, in the computer they are square.
Well, this alone would not be too hard to deal with except for the fact that different video formats have different shaped rectangles. Some are short and fat, some are tall and thin. So we've got a flock of different image sizes that we've got to cope with, and that's with this next chart comes in. This screen is probably the most important screen of any one we're going to show in this entire title. It shows the standard-def image sizes that you need to make your images look good in all the different video formats we have inside standard-def. The problem is those stupid rectangular pixels.
For instance, if we're working with DV NTSC creating a 4:3 image and you just want the image to be full-screen with no black around the edges, circles are circles, squares are squares, and all your logos look like you're going to stay employed for more than a week. A full-screen image needs to be 720x540x72 pixels. Same thing for 16:9 image. 853x480x72, and the rest of these numbers I won't read. You can just write them down, but write them down because you're going to find them to be very, very helpful as you start to prepare your still images.
What about this other column here, 2.5x movement? So you remember -- it seems like a lifetime ago -- I said that an image only looks good at 100% or smaller. What happens if you want to pan around inside the image? Well, panning around inside the image means that the image can't be bigger than 100%, which means we have to make the image bigger than full-screen. So that's what that column is. If you want to zoom into an image and then pan around from the lower left corner to the upper right corner then make the image 1800 pixels by 1350 pixels by 72 pixels, and now you can move around inside the image and explore it.
So what this table of numbers gives you is the center column is full-screen, no movement, and the right-hand column gives you full-screen and movement, the ability to move around pan and zoom in and zoom out. Remember keep your image at 100% size or less. By the way Apple uses a different calculation for DVDs than it does for Final Cut. Final Cut uses one set of calculations and DVD Studio Pro uses the second. So these bottom two lines are, if you're creating full-screen slides, full-screen backgrounds for a DVD, you need to create it at a different image size than if you were creating a full-screen image for Final Cut or Motion. Don't ask me why, it just is.
Let's take a look at what high-def looks like. High-def, thank goodness, is easier. You have only two choices. If you're creating images for a high-def 720p, whether it's DVCPRO HD or whether it's XDCAM HD or XDCAM EX or whatever it is, a full-screen 720 image is 1280x720x72 and a full-screen 1080 image is 1920x1080x72. Just drop that image into your Timeline, Final Cut will automatically size it appropriately. Again, if you want to move around it needs to be 3200x18 for 720 image or 4800x2700 for a 1080i image.
Your images should be sized to match the size of your sequence. Don't worry about what the size of the video is, remember video is working with rectangular pixels, you're working with square pixels. Also you want to avoid using images that are much larger than about 5,000 pixels when you're working in Final Cut. Final Cut likes your images to be along the smaller side. Just a couple of more cautionary notes and then we're done with all this initial getting ready stuff. To avoid moire or interlace flickering you want to make sure that your images don't have thin lines and avoid lines that are almost vertical or almost horizontal.
Avoid very fine detail like interwoven patterns or leaves in the background. They all tend to set up flickering or moire. Avoid extremely saturated colors. You want to check them using the vector scope inside Final Cut. There is no vector scope inside Motion, but there is inside Final Cut. Use it to make sure that your colors are broadcast safe. And it's not if you're going to broadcast, it's whether you're burning to a DVD. If it's going to be played on a TV set your colors have to be broadcast safe because over-saturated colors cause your TV to cause problems.
We don't need this level of abuse in our life. You want to make sure that your colors don't get over-saturated and you want to make sure that the color saturation is checked for the video format that you're using. I mentioned this earlier, but I want to stress it again. Use higher-quality image formats like TIFF or PNG or PSD. JPEG images tend to have lower image quality, they tend to have lines, thin lines that start to break up and we've got artifacting, blocky colors, doesn't look as good, and please, oh please! Keep your image size to 100% or smaller because your images are always bitmapped, and once they get bigger than 100% all we're doing is making fat ugly bitmaps and we don't need that.
Boy! It's a lot to cover in a very short period of time, but with that as our background, now the fun stuff starts. We get to start Photoshop and we get to start moving our images around and making them look great, and then we get to animate them. All of that is coming up next.
- Preparing images in Photoshop
- Building an image sequence and synching to a music track
- Adding handles and transitions to images
- Adjusting scale, rotation, cropping, distortion, and other motion parameters in Final Cut Pro 6
- Understanding effective use of keyframes and Bezier controls
- Sending a project back and forth between Final Cut Pro and Motion
- Adding effects with behaviors or keyframes in Motion