[ This article was first published in the June, 2010, issue of
Larry’s Monthly Final Cut Studio Newsletter. Click here to subscribe. ]
I received several emails this week from people concerned that their videos had suddenly developed a serious case of the “jaggies.”
They hadn’t. What they were seeing was interlacing.
There are two ways an image can be captured: progressive and interlaced. Which version you capture depends upon your camera and the video format you are using.
The original way video images were captured was interlaced. All NTSC, all PAL, and all 1080i HD formats are interlaced. They record an image in two groups; one group is all the pixels in the even-numbered lines, the second group is all the pixels in the odd-numbered lines.
In the case of NTSC, the even-numbered line group is recorded first. Then, 1/60th of a second later, the odd-numbered line group is recorded. (HD formats generally record the odd-numbered lines first.) This means that anything in your image that’s moving has half its pixels time-shifted by 1/60th of a second.
NOTE: If you work with PAL, the process is the same, but the time-shift is 1/50th of a second.
Interlacing had its genesis at the very beginning of television as it solved technical problems related to the display of images on a phosphor-coated TV screen, as well as minimizing bandwidth during transmission.
The entire television industry was founded on interlaced images which worked great, as long as you were viewing them on a TV set.
The second way video images are captured is progressive. Progressive images are indicated by the letter “p” in the format name: 720p, for example. Here, the camera captures all the image pixels at one time. Film has always worked this way, but this process is relatively new to video.
While interlacing is traditional, it totally breaks down when you view images on a computer, because the computer always displays images progressively.
If you are shooting video to burn to a standard-def DVD, interlacing is fine. Its part of the format. It may look a bit strange in Final Cut, but it will look fine on your DVD.
If you are creating video for the web, interlacing will drive you nuts.
Ideally, you should shoot progressive video – because progressive footage guarantees no interlacing. It is far easier to convert a progressive image to interlaced than an interlaced image to progressive. We’ve been doing this successfully with film for decades.
The ideal format to shoot in is 1080p. However, that gear tends to be expensive. For me, I do all my work as 720p. Since everything I do goes to the web, 720p gives me all the quality I need using affordable cameras. Even high-def websites only display 720p video.
But, what if you need to go to the web and all you have is interlaced footage? Well, the answer is easy, but not ideal — you need to deinterlace.
Here’s my suggestion: never deinterlace in Final Cut Pro. Only deinterlace when you are compressing your final image for the web.
NOTE: Well, OK. You can deinterlace in Final Cut Pro if you have a freeze frame, or a still image that is vibrating badly. What I mean is that you should not deinterlace your entire program in Final Cut Pro. First, the render times are LOOOONNNNG, and second, the quality is not the best.
Edit the format you shot in Final Cut Pro and export it in that format using File > Export > QuickTime Movie. This gives you the highest-possible quality throughout the entire editing process, including export.
Here’s why. In the image above, the red lines represent the time-shifted pixels which we need to remove to eliminate interlacing.
However, if we remove half the lines, we automatically cut our vertical resolution in half — because half the lines are now missing. For the web, where we traditionally reduce the size of our images, this may not be bad. But for projection, it creates an automatic drop in image quality.
This is why it is useful to think about where your finished video will be viewed before you pick the camera and video format to shoot.
Before you deinterlace your image – and remember, don’t deinterlace inside Final Cut Pro – think about what your final image size will be.
Since deinterlacing removes 1/2 the lines in your image, if your final image size is smaller than 1/2 the original, you don’t need to deinterlace at all – it will happen automatically.
In most cases, deinterlacing is handled automatically by your compression software. However, deinterlacing will be required for NTSC or PAL images displayed at full screen size: 640 x 480. The best place to deinterlace is in your compression software.
If you want to know how to do your own deinterlacing, here’s a video tutorial I recorded on how to deinterlace inside Compressor.
Steve Schafer asks:
I was at a Gnomon event and asked them what is the format I should be shooting with my HVX200 for FX later in FCP or Motion or after Effects. The told me 1080i was better than 720P. This seemed odd to me as interlaced seems inferior. I was later at an Creatasphere event and there was a well know special effect house there (located on sunset in hollywood) and I asked the same question they also told me 1080i was preferred to 720P. They claimed there was more information from which to do the effect. You stated in your presentation that Progressive was preferred and so now I am still left baffled as to what I should be shooting to my HVX200 at for the best resolution and look in FCP, Motion and After Effects.
Larry replies: Steve, there is more horizontal information in a 1080i image, but more vertical information in 720p due to the problems caused by interlacing.
The ideal format to shoot is 1080p, as that can be converted to anything. However, I would disagree with their opinions that 1080i is better than 720p. Especially with an HVX200, which stretches a horizontal resolution of 1280 pixels out to 1920. In other words, the horizontal resolution of the HVX-200 is the same whether you are shooting 1080 or 720!
NEW & Updated!
Edit smarter with Larry’s latest training, all available in our store.