Dots and Lines

Dots and Lines

Why 2048 (More or Less) Equals 1080 in Digital Video

When industries converge, there is often a collision of terminologies. Different traditions have their own ways of describing the same thing. In the latter half of the 1980s, personal computing converged with the typesetting industry. While the latter talked about points and picas, the former saw everything in terms of pixels.

Now, computing is converging with entertainment media and, again, we find a heritage of looking at the same thing in different ways.

Consider the resolution of digital video.

Anyone who has recently purchased a Blu-ray player or a large-screen television is no doubt familiar with the numbers used to characterize high-definition television — 720p, 1080i, and 1080p — even if the exact meaning of these figures may remain elusive. And most people know that, for a given resolution, ‘p’ (progressive scan) is better better than ‘i’ (interlaced).

What may be less clear is: 720 or 1080 what? What do these numbers measure? As many of you — the technically savvy readers of this blog — are no doubt aware, these figures refer to the number of horizontal scan lines in the image — in other words, its vertical resolution.

If you’ve been following developments in digital cinema, you also know that movie theaters may use what is termed a “1K” or a “2K” digital projector. My recent post on the IMAX Conundrum, for example, mentioned that IMAX Digital uses dual 2K Christie projectors. But what is the “2K” (that is, 2,048) measuring? 2,048 what? Is it the same vertical resolution described above?

It is not. In this case, the number indicates the number of pixels in each scan line — the image’s horizontal resolution.

Why are consumer displays indicated by their vertical resolution and professional projectors by their horizontal resolution?

Television’s focus on vertical resolution may come from the legacy of traditional TV signals which, although analog, contain a fixed number of horizontal scan lines: 525 for NTSC and 625 for PAL. The new digital formats continue this tradition of thinking of image quality in terms of the number of scan lines — the vertical resolution. Why cinema projectors took the opposite tack of specifying the horizontal resolution is anyone’s guess.

The upshot is that a projector with a resolution of, say, 2048 pixels horizontally and 1080 pixels vertically could be described as either 2K (‘2048’) or 1080p.

Differences in aspect ratios add to the confusion. A 1080p digital television may have a resolution of 1920 by 1080 pixels, while a 2K cinema projector may be 2048 by 1080. Sony has a 4K projector with a native resolution of 4,096 by 2,160 pixels. The RED ONE digital camera is also termed a 4K device, even though it has a native resolution of 4520 x 2540 pixels.

Computer monitors and video display cards have long avoided this confusion by simply indicating both the horizontal and the vertical resolution. Thus early VGA displays were characterized as “640 x 480,” that is, a horizontal resolution of 640 pixels and vertical resolution of 480 pixels. An XGA display is 1024 by 768. And so on.

The computer geeks have it right in this case. The convenience of having a single number to indicate resolution — 2K, 1080p — is outweighed by the clarity of giving both dimensions. Let’s avoid the confusion and state video resolution as two numbers — the horizontal by the vertical resolution.

7 thoughts on “Dots and Lines

  1. A great example of how the fixed-pixel world clashes with the scanline world was with (mostly sub-50″) 720p Plasma displays having an oddball resolution of 1024×768.

    So, while it met the scanline requirements for 720, the it didn’t seem to meet the 16:9 aspect ration requirements; 1024×768 is a 4:3 aspect ration. But, those plasmas were being billed as HD! What gives?

    Turns out, the vendors used *rectangular* as opposed to square pixels to get the desired aspect ratio. Try explaining that to an average consumer!

    I agree, especially since we have moved entirely to fixed-pixel displays, that we should be getting the pixel numbers right.

    A strange side-effect: I bought a new laptop for my brother, and it had an HD-display @ 1920×1080. The “normal” wide-screen display aspect ratio for a computer is 16:10 (so 1920×1280), so I couldn’t help but feel that I was getting *less* for my money than before.

    Like

  2. The per pixel cost you bring up is a good one, Joe. When rendering 3-D content into video for television, you have to take into account the 16:10 versus 16:9 during the capture process. Fortunately, most 3-D rendering software has settings for this, so you can choose to display them through a HD television, a projector, or on your monitor without too much hassle.

    Like

  3. Interesting post, totally agree about the silly naming conventions.

    What hooked me and got me to read the article not just once but twice was the image. What are those white dots? Please do explain!

    Like

  4. Jin:

    The short story is that it’s just an abstract image. The dots and lines aren’t intended to represent anything beyond the general concept of thinking about video as either a series of scan lines or an array of dots.

    As I was working on the post, I had a picture in my head of what sort of image I wanted, but I wasn’t sure how to create it. Rather than drawing everything by hand — which would have been extremely tedious — I concocted a simpler way to produce the image I wanted.

    What you see here started as an image — actually a set of images — created by Social Collider. (See “Social Collider: Visualizing Twitter” for more on this.)

    To create the scan lines, I took several different Social Collider images, dropped out the black background, and then used Photoshop’s “motion blur” filter to convert the Social Collider swirls into scan lines. By layering several images created like this, I got the density of lines I was looking for. Then I took one more Social Collider image and dropped out everything except the white dots and layered that on top of the lines.

    So, all the lines and dots you see here were originally generated by Social Collider graphs that were subsequently modified and merged together.

    Like

  5. Ah! I knew that image looked familiar in some way! I was racking my brain but I thought it was some image I saw having to do with tv or something so I could not place it.

    Have you seen this site? http://flowingdata.com/ You seem like you might enjoy it.

    Like

  6. From my perspective much of the confusion and a lot of the problems comes from the fact that thee word “standard” is now plural.

    Digital is wonderful, but having many “standards” is not.

    J

    Like

  7. Jay:
    Your comment brings to mind the ironic quip, “The nice thing about standards is there are so many to choose from!” And then there’s the dialog between Ian (Joe Mantegna), Pip (Adam Sandler), and Chazz (Brendan Fraser) in the movie Airheads:

    Ian: “The Lone Rangers”? That’s original. How can you pluralize the “Lone Ranger”?
    Pip: What’s wrong with that?
    Ian: Well, there are three of you. You’re not exactly lone. Shouldn’t you be the Three Rangers?
    Pip: No idea what you’re saying right now.
    Chazz: You lost me…

    So, yes, the concept of “multiple standards” is, in a sense, an oxymoron. But it’s nonetheless a common situation. In some cases there are benefits from having alternate approaches to a given problem, in others it just creates confusion. I think the various ways of describing digital video fall squarely into the latter camp (and, hence, the blog post).

    Like

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.