Comparison of progressive and interlace scanning

The advantages of progressive scan are:

• Higher vertical resolution than interlaced video with the same frame rate. The perceived vertical resolution of displayed video is traditionally adjusted using a Kell factor coefficient. This coefficient has no fixed value and depends on display device. Its value for interlaced video is usually lower than for progressive video, when the same display device is used. When interlaced video is compared to progressive video with the same number of scan lines, interlaced video delivers lower perceived vertical resolution at a lower frame rate.
• Absence of visual artifacts associated with interlaced video of the same line rate, such as interline twitter.
• No necessity in intentional blurring (sometimes referred to as anti-aliasing) of video to reduce interline twitter and eye strain. In the case of most media such as DVD movies and video games, the video is blurred during the authoring process itself to mask flicker artifacts when used on interlace displays. As a consequence, recovering the sharpness of the original video is impossible when the video is viewed progressively. An excellent, but rarely employed countermeasure to this is when display hardware and video games come equipped with options to blur the video at will, or to keep it at its original sharpness. This allows the viewer to achieve the desired image sharpness with both interlaced and progressive displays. An example of a video game with such a feature is Super Smash Bros. Melee, where a “Deflicker” option exists. Ideally it would be turned on when played on an interlaced display to reduce interline twitter, and off when played on a progressive display for maximum image clarity.
• Offers much better results for scaling to higher resolutions than equivalent interlaced video, such as up converting 480p to display on a 1080p HDTV. Scaling works well with full frames, therefore interlaced video must be deinterlaced before it is scaled. Deinterlacing can result in severe “combing” artifacts.
• Frames have no interlaced artifacts and can be used as still photos.

However, the only disadvantage of progressive scan is that it requires higher bandwidth than interlaced video that has the same frame size and vertical refresh rate.

Example of interlace scanning field

Interlaced scan refers to one of two common methods for “painting” a video image on an electronic display screen by scanning or displaying each line or row of pixels. This technique uses two fields to create a frame. One field contains all the odd lines in the image, the other contains all the even lines of the image. A PAL based television display, for example, scans 50 fields every second (25 odd and 25 even). The two sets of 25 fields work together to create a full frame every 1/25th of a second, resulting in a display of 25 frames per second.

The interlaced scan pattern in a CRT (cathode ray tube) display completes such a scan too, but only for every second line. This is carried out from the top left corner to the bottom right corner of a CRT display. This process is repeated again, only this time starting at the second row, in order to fill in those particular gaps left behind while performing the first progressive scan on alternate rows only.

Such scan of every second line is called interlacing. A field is an image that contains only half of the lines needed to make a complete picture. The afterglow of the phosphor of CRTs, in combination with the persistence of vision results in two fields being perceived as a continuous image which allows the viewing of full horizontal detail with half the bandwidth that would be required for a full progressive scan while maintaining the necessary CRT refresh rate to prevent flicker.

Interlace is a technique developed for improving the picture quality of a video signal primarily on CRT devices without consuming extra bandwidth (bandwidth is a cost, higher bandwidth require more data storage). Interlacing causes problems on certain display devices such as LCD TV. It was invented by RCA (Radio Corporation of America) engineer Randall C. Ballard in 1932, and first demonstrated in 1934, as cathode ray tube screens became brighter, increasing the level of flicker caused by progressive (sequential) scanning. It was ubiquitous in television until the 1970s, when the needs of computer monitors resulted in the reintroduction of progressive scan. Interlace is still used for most standard definition TVs, and the 1080i HDTV broadcast standard, but not for LCD, micromirror (DLP), or plasma displays; these displays do not use a raster scan to create an image, and so cannot benefit from interlacing: in practice, they have to be driven with a progressive scan signal. The deinterlacing circuitry to get progressive scan from a normal interlaced broadcast television signal can add to the cost of a television set using such displays. Currently, progressive displays dominate the HDTV market. Only CRTs can display interlaced video directly – other display technologies require some form of deinterlacing.

Broadly speaking, Progressive video will look better on an LCD TV because these panels are progressive in nature. Any Interlaced content is converted on the fly to Progressive (it ‘fakes’ a progressive picture, which never looks as good as a real progressive one).

1080p is the shorthand name for 1,080 lines of vertical resolution progressive scanning signal (1080 horizontal scan lines). The letter p acronym for progressive scan. 1080p can be referred to as Full HD (Full High Definition) to differentiate it from other HDTV video modes (example, HD Ready LCD TV). The term usually assumes a widescreen aspect ratio of 16:9, implying a horizontal resolution of 1920 pixels. This creates a frame resolution of 1920×1080, or 2,073,600 pixels in total. The frame rate in Hertz can be either implied by the context or specified after the letter p, such as 1080p30, meaning 30 Hz.

Common Video Resolutions (1080 / 720 / 480)

1080p is sometimes referred to in marketing materials as “Complete High-Definition”. However, 2K/4K digital cinema technology is commercially available, and ultra-high definition video is in the research phase.

In addition to the meaning of 1080p as a display resolution, 1080p is also used to describe video equipment capabilities. Use of 1080p and the closely related 1080i labels in consumer products may refer to a range of capabilities. For example, video equipment that up-scales to 1080p takes lower resolution material and reformats it for a higher resolution display. The image that results is different from the display of original 1080p source material on a native 1080p capable display. Similarly, equipment capable of displaying both 720p and 1080i may in fact not have the capability to display 1080p or 1080i material at full resolution. It is common for this material to be downscaled to the native capability of the equipment. The term “native 1080p capable” is sometimes used to refer to equipment capable of rendering 1080p fully.

On August 25, Dish Network announced it would begin transmitting high definition TV programming using the MPEG4 Advanced Video Codec (AVC) in 21 markets. By itself, that’s not much of a news story.

Dish Network HD TV Broadcast

What did make the headlines was the second part of the announcement. Dish now plans to deliver HD video-on-demand (VOD) programming in the 1920×1080 progressive scan HDTV format. The Dish press release promises (in their own words “…the availability of movies in Blu-Ray Disc quality 1080p resolution” but goes on to state later that their TurboHD program service offers “…the highest quality HD available, including 1080p where applicable.”(Doggone it, there’s always a catch!)

Aside from that, you’re probably wondering: Can Dish really deliver 1080p video to my home? If so, then why not Comcast, or Time Warner? Why not CBS and NBC?

The answers aren’t simple, but yes, Dish really can deliver 1080p. So can DirecTV, and every cable company currently transmitting HD. For that matter, so can any TV station currently transmitting over-the-air HD in the 1080i format.

The really question is; would the HD picture quality be acceptable to you? Maybe; maybe not. And would it be as good as Blu-ray? That depends. How much “time” do you have to spare?

IT’S ALL IN THE NUMBERS

To get a better handle on the viability of 1080p over satellite, cable, and terrestrial links, we need to stop thinking about “P’s” and I’s” and concentrate on bit rates instead. Currently, TV stations transmit digital video – standard definition and high definition — using the MPEG2 compression/decompression (codec) standard. So do cable companies and DirecTV (and to date, so has Dish).

MPEG2 is getting pretty long in the tooth. It was originally developed for use almost 20 years ago with standard definition DVDs, but works fine with its many profiles and levels for 720p and 1080i HDTV, too. You just need a high enough bit rate to preserve image quality.

For satellite HD feeds from networks to local affiliates and cable head ends, that would be nominally 40 megabits per second (40 Mb/s). And of course, the maximum bit rate that can be used for over-the-air broadcasts using the 8VSB modulation standard is 19.39 Mb/s, although stations usually cap that at a maximum rate around 18 Mb/s.

Cable companies have more real estate with their 256QAM modulation system. It yields a maximum bit rate of 38.8 Mb/s, although that’s usually divided between two HD programs, or multiplexed across 10 to 12 SD programs and one HD program.

DirecTV and Dish have similar bit rates but use yet another modulation scheme, known as quadrature phase-shift keying (QPSK). So all three delivery systems can pump 720p and 1080i HD signals to your home with reasonable quality.

Notice I said, “can.” Think of a TV channel or satellite transponder as a digital shoebox. You can fill it up any way you like, using just one HD channel, a bunch of SD channels, or a mix of HD and SD channels. The only rule is; you just can’t have anything spilling outside of the digital shoebox.

The problem with digital compression is that it can get out of hand. Just because an HD program can be packed down 100:1 doesn’t mean it’s a good idea. For reference, consider that a 720p/60 HD program transmitted to your HDTV at 18 Mb/s has been compressed by a ratio of 49:1 (uncompressed 720p/59.94, using 4:2:0 encoding, has an uncompressed data rate of 885 Mb/s.)

If you think that’s a lot, consider that a 1080i/29.97 signal is packed down by 55:1 when transmitted at 18 Mb/s (its uncompressed bit rate is 995 Mb/s). Theoretically, MPEG is a “lossless” compression system, and it does work very well.

But fast motion really strains the limits of 1080i and 720p. If you have a big screen HDTV, particularly one with 1080p resolution, you’ve no doubt noticed weird compression artifacts from time to time, particularly during the recent Beijing Olympics.

Did it look like clouds of bugs were buzzing around the USA basketball team on a fast break? Did you see strange, square-shaped objects flying around with water droplets during the swimming competitions? Those are MPEG compression artifacts commonly known as “mosquito noise” and “tiling.” And they’re present in any digital images with lots of motion and compression.

THE THREE Bs

We all know that HDTV is a business, and that when it comes to the three “Bs” – bit rate, bandwidth, and bucks — that the bucks always win out. What that means is that a given HD content provider will generally look to maximize channel space (bandwidth) by stuffing as many HD signals into that channel as they think they can get away with.

Sharp 37-inch HDTV

The gray area is in that last sentence — “get away with.” You aren’t as likely to see compression artifacts on a 34-inch CRT HDTV or a 37-inch LCD, as you will on a 50-inch plasma set or a 61-inch DLP rear projection TV. And there are a whole lotta big screen HDTVs sitting in homes now with some critical eyes watching them.

Internet forums are full of commentaries on HD picture quality from TV stations, cable, and DBS. In general, it seems to be the satellite companies that get hit hardest for poor HD image quality, and in fact DirecTV was sued a few years back for allegedly delivering HD images with only 1280×1080 picture resolution.

But broadcasters and cable systems aren’t immune, either. While CBS generally gets kudos for refusing to multicast their HD programs with other content on their owned-and-operated (O&O) stations, other O&Os and network affiliates see nothing wrong in multicasting two and even three SD programs along with their main HD service, packing that 720p or 1080i channel down as low as 10 Mb/s. Yecchh!

And cable MSOs have to be careful how they bundle up SD and HD channels in a 256QAM stream, so as not to impair HD video quality and lose any marketing advantage against their age-old nemesis, direct broadcast satellite.

In the satellite biz, the race is on big time for DirecTV and Dish to offer the most HD channels and gain a competitive advantage in the market, not only against each other, but against the likes of Comcast and Time Warner, too. And satellite transponders (channels) are pretty darned expensive! Think of rockets, launching satellite into specific earth orbits, etc.)

Even a reasonable person will only try compressing HD so far before the signal has been changed from chicken salad into chicken s—t. But transponder bandwidth is fixed, and expensive. What’s the answer?

A BETTER MOUSETRAP

A new codec, of course. MPEG4 is a more advanced way to compress digital video, not only compressing from frame to frame, but on the sub-pixel level. As such, it promises greater compression efficiency with comparable picture quality.

So, how efficient is MPEG4 in its Advanced Video Codec (AVC) implementation? It depends on the content and the refresh rate of the electronic images. Perhaps for one type of program, it’s 25% more efficient. For another type, it’s 50%. (The less motion from frame to frame, the better, when it comes to compression.)

That would mean that, in theory, a 1080i/29.97 HD program that looked acceptable at 18 Mb/s using MPEG2 should present with comparable image quality at 9 Mb/s using MPEG4. And a 720p/59.94 HD program running at 16 Mb/s in MPEG 2 should only require 8 Mb/s with MPEG4.

Alas, that also means that a not-so-spiffy 1080i broadcast that is suffering at 13–14 Mb/s in MPEG2 will experience as much degradation at 6.5–7 Mb/s in MPEG4. The sad truth is; live 1080i programming, particularly sports, doesn’t hold up well at such low bit rates, no matter what anyone tells you. (Are you aware that AT&T’s U-Verse system carries 1080i HD programming over DSL lines at just 6 Mb/s, using MPEG4?)

DOUBLE THE FUN

You can see that MPEG2 is already straining at the bit to deliver quality 1080i content at existing home-delivery bit rates. Simple math will tell you that switching to a progressive scanning system would mean double the number of pixels traveling in the same time interval! So, if 18 Mb/s were considered a sufficient bit rate for decent HD quality on a 1080i/29.97 program with MPEG2 coding, a 1080p/59.94 version would need 36 Mb/s! (Incidentally, that’s the maximum bit rate supported in the Blu-ray specification. V-e-r-y interesting!)

Switching from MPEG2 to MPEG4 in theory should cut that bit rate in half again, to 18 Mb/s. But that’s still too much data for a Dish channel, and would make adding more HD channels problematic. A more realistic bit rate target might be half again — 9 Mb/s, which would be make for pretty crappy picture quality with a 1080i program using MPEG2.

OK, 1080p/24 playback from Blu-ray is all the rage now. How’s about we drop the bit rate to 24 frames per second for our Dish broadcast, instead of nearly 60 fps?

That’s a reduction in the pixel payload of 60%. If we agree 18 Mb/s is the minimum for 1080p/59.94 using MPEG4, then that means we can probably get away with about 7 to 8 Mb/s for 1080p/24. (One characteristic of shooting and encoding HD content into a 24-frame format is that motion blur becomes inherent, and not the result of artifacts caused by over-compression or LCD blurring.)

Is 7 Mb/s 1080p HD Blu-ray quality? Not even close. A typical Blu-ray feature, encoded into MPEG4, will stream data around 14 to 16 Mb/s using constant bit rate encoding. 6 to 7 Mb/s may be fast enough to transport 1080p HD content in the MPEG4 format, but it may not look very good when you actually watch it. (Just because something’s possible doesn’t mean it’s always a good idea!)

TIME IN A BOTTLE

There is one way to get around the low bit rate conundrum: Spread the bits out over time. That’s how broadband pay-per-view and VOD services like VuDu do it. As long as you don’t have to stream the 1080p HD program in real time, you can cut way back on the bit rate and still preserve image quality.

VuDu gets by with a minimum bit rate of 4 Mb/s, but that’s because the first 30 seconds or so of every movie is already stored on its hard drive. When you start watching, it’s got a half-minute head start to frantically write the rest of the data to the HDD so you can enjoy The Bourne Supremacy uninterrupted. And it works!

The problem is, that model won’t work at all for real-time terrestrial broadcast HDTV. However, it might work for DBS and cable services, if their set-top boxes contain large hard drives that can use the same buffering trick as VuDu: Store the first 30 seconds or minute of a movie or TV show on the HDD, then start streaming the rest as you watch.

To sum up, if an HD content delivery system can tolerate some degree of latency, then lower bit rates for 1080p content are practical. But if you are surfing channels and want top watch programs that aren’t already buffered, you’ll either have to put up with some latency or suffer with over-compressed images.

The jury’s still out on what bit rate Dish will use for their 1080p/24 TurboHD service, and how they plan to address the image quality vs. bit rate vs. latency puzzle. (You can’t put 10 gallons of water into a 5-gallon bucket without spilling it!)

Something’s gotta give along the way, and the “3 Bs” economic model tells us it’s usually the bit rate that suffers while the bucks are always maximized. Only time will tell…

Article by Peter Putman