Infobase: About High Definition Video (Part 1)


High definition is more than just a larger flat screen – it’s a complete change in the way we record, edit and share our video clips using HD camcorders. But what, exactly, is this thing called High Definition and how does it influence the way we shoot our movie projects?


It seems only a short while ago that we regarded high definition as being the domain of the professional camera operator, yet thanks to the proliferation of so-called “HD Ready” Plasma and LCD television displays in the shops the whole market has changed to the point where people are rapidly turning their backs on standard definition (SD) in favour of the largest HD displays they can lay their hands on. But proper HD is more than just a TV or DVD picture being displayed on an impressively large flat screen.

What influences people’s buying choices for TV displays has a natural knock-on effect for camcorders, too; somehow, all those old analogue and even digital video clips don’t look so good on that super new TV screen – even if they’ve been recorded digitally and in widescreen. Today’s camcorder user wants High Definition, and who can blame them? But what, exactly, is High Definition – and how is this likely to influence the type of camcorder we buy?

Video by definition


When people see “HD Ready” stickers on those nice big flat displays in the shops do they actually know what it all means? In order to see just how radically different is HD or SD (standard definition) let’s first analyse the two. The TV system we’ve grown up with since the medium’s very beginnings is regarded as being SD. It’s only since the emergence of HD that it’s had that moniker, but we now regard SD as being the traditional format TV picture – which has a picture aspect ratio of 4:3 (horizontal:vertical).

In most of Europe, our TV system uses a colour system call PAL (Phase Alternating Line) which describes TV picture sequences as having 25 frames per second, each made up of 625 lines. However, each picture is scanned in two passes – odd and even lines separately - thus involving 50 scans per second (50 cycles per second = 50 Hz). Of the 625 scanning lines, only 576 lines are of any use (the rest being “redundant”), and when we’re handling SD video in the digital domain the picture sequence is converted into pixels measuring 720 x 576 pixels per image.

In the USA, Japan and other parts of the world where the NTSC (rather than PAL) system is used, a standard TV picture is scanned by 525 interlaced lines – 525i rather than PAL’s 625i. Taking into account redundancy, the standard definition 4:3 NTSC digital video frame is constructed using 720 x 480 pixels.

Defining High Definition


High definition is a completely different beast to SD for two reasons. The first is that it is based upon a specification that employs a far greater number of scanning lines, and secondly because it’s a much wider picture by default. The full specification HD image is defined as 1920 horizontal pixels by 1080 vertical pixels (scanning lines). However, there are two types of HD – one of which uses a lesser 720 vertical lines. So which is best?

Well, it depends on how the picture is constructed. When the benchmark specification for HD was laid down, the vast majority of us were still using CRT (Cathode Ray Tube) TV sets which relied upon interlaced frame scanning. However, the increased use of Plasma and now LCD flat-screen displays is such that interlaced sequences can look jittery and unpleasant when viewed on flat displays, and tests have show that it’s better to combine each of the two alternate “fields” into one and a single solid frame in one pass. We call this a “progressive” frame (as in 720p, 1080p, etc).

The majority of current high definition camcorders shoot interlaced frames made up of 1080 lines – denoted as 1080i. Of these, an increasing number are employing the full 1920 horizontal pixels, though many use a cheat by scanning 1440 pixels (as do many professional TV cameras) and then scaling up using what’s known as the Pixel-Shift Technique (see “Shifty Pixels”). Whilst lots of consumers are starting to demand 1080p camcorders, it’s possible that audiences will prefer 720p sequences if recent research in Austria is anything to go by; according to tests by Hans Hoffman – Senior Engineer with the European Broadcasting Union – 720p HD sequences looked better than 1080i/50Hz and even 1080p/50Hz at all levels of compression, so perhaps the holy grail that is 1080p is superfluous.

Shifty Pixels


Some recent HD camcorders, such as JVC’s GZ-HD7 hard-disc recording HD camcorder and Panasonic’s latest HDC-SD5 and SX5 AVCHD models, achieve the full 1920 x 1080i specification by a nifty little technique called “Pixel Shift”. Whilst it can be considered to be a bit of a cheat, it’s not new – having been around for 20 years or so, and which has been employed by Canon on its 3CCD XL, XM and XH range models. But what is it?

Pixel shift is a technique that was developed by Panasonic and is used also by Canon, Sony and JVC to effectively increase the luminance (brightness / saturation) and overall resolution of the images produced by a camcorder’s CCDs that originate a lesser number of pixels than required by high definition. What it effectively does is to shift the green CCD’s output diagonally upwards (or sometimes just sideways) by one half a pixel; this – combined by some other complex maths – effectively increases to 1920 x 1080 the true CCD output of, for instance, Panasonic’s HDC-SD5 from 1440 x 1080i. This improvement in luminance and resolution is most noticeable in images that don’t contain fully saturated colour, such as red. Even some camcorders that record HD to tape (eg: HDV) rely on Pixel Shift despite the fact that the image is then reduced to 1440 horizontal pixels prior to recording.

Relative sizes - How Standard Definition compares to High Definition


It's only when you compare like-for-like video frames that you begin to appreciate the substantial differences between standard definition and high definition television images. Even standard definition widescreen images look small when compared at exactly the same scale on a pixel-for-pixel basis as their true high definition counterparts. Here’s a standard definition picture shot in 4:3 aspect ratio and conforming to the PAL specification of 720 x 576 pixels:

720x576_sd

The wider 16:9 aspect ratio afforded by some SD cameras can produce 1024 x 576 pixel images (usually by squeezing the optical image before it hits the CCD), but at least it provides us with increased creative potential where composition is concerned:

1024x576_sd

It’s when you consider the true HD image at 1920 x 1080 pixels that you begin to see how advantageous it is to shoot in HD. The result is greatly improved picture clarity on similar-sized HDTV capable displays:

1920x1080_hd

Please note that for ease of presentation on this page, the actual-size images (as denoted in pixels in each case) have been reduced by a common factor of 35%. Please be aware that the SD picture measurements relate to images produced using PAL equipment. In areas where NTSC is used (eg: USA, Canada, Japan, etc) the vertical resolution will be 480 pixels rather than the 576 pixels referred to above.

In the second part of this feature on HD we consider the practical considerations of choosing your first HD camcorder and the means of then connecting it to a suitable television display.

See: High Definition Part 2


Words: Colin Barrett, SimplyDV. Aspect ratio pictures (above): Brian Thomas. Used by kind permission. No unauthorised reproduction or distribution. Copyright 2002, 2005, 2008.

Print This Page