You know what editing is: every movie needs it.  But today's talk is all of "NLE", computers and a baffling array of technical jargon.  DAVID BRINDLEY, Course Director for Swan Rose offers a clear introduction to

DIGITAL EDITING FOR BEGINNERS

CONCEPTS OF ANALOGUE & DIGITAL

A video camera converts patterns of light and sound into an electrical signal in order that the pictures can be viewed on a TV screen, and the sound heard through either the TV speakers or a separate Hi-Fi sound system. The electrical signal produced by the camera varies in sympathy with the picture or sound it represents. It is therefore, if you like, an analogy of the original subject. Such electrical "signals" are thus referred to as ANALOGUE. If we could view these signals at minute intervals of time, we could measure the voltage of the signal at that point and establish a value for it. In other words, we could represent it by a number expressed in digits.

If we took this a step further and "sampled" the value of the signal many thousand times per second, we would produce a whole string of digits representing the value of each sample through time. This "signal", representing the original pictures and sound in terms of digits is referred to as DIGITAL, To take a non-video example an analogue thermometer measures temperature by the expansion or contraction of a column of mercury. The higher the temperature, the higher the mercury rises. A digital thermometer uses a sensor to determine the value of the temperature in terms of an electrical voltage which is then expressed as a number and displayed using digits.

ANALOGUE TO DIGITAL CONVERSION

This is not intended to be an in-depth technical explanation but we include it because it introduces technical terms which you may come across when studying manufacturers' literature. It also will provide a base for further study should you wish. We have seen in the previous section how video starts and ends its life cycle as an analogue signal. The original subject reflects a complex pattern of light, shade and colour and produces sound waves by pushing the air also in complex patterns. When these patterns are converted into electrical signals the challenge is to be able to transmit these faithfully and display them as glowing phosphor dots on the TV screen. (When you look at it this way it's amazing the thing works at all don't you think?) However, in video work the most difficult challenge is recording these complex patterns onto tape and then getting them back again.

The tape has its own "video noise" (that nasty white stuff you see when you play a blank tape) and the analogue recording processes struggle to keep the luminance (black and white picture) and the chrominance (the colour component) together, resulting in the only too familiar "colour bleed". This is where digital comes to the rescue. If we can describe this complex signal accurately as a stream of numbers (and particularly if we restrict the numbers to a choice of either the number one or zero , then the record and playback task is made so much easier. Instead of having to copy complex electrical patterns we merely have to transmit a "one" or a "zero".

Modern technology enables us to do this with a high degree of accuracy and reliability. Now, here come some of those technical terms: SAMPLING, QUANTISATION, CODING, COMPRESSION. The analogue waveform is SAMPLED many thousand times per second. Each sample is measured and its value expressed as a number which is rounded up to the nearest value on a pre-set scale of "quantum levels". This process is known as QUANTISATION and the resultant error due to the "rounding up" is referred to as QUANTISATION NOISE. So, you may be able to gather from this that the higher the SAMPLING RATE and the greater the number of QUANTUM LEVELS, the better the quality of the conversion process will be. In practice the sampling frequency and number of quantum levels are such that no quality loss is perceptible. The CODING process converts these decimal values into BINARY NUMBERS (made up of "ones" and "zeros"). The individual "ones" and "zeros" are known as BINARY DIGITS (or BITS). As you might imagine this generates one heck of a lot of ones and zeros - about 250 million of them every second. In order to reduce this to something more manageable, the digital signal is COMPRESSED. With very light compression (2:1) this has no perceptible effect on the final picture and delivers BROADCAST QUALITY. The more the signal is compressed however, the more information is discarded and so BROADCAST quality is no longer maintained. A rough guide would be 5:1 for DV and DVCAM, 8:1 for good S/VHS or Hi8 quality and about 10:1 for good VHS. Many amateur users employ 13:1 compression and still consider the VHS quality to be satisfactory.

COMPRESSION METHODS

Again, this is not an in-depth explanation of compression but merely an attempt to explain the difference between four compression methods that you will read and hear about whenever digital editing is discussed. These are MJPEG, MPEG1, MPEG2 AND WAVELET. The first major difference is between MJPEG and the other three. MJPEG compression is applied to every single frame of the video signal. The picture is divided up into blocks of 8 x 8 pixels and redundant information is discarded thus reducing the number of pixels required to describe that part of the picture. With the MPEG method a similar compression is used but not to every frame. Instead, the full compression is applied only to certain key frames. In the intermediate frames only the difference information is retained (non-moving background information is discarded). The total group of pictures between and including the key frames is known as a GOP (Group Of Pictures).

From here on MPEG gets rather involved. Very briefly and simplistically, the intermediate frames can be either "B-frames" (bi-directional) or "P-frames" (predicted). The key frames where full compression is applied are called "I- frames". "B" and "P" frames contain only "difference" information. MPEG1 is the version of MPEG used primarily to compress video sufficiently to allow playback from CD or with applications such as PowerPoint for business presentations. MPEG2 is a higher quality version of MPEG associated particularly with DVD and digital TV transmission. The main advantage of MPEG is the greater compression available with minimum loss of quality on playback, but the lack of information in the B & P frames makes it tricky for editing.

As far as video editing is concerned WAVELET is the new kid on the block. Like MPEG it offers heavy compression with minimal loss of picture quality but, instead of applying it in blocks, it treats the whole picture. Because of this it is claimed that WAVELET provides a softer and more acceptable TV picture not suffering from the occasional "blockiness" experienced with MJPEG and MPEG.

INTRODUCING DV AND DV/CAM TAPE FORMATS

Digital camcorders carry out the analogue to digital and compression within the camcorder and so write digital information direct on to the tape (ones and zeros). I have been told that digital camcorders are now selling faster than high band domestic analogue ones. In fact, I believe Sony no longer produce Hi8 domestic camcorders which they have recently replaced by new Digital 8 models recording digital signals onto Hi8 tape. When the DV (Domestic Video) format was introduced a few years ago, professional users were astounded by the quality and many claimed they couldn't tell it apart from Betacam. This is mainly due to the complete absence of video tape noise, no chroma shift (colour bleed) and a 25% improvement in resolution over S/VHS and Hi8. But, it must be understood that DV is NOT BROADCAST QUALITY (at least not in theory). DV recording samples the colour components at only half the rate required for broadcast (i.e. 4:2:0 instead of 4:2:2); the compression is 5:1 whereas broadcast is 2:1; and the resolution of 500 lines falls short of the required 600 or 700 for broadcast. BUT… to the eye the quality is brilliant. Copying to VHS however does, I believe, constitute the ultimate test and it is at this point that I think you will find it does not hold up against true broadcast. I edit entirely in DV and so I am used to watching my footage at that quality. When it finally goes down to VHS I could cry. Roll on DVD!!

DVCAM is similar to DV and produces no better quality (in my humble opinion!!). The recording process is much the same but the tape runs at 1.5 times the speed of a DV recorder and so the ones and zeros are written on a longer section of tape reducing the risk of errors due to drop outs. The big plus for DVCAM comes with the superior camcorders produced for this format and the better quality tape (which also comes in a much 'nicer' box!). But remember, theoretically this is NOT BROADCAST QUALITY.

STORAGE FOR DIGITAL VIDEO

PAL TV and Video produces 25 still pictures per second, creating the illusion of moving images. Each picture in turn is made up of 768 dots horizontally and 576 vertically making a total of 442,368 dots per picture. Each dot is described in terms of its three primary colours RED, GREEN and BLUE each colour requiring 8 bits. So, each dot requires 24 bits to reproduce true colour.

If you do the maths, you will discover that this comes to 442,368 x 24 = 10,616,832 bits. And that's just for ONE PICTURE! A group of 8 bits make up one BYTE. So, expressed in BYTES, we could say that one frame of video requires 10,616,832 / 8 = 1,327,104 BYTES. This is still a very large number and awkward to deal with, so we tend to work in units of MEGABYTES (i.e. a million Bytes) and ignore the odd few hundreds of Bytes. e.g. 1.327MB.

So, you can see from the above that, at a rate of 25 frames per second, we would need to capture, and store, something in the order of 33MB for very second of moving video. At the time of writing, despite recent massive advances in technology, this is just too much for the humble PC (or the MAC for that matter) and so now maybe you can see why it must be compressed to something more manageable. But even after compression, video takes up a lot of hard disk storage. Just to give you some idea, here are a few figures for you: DV footage (compressed at 5:1 in the camera) takes up 3.6Mbytes per second; S/VHS and Hi8 compressed at 8:1 using an 'MJPEG' capture card (such as the Pinnacle DC30PLUS or Casablanca) takes up about 2.7Mbytes per second. If you're shooting on DV you can reckon on storing around 4 minutes of video for every Gigabyte (one Giga-byte = a million Mbytes); on S/VHS or Hi8 you can expect around 6 or 7 minutes. Systems offering MPEG2 and WAVELET compression promise about a 20% improvement on these figures (e.g. Pinnacle DC1000, ScreenPlay and Casablanca Avio).

FIREWIRE, IEE1394 & I-LINK

As mentioned earlier, DV camcorders digitise and compress the video signal within the camera so the information recorded onto the tape is in digital format (i.e. it is a series of ones and zeros). Firewire is a method of transferring digital information between different pieces of equipment. The system was originally 'invented' by Apple (the Mac people) and from this an international standard known as IEE1394 was developed. Just to confuse matters, the IEE1394 standard has at least 40 variants to accommodate different methods of digital transfer. Sony jumped in, recognising the usefulness of firewire, for domestic home entertainment products in particular, and they came up with a simplified sub-set of the IEE1394 which they call "i-LINK". However, from a digital editing point of view, you can consider firewire, IEE1394 and I-Link as the same. They are all methods of transferring digital data from one point to another. DV camcorders all have firewire sockets (some unfortunately are input only) and so do VCRs and DV editing systems. The sockets are connected together by a very simple two-wire lead and data flows between. The lead carries (in digital form) not only the video and audio, but also timecode and machine control information. The major benefit of using firewire is the ability to transfer data from camcorder to VCR tape, or from computer to hard disk and back out to a digital VCR with no loss. The ones and zeros are faithfully copied across with no distortion and with no added noise or chroma shift of the final video output. It is therefore possible with a DV editing system to produce an edited master tape having the same quality as the original camera tape.

POPULAR DIGITAL EDITING SYSTEMS

Do not be confused by the term "digital" editing. This does not mean editing of digital footage only. It refers to a digital editing process which can be applied to any type of footage (analogue or digital) which has either been converted from analogue to digital form, using the SAMPLING, QUANTISING, CODING, compression process described earlier, OR has been digitised within the camcorder and recorded to tape as ones and zeros. So, it doesn't matter whether you are shooting on a V6000 using Hi8; an MS5 using S/VHS; a VX9000 using DV; or a DSR200 using DVCAM … YOU CAN STILL EDIT DIGITALLY!

All digital editing systems store the video and audio as digital files on a 'hard disk' (in the same way that a letter produced on a word processor is stored as a file on a computer hard disk). This has the advantage that the material can be accessed in a NON-LINEAR way. Instead of having to fast forward a tape to access a scene near the end of the tape, disk based material can be accessed immediately. Think of an LP record (remember those?). If you wanted to play a track near the centre of the record, you wouldn't drag the stylus round all the grooves to get there. You'd simply pick up the arm and move it across to the required track. THAT'S NON-LINEAR ACCESS!

The most popular digital editing systems are PC based and use an appropriate 'capture card' to bring the footage into the machine and get it out again. But, there are also "stand-alone" systems such as CASABLANCA and SCREENPLAY employing similar technologies but cunningly disguised to look like a common or garden domestic VCR. But don't be fooled by the apparent simplicity. These are very powerful video editing machines and deliver excellent quality. PC based systems tend to be more complex and take longer to learn (which is of course why we at Swan Rose love them so much), but they can do a lot more than a stand-alone system.

HOW TO CHOOSE THE RIGHT SYSTEM FOR YOU

If you are already reasonably PC literate (or at least not unduly scared of the things), and you are likely to need more advanced graphics and special effects, then I would certainly encourage you to consider one of the PC solutions on offer.

If you want an easy to operate, high quality, (and very portable) system, without the need to get to grips with a computer, then get a demonstration of CASABLANCA and of SCREENPLAY. Make sure you see them both.

You really do need to see these things in action to appreciate what they can do for your business (or just for pure pleasure for that matter). Remember however, if you are in business, you're not just buying an edit system - you're buying a 'money machine' if you use it right! So make sure it will do what you want it to do.

If you are close to us, then do give us a ring at Swan Rose on 01772 458300 or 623608. We can arrange a demonstration of any of these options at our Training Centre in Leyland Lancs (5 mins from M6 J29).

- DAVID BRINDLEY, Course Director, Swan Rose

Adapted from the text of a seminar held at the APV Show 6 October, 1999 and  used with David's kind permission.

Visit the Swan Rose web site.

"Experience is what causes a person to make new mistakes, not old ones." - Lee Prescott

Page updated on 21 March 2008

Authors' views are not necessarily those of The Institute of Amateur Cinematographers

Free JavaScripts provided
by The JavaScript Source