 |
|||||||||||||
|
|
 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
HD Production Guide August 2006
If you’re making the move to high definition, then you’ll be faced with a lot of choices. This short guide aims to explain some of the jargon and technology involved in those choices, and help you in getting the most out of your budget, and maximising your distribution and potential sales.
Unfortunately there’s no universal standard for a high definition production. The way technology develops, and the commercial pressures of the real world, mean that decisions will have to be made in terms of choosing the right tools for the job. Some production standards reach market dominance, but because of the speed of technological change there are always likely to be several different systems in use at the same time. They’re all capable of making high definition TV of one flavour or another; in different ways, at different quality and budget levels, and making different compromises along the way. There is “no one size fits all” that will suit all programme needs without compromise, even with an unlimited budget…so choices need to be made.
The nearest we have to a universal standard is the 1920 x 1080 pixel structure for an HD image. This is known as the Common Image Format (CIF 2k), as standardised by the ITU. Even so 1280 x 720 pixel HD images are also used in some areas of the world. The choice of the 2 pixel counts (Usually abbreviated to ‘1080’ and ‘720’) is largely associated with the geographic release area, but it helps if you can stick to one or the other. Sony professional equipment is based around the 1920 x 1080 Common Image Format.
There are three key questions that need answering before you start any HD production…
Having answered these questions it will decide your deliverables and will probably lead you in your choices of format. However there are other things to consider which we cover in the following guide.
Destination…where will you sell your programme?
Theatre releaseResolution is king. The bigger the screen, the more resolution you need. Ideally a 1080 line format. Ideally a progressive scan frame rate that converts easily to the cinema frame rate of 24fps. Film cameras and projectors always acquire and display their images a complete frame at a time, as you’d imagine. In the video world this is known as ‘progressive scan’ in order to distinguish it from ‘interlace scan’ traditionally used in the standard definition TV world. With interlace scanning you typically capture twice as many images, but at half the vertical resolution. This gives good temporal resolution (smooth movement) but means each frame consists of two fields (half frames) from different instants in time.
If you want to make your two fields of interlaced video into a single frame in order to transfer to film you have to deal with the fact that the two fields are different. Much easier if your video system can work in progressive scan, so a frame of video is transferred to a frame of film.
Japan / US / NTSC worldIn these regions standard definition TV runs at 59.94Hz*, and they prefer their HD formats to work well with that standard. This means either shooting at 60i, or shooting at 24P and applying a 3-2 pull-down as you would with a movie. (3-2 pull-down is a means of converting 24fps to 60i You could shoot 50i and do a HD standards conversion to 59.94i. You could shoot 25P and play your tape at 24P, and then do a 3-2 pull-down. Both these latter two options will need to be discussed with the co-producers first.
DVD releaseHD delivers a high quality, low noise master that responds well to the MPEG encoding required for a DVD. This encoding will exaggerate flaws and noise in the original, so a high quality source like HD will give better results. If you have an HD master it can be used for both standard definition DVD release and a later HD release. Don’t forget, you might want to re-edit for an HD release, as it is often better to hold a shot longer in HD and cut less frequently, because the amount of information on the screen needs time to be absorbed by the viewer. Your shooting frame rate decision will make delivery to region 1(USA) or region 2 (Europe) easier or more complex. 50i is simpler for region 2, and region 1 will require a standards conversion of the HD master. 24P / 25P will work well for either region assuming the lower frame rate is appropriate for the content.
Blu-ray / HD-DVD A new generation of high definition DVDs is upon us. Requirements for a Blu-ray authoring and mastering are still being finalised, but as it is extremely high quality with 5.1 surround sound capability, you should consider mastering and delivering in the highest possible quality such as HDCAM SR
Web Shooting for the web shouldn’t affect your choice of format or frame rate. HD encoding systems from Real, Windows Media, QuickTime etc. will all run at various frame rates, and the high level of compression required is likely to mask any 1080 line / 720 line issues.
BSkyBThe preferred delivery format for BskyB is HDCAM SR, partly because of its multiple audio channels, but also because of the low data compression. This doesn’t necessarily mean you need to shoot on HDCAM SR though. At present, Sky satellite transmission is 1080i line, but the platform also supports 720P lines as well.
What type of programme content will you create and use?
Content
Archive material Consider the proportion of archive material that needs to be introduced into your program. Can you access original film negatives? Can you find a version that has not been cropped for 4:3 television if originally shot in a wider format? If importing Betacam / BetacamSP / BetacamSX / Digital Betacam material, the HDW-M2000P/20 series VTRs may be useful. They will up-convert and aspect ratio convert tapes from all these formats, and give you an HD-SDI output from the VTR. Just put your tape in press play. Menu settings give you access to crop and letterbox options if required.
Scientific / technical material The ‘film look’ of 24P or 25P acquisition may not be appropriate for this kind of programme. 50i would be the natural choice. Any item when you’re analysing what happened in a fast moving event needs to be shot as for a slo-mo, with a faster frame rate. 25P would not usually be appropriate for a documentary about analysing performance in sport. (Except perhaps for your flashback sequences!) You may also need to use mini-cams, crash cams etc, and be able to record for long periods while waiting for an event to occur.
ReconstructionsIf reconstructing past events, it may be useful to use a different frame rate to emphasise the passage of time. Typically, if shooting the main programme at 50i, you might shoot reconstructions at 25P. This would give the reconstructions a more ‘filmic’ look to motion, with a bit of the blur and judder you see with film.
Super Slo-mo Producing a good slow motion effect relies on running the camera at an increased number of pictures per second. When slowed down for playback, there are still enough pictures per second to give the impression of smooth movement. Any material played back at less than 25fps starts to look ‘juddery’ and discontinuous. This means that if you want to play back at 1/3 real speed for example, then you’ll need to shoot at 75fps. Most HD cameras will work at a maximum speed of 60 pictures per second. These pictures will usually be 1080i, or 720p. We are starting to see cameras that will record 1080p line frames at 50 or 60 Hz (e.g. HDC-1500 recording to SRW-1) but this option may be outside the budget limits. A simple option if shooting with an HDW-750P at 25P would be to change to 50i for slo-mo sequences, and play back at half speed. This will give you effectively 25i, which will match the frame rate of the main program perfectly, but at reduced vertical resolution. (You have got 1080i lines to start with though) There are also specialised cameras that will record very high frame rates, usually at a cost in terms of picture quality and ease of operation. If you need a very specific frame rate, the PDW-F350 HDCAM XD camcorder will operate at any frame rate between 4 and 60 fps. Film cameras of course may be the easiest option, and 35mm high speed film will integrate well into an HD production. (It will of course cost more to shoot high speed film, as you’re running more film through the camera…probably not for too long though.)
Time lapseA time-lapse sequence is the opposite of shooting for slow motion. You need to record at a low frame rate, for instance 1 frame per minute, and play back at normal speed. Traditionally this has been a difficult task for tape based formats, as control of a mechanical tape drive and head drum to record single frames at a time is very difficult. More recently, memory cache type options have become available, that store frames in a solid state memory until the memory is full. The accumulated frames are then all automatically saved to tape, and the cache is ready to continue. This can make time-lapse very simple, for instance with the memory cache option installed in the HDW-730S/750P/F900R camcorders. The inter-frame compression systems used in HDV camcorders make them unsuitable for time-lapse. Digital stills cameras can also be used.
SFX Special effects often require the use of chroma-keying against a blue or green screen background. In order to get the best separation from the background it helps to have plenty of resolution in the individual colour channels. Many recording systems rely on the fact that the human eye cannot resolve colour information as well as it can resolve luminance information, and consequently allocate less data to describe chrominance than luminance. (Phrases such as ‘Colour sub-sampling’, ‘4:2:2’, or ‘4:2:0’ will be typical of the jargon associated with this process). If chroma-keying, multiple layers of effects, and a possible cinema release are parts of your plan, then you may wish to consider 4:4:4 acquisition. In this mode there is no reduction of data about the colour information. You will need special cameras and recording systems though: for example a Sony HDC-1500 with 4:4:4 option or an Arri D20 recording to an SRW-1 HDCAM SR tape deck.
Vox pops / self opIf you’ll be recording a lot of material from members of the public keep in mind the likely shooting ratio. Media cost and length of available tapes may be an issue. Also if there are diary cameras involved, simplicity of operation will be important. HDV is the obvious choice, especially if you need a non-threatening or discrete camera.
Workflow…how will the production process work?
Turnaround timeIf you need a very fast turnaround time, formats such as XDCAM HD may be of interest. Although tape-less formats may allow rapid ingest onto a server or hard drive, once ingest is completed there is no further difference in the editing process. Tape ingest may also be a useful opportunity for review and logging. Formats such as HDV can be edited on location via suitable NLE platform, but if you need to get material back to an edit suite then a tape in a jiffy bag is often still the best way. In order to send native HD files over an internet link you’ll need a very fast connection indeed.
Delivery formatEach broadcaster will have their own requirements. BBC delivery for international use for instance should normally be 1080 / 25P or 1080 / 50i.
Discovery channel requirements can be found at: http://www.discoverychannel.ca/_includes/disclaimer/producers_guide/docs/HDSpec.logo.doc
You should always check carefully with all production partners that your shooting and delivery formats are appropriate. HDCAM SR is becoming well accepted as a delivery format because of its very low data rate reduction, and 12 channels of audio. It’s unlikely that you’d shoot entirely on HDCAM SR though, unless you have a very large, 35mm, type budget.
Archive / StorageTape is still the best all round storage format. If you need to keep a secure copy of your rushes or your finished programme, there’s nothing quite as reassuring as a row of tape boxes on a shelf. Cost wise, this is also the cheapest way by far of looking after your assets in terms of £ / Gbyte. Tape formats such as HDCAM-SR are right at the cutting edge in terms of quality and technology, and because of their ease of use and practicality will be around for many years.
Editor choiceIn the early days of HD, the only way to edit was in a linear edit suite with a couple of VTRs, an edit controller, and maybe an HD switcher. We’re now getting close to being able to edit an HD programme on a laptop. All the major NLE manufacturers can offer a range of HD solutions. You really just need to know that you have access to play in machines that can handle all the formats you will be using, and that have the transport mechanisms and features that suit your budget. (In general the more expensive machines will have faster more responsive mechanisms, RS-422 control, more outputs etc.)
Grading All HD formats can be graded to some degree, though in general you’ll have more room to manoeuvre with higher quality recording formats. A lot of HD cameras have the capability to pre grade in the camera by loading default set ups into the camera in advance of shooting, even on multiple camera set-ups. This enables a more consistent look to be achieved up front. Grading can still of course be carried out at a latter stage, but time spent matching up cameras can be minimised. Camera set-ups can also be optimised for grading, though it would be useful if the DoP, the director and the colourist all had a unified vision of what is trying to be achieved, as such a camera set-up may not give a very impressive look on the camera rushes.
Metadata Metadata is information about the programme that is stored side by side with the picture and audio data. It could be as simple as the date of the recording, through to a file containing edit decisions or voiceover script. Tape based systems tend to have less space available for meta-data than for instance an optical disc system like XDCAM HD.
Shooting… how will you get the images you need? Extreme conditionsExtremes of heat, cold and humidity can harm any production equipment. HDCAM is based on the mechanics of Digital Betacam, and HDV is based on the mechanics of DVCAM. You’ll get similar levels of performance to their predecessors…which seem to have served in every conceivable situation at one time or another and hence achieved their benchmark status.
Mini-CamsMiniature cameras tend to be of lower quality, and there are few HD miniature cameras around at present. Having said that, HDV camcorders such as the HVR-A1E can be useful, and remote cameras such as the HDC-X300/X310 will give a high quality HD output, though it’s not that small.
UnderwaterUnderwater housings are available for most HDCAM and HDV camcorders. Most of the underwater footage for ‘Planet Earth’ was shot with an HDW-750P
Remote controlSome operators prefer to have a remote control panel (RCP or OCP) attached to the camera to control iris and the pictures. This gives you the freedom to change detail, knee, matrix, colour temperature, black level etc. but it may not be the kind of freedom you want on a fast moving shoot.
HelicoptersCheck you can find the correct mount or harnesses for the camcorder you have chosen. Safety is always the first concern. The HDCAM camcorders are very similar size and weight to Digital Betacam camcorders that helicopter companies may be more familiar with.
The Look…camera set-up, post and grading to give the image quality you want
The most important factor in the look of your pictures is how the camera is used; the lighting, framing, positioning and movement of the camera, and the choice and use of the lens. Unfortunately there is no set-up card that will do this for you.
Also vital to the look is the choice of frame rate. 24 or 25fps will create motion artefacts similar to 24 or 25fps film. 50i or 59.94i will give a smooth look to motion, more associated with video.
Note: Progressive film scanning “look” is only relevant at low frame rates of 24P, 25P and 30P. If progressive scanning is done at higher frame rates such as 50P and 60P then the film look is not present and instead you get clean television type motion. The only difference at these frame rates compared 50i or 60i is the fact that you will achieve slightly higher subjective vertical resolution. However the eye is more sensitive to increased horizontal resolution which is why the horizontal pixel count is higher in all formats.
As long as the camera is not badly set up, the next most important stage in terms of its effect on pictures can be post-production. This depends whether you go through a grading stage, in a similar process to film, and possibly a film effect process such as field doubling. Different cameras allow different levels of in camera picture control. The new HDW-F900R for instance allows gamma curves to be designed on a PC and downloaded to the camera.
Post Production Most things are possible in Post. It depends how much time and money you are prepared to spend. It is often worth considering a grading process similar to that used in film production. Colour correctors of a wide range of quality and sophistication are available, but if you feel you got the colourimetry right in the camera (see grading section) it may not be necessary.
Field doubling or film effect is a more controversial process. Film cameras have a temporal resolution of 24 or 25 pictures per second. This isn’t quite fast enough to completely convince the human eye that it is seeing a moving image, and a trace of flicker is visible on movement. This distortion is part of the film look, and may in some way distance the viewer from the action and give a sense of unreality. Distortion though it may be, this is viewed as a good thing. HD cameras can often shoot at low frame rate (25P or 24P) or a higher rate (50i or 60i) If you shoot at 25P it’s not possible to accurately recreate the missing pictures and convert your material to 50fps. If you shoot 50i you can generate a 25P version, by discarding every other field of your 50i and doubling up the lines. This will reduce the vertical resolution A video camera produces 50 interlaced fields per second, two fields going to make a frame. If the second field of each frame is discarded the temporal resolution is reduced to 25Hz, the same as film. This is essentially what field doubling does, and is a real time process in post production. The downside is that half the vertical resolution is also lost, though as discussed above, we tend to have a surplus of resolution. The process needs to be used with care as on some shots it can appear to flicker too much, and it may be worth looking for a facility that can effectively turn the effect upand down by mixing back in part of field 2. Some like the effect, others don’t. The film look is a very subjective thing.
Menus and set-ups
The set-up of the camera is perhaps less important than other factors, as long as the camera is not badly set up. Don’t try and do too much in the camera until you are happy that you know what the end result of a particular adjustment will be. Approach your film look a step at a time, updating your set-up card as you go.
Depth of Field Depth of field is inversely proportional to image format size. The size of a frame of Super 16mm film is about 35% larger than the active area of a 2/3” widescreen CCD. As a result the depth of field is narrower, making it easier to isolate a subject within a frame by keeping background out of focus, or to pull focus from one part of the scene to another. Smaller format CCDs will give an even greater depth of field, and everything in shot being in focus is often a tell-tale sign of material shot on video. There are two ways to minimise depth of field on a video camera. One is to use ND filters to ensure that the camera aperture is at the wide end of its range, around F2.8. The other is to use longer focal length lenses where possible, though this may necessitate positioning the camera further from the subject than normal.
Lenses In general, standard definition lenses can be fitted to HD cameras, and will work perfectly well, though it’s a bit like buying an expensive hi-fi system, and using a cheap set of speakers with it. Not recommended for material that may be viewed on large HD displays. You may get away with it if your programme is purely for SD transmission. There is now a wide range of prime lenses and film style zooms available from most manufacturers, as well as a number of adapters to mount film lenses on video cameras.
Filters and effects If you want a heavily filtered kind of look to your pictures, with coloured grads and effects the same wide range of filters and optical effects is available for the front of your camera as in the SD world. Soft filters aren’t usually necessary for an HD shoot though, as detail correction tends to be a lot gentler in HD because there’s a lot of resolution there to start with. Remember you can’t get that resolution back again once you’ve put a ½ pro-mist on your camera. Colour grads can be ‘undone’ during grading to a certain extent, but shooting clean pictures and adding effects later is not usually a bad policy. You’ve usually got that convenient ‘undo’ feature in post!
Sets and costumes Your choice to use HD has some implications here, but the frame rate and format is less important. HD will show flaws in costumes, make-up and sets, and people do tend to get nervous about this. In practice, it’s not very different from shooting for 35mm. Usually sets will not be the focus of attention, and careful control of iris and focal length of lenses can often keep the depth of field narrow enough to keep therough edges out of focus. None of thisisnew if you have experience in 35mm and maybe 16mm film production, it is just about attention to detail.
Budget
If you’re thinking of your budget in terms of what you would have been shooting on in standard definition, then you can very roughly translate that budget to the HD world.
Obviously this is a simplification, particularly when equating a film shoot with an HD shoot, but it will at least give you an idea. The cost of HDV tape stock is similar to DVCAM. HDCAM stock costs about 15% more than Digital Betacam, SR stock, though expensive by tape standards costs a lot less than 35mm film stock.
Factors to be taken into account include:
The cost of shooting HD is coming down all the time, and there’s usually a format that suits your budget, though you may have to choose a suitable compromise in terms of data compression, optical quality, or colour resolution.
Examples:
Conversion Chart
As there’s no single shooting choice that will be perfect for every delivery specification, there must inevitably be compromises if you want to deliver to more than one ‘zone’. This has always been the case even with SD, and processes such as off-speed playback and 3-2 pulldown have been used for many years. Every time you watch a movie on NTSC TV it has gone through a 3-2 pulldown. In PAL your movies have always been played back at slightly the wrong speed!
The followingchart shows you the process you will need to go through to deliver to the various transmission formats from a given shooting frame rate.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
[Home] [Contacts] [Training] [Background] [Links] [HDC Systems Cameras] [Time for some filing] [HD Production Guide] [Decoding the Jargon] [Downloads] [Viewfinder Adapter] [Stuff you can buy] |