pegbar

Widescreen TV in the UK
 

All new TV programmes in the UK are now broadcast in "letterbox" or widescreen format. Unfortunately, this can still be a source of confusion as can the issue of anamorphic.

These notes are primarily for 2D animators working in the UK.


 

note: click on illustrations to enlarge them

Fig. 1


Standard 12 field chart with 4 x 3 (12 x 9) aperture

The red area shows the shape of the standard 4 x 3 TV picture superimposed on a 12" animation field-chart (also known as a graticule). As you can see, they are not exactly the same shape, but they are close enough and we've lived with this discrepancy without problem for many years. Most cinema films were this shape until around the middle of the last century.

Screen shapes are described by a ratio, width to height. In this case, the actual ratio is 1.33:1 . This shape is more commonly referred to as "four by three", written 4 x 3 (that is 4 units wide by 3 units high). It's also sometimes referred to as 12 x 9.


Fig. 2


Standard 12 field chart with 16 x 9 aperture

Widescreen (also referred to as letterbox) TV shape is 1.77:1 or 16 x 9. This is close to the normal cinema screen shape of today (in the UK 1.85:1). Work for TV in the UK now has to be delivered in widescreen. When you work on a 12" field, the area that will be recorded to tape is the blue area in Fig. 2.  The width of a 12" field animation remains the same, but in letterbox, the height is just over a 9" field. This is the shape to base all your artwork on.

There is one immediate advantage from working in widescreen, namely that the top and bottom parts of your artwork are no longer in danger of being cropped. Cropping happens because every 4 x 3 TV set displays the picture differently - this is due to the design of the picture tube and the set's cabinet. The problem is less with modern plasma and LCD displays, but is still there. The good news is that everyone will now see the top and bottom edges of your artwork correctly regardless of whether they are watching on a 16 x 9 or an older 4 x 3 TV. The bad news is that the sides of the picture are still a problem, and that problem is aggravated by the use of the compromise 14 x 9 shape which is discussed below.


Let us consider for a moment the choices available to broadcasters regarding picture shape, having accepted that 16 x 9 is a fact of life and not just a passing fad. For some years, they must broadcast to people watching on a mixture of 4 x 3 and 16 x 9 sets. On reflection, you will find the only straightforward way to display a 16 x 9 picture on a 4 x 3 TV would be to show the widescreen image in the middle of the squarer screen - this results in heavy black bars top and bottom - the so-called "letterbox" shape.
Unfortunately, there has been large consumer resistance to this, so a half way compromise shape of 14 x 9 has therefore been chosen. It may seem like a crude solution, but it has been surprisingly successful in terms of acceptance and the majority of analogue terrestrial broadcasts are now transmitted this way.


 

Fig. 3


Standard 12 field chart with 14 x 9 aperture

For analogue terrestrial broadcasts, the 14 x 9 compromise means that both 16 x 9 and 4 x 3 viewers will receive the same green area in Fig. 3..
16 x 9 viewers will see black bars to both sides of the image, (the so-called "pillar-box" shape) - these are shown in blue in
Fig. 3.
4 x 3 viewers, meanwhile, will be receiving the green area but the sides will be cropped depending on the design of their TV set. They will have thin black bars top and bottom of picture. To add to the headache, however, you must bear in mind that most broadcasts on digital formats, that is satellite, cable and digital terrestrial (eg Freeview) will broadcast
true widescreen and the whole image will be broadcast, (that is the blue and green areas together), though, again what any one viewer will actually see at any time will depend on the design of their TV.

This is where the real problems for you as a designer begin. In Fig. 3 above, the whole of the coloured area will be recorded to tape - that's the full 16 x 9 area. For the reasons I have just outlined, however, you really cannot be sure what any of your viewers are really going to see. You can be sure that the top and bottom edges of your picture will be seen pretty well by everyone, but the degree of cropping of the sides of the images can vary widely.

The solution is to design for the full 16 x 9 shape, but allow for the possible cut-off that will occur on the sides of your pictures. The way to do this is shown in Fig. 4

remember: to see any picture more clearly, click on it to enlarge it.

Fig. 4


Standard 12 field chart: 14 x 9 with safe areas

pink area: the full 16 x 9 area
red line: the full 14 x 9 area
blue line: the safe action area for 14 x 9
green line: the safe titling area for 14 x 9

As can be seen, cut-off can be severe when safe titling areas are allowed for. Until letterbox shaped TV sets become the only ones used, you will have to get used to compromising the framing of your animation, never putting anything important anywhere near the sides of your picture.

As a very rough and ready guide, consider the safe title area within a 12 field on a traditional graticule as being at an 8.3 field. (That's roughly the green line in Fig. 4)

 


 

 Anamorphic for SD (standard defintition)

 

Fig. 5


Normal 16 x 9 image

 

Fig. 6

same circle, squashed anamorphically into 4x3 frame
Anamorphic version - widescreen image squeezed into 4 x 3 shape

Most clients will ask for work to be delivered in anamorphic format. This involves taking an image such as in Fig. 5 and "squashing" it to make it look like Fig.6.

This is necessary because all normal resolution tape formats record raw images that are always 4 x 3. If you are using a digital ink and paint bureau, it isn't your problem; there are no design considerations to take into account other than preparing artwork letterbox shape and allowing for the cut-offs at the sides as shown in Fig.4. You need know only to ask the edit house or ink and paint bureau to supply the finished rushes anamorphic - just blanking off the top and bottom of the image to make it letterbox shape is not acceptable.
Should you be generating your own computer output and going straight to tape, you will need to know how to squash your images to fit into the 4 x 3 shape. The article
here will help.

Displaying Anamorphic

If you have a video tape in anamorphic format that is going to be shown publicly, (such as an entry to a film festival), check with the organisers that their video projector can display anamorphic, particularly if the work is being shown abroad. I am advised by Video Europe that most video projectors are switchable but an old model may not be. Should this be the case, the only solution, unfortunately, is to have a non-anamorphic dupe tape run off with letterbox banding.


High Definition TV and Animation

HD (High Definition) TV is becoming increasingly important and any new animated project should take this into account. The problems we have with SD above regarding 4x3 versus 16x9 are no longer a worry, since all HD work is delivered in letterbox 16x9 shape.

You should be aware that there is not just one HD standard for TV - around the world are many variations in screen resolution, frame rate and issues of interlace versus progressive scan versus sequential field scan.

Now you are sufficiently worried, here is the good news; the technical issues raised by high definition that beset live-action film makers are not really anything for animators to worry about. But please read on so that you can be aware of issues that may arise when you take on a project to be delivered in high definition.

To be sure that you don't make any expensive mistakes, best advice is: don't do any production work until you have got your client to specify, in writing, the exact form in which your work is to be delivered.

The specification should tell you:

1 - the resolution of the animation in pixels (of aspect ratio 1.0 - "square pixels"). If you offer to do the work at 1920 x 1080, this will almost certainly be acceptable.

2 - the running speed in frames per second. This could be 24,25,30 or even some weird number like 29.976 (American TV). It really won't affect your animation, so don't worry about it, except that you must know what it is so that you can deliver the exact length of animation needed.

3 - how you will deliver the animation, that is whether it is to be on a video format like HDV or 35mm film negative etc. My advice is: try to insist that you deliver the animation as bitmap file sequences and not as tape or film media. This means that you burn the sequences as numbered bitmap files on CDs or DVDs and deliver these to the post-production house or editor(s). If they want the work delivered as, say HDV tape, still do it as bitmap sequences and have them transferred at a post-production house. Don't think you can get away with using your HDV camcorder. have it done professionally and allow for the cost of these transfers when you are budgeting.

At some point, a client or other person involved may ask innocently about whether you can provide progressive scan or interlaced. You will need to explain to your client that animation is always delivered as progressive scan*. If the finished production is to be delivered in something other than progressive scan, the post production house will need to take care of this and it really is not your problem.

Then go and talk to the post-production facility house that will be involved and make sure you know exactly what they will be expecting you to give them - if there are any doubts, do a few test frames for them. I used to recommend TIF or TIFF bitmap files but PNG files are probably now the best all round file format to use. If your animation is to be layered over live action or over other animation elements, be sure that your bitmaps have a clear alpha channel as a background. Once again: any lingering doubts, then discuss with the post-production house or editor and do a few trial frames.

The reason I am labouring these points is that although practically any technical problems can be fixed later, this is only at a cost and possible loss of quality. If it's your fault that this extra expense is incurred, it may well be that the cost will come out of your own pocket.

Once you are certain that a specific HD format is required and that the costs of any later changes to this will be borne by someone else, is it safe to proceed.

If you don't have a client or post production house to ask, use the following specification:

resolution:  
1920 x 1080 pixels
frame rate:  24 frames per second

The logic to this is that the resolution is the highest possible for broadcast television use and will be for probably at least the next 20 years. If a lower resolution copy is needed, this is not difficult. Indeed, many people in broadcasting insist that SD (standard definition) material down-converted from original footage shot in HD looks much better than material originated in SD.
It is
always better to reduce resolution than try to increase it from a low resolution original.
Regarding the frame rate; 24 frames per second is not used by any TV system anywhere in the world, so this may seem an odd choice. However, countless millions of movie films have been shot (and are still being shot!) at this rate, so every TV station in the world also has a means of transmitting them without problem.

A further advantage to using this specification is that your work can be used without expensive conversion for cinema release.

Interlace is not really an issue with animation, since it is not normal to use field rendering. One way to completely circumvent issues of interlace is simply to output your work as bitmap file sequences using recordable DVD media. Any facilities house with HD capability will then have no trouble in outputting to the required format. For testing animation in-house, you can of course consider working in lower resolution and only outputting at high resolution when you are ready for final rendering.

*regarding interlace versus progressive scan: in fact, it is technically possible to output animation as an interlaced image and this is occasionally done for certain problematical scenes involving pans that look too juddery. It is also quite simple to output 3D cgi as interlaced, so it is not impossible to do so much as a pain - rather than get bogged down in pedantic arguments, you really are best advised to simply tell your clients that animation comes non-interlaced and leave it at that

 

Ultra High Definition

OK, you've read the above, you've bought the latest and fastest computers to render the larger images and you are now ready to make future-proof animation footage.....sorry, but there's yet another twist to the tale!
As I write this in June 2008, work is already proceeding on the next format being designed to supercede high definition. With the backing of major broadcasters such as BBC and NHK, it does seem likely that ultra high definition will eventually become reality. This is made likely because of enormous improvements in the technologies used to compress video to get even very high definition images down to a practical bandwith. At present, storage and broadcast of ultra high def. is not practical and of course consumers would be less than happy to be told they have to upgrade their HD TVs and Blu-ray DVD players.
So, just in case you think that your film is so important that future generations will need to see it at its optimum quality, you will need to make your animation at a resolution of:
7680 x 4320 pixels  
I assume a running speed of 24 frames per second will be fine.

 

Shooting on Film, TV Usage

If you are originating material on film, such as on a rostrum camera, design with Fig. 4 above in mind. The extra picture that will also get shot above and below when you use an open or Academy gate will be "lost" at the telecine stage. Film has a far greater resolution than broadcast video, so it is not necessary to shoot film through an anamorphic lens and the part of the picture that is thrown away does not lead to any discernible picture degradation when broadcast.

 

Shooting Video for Cinema Usage

Reasonably good results can be got from taking an anamorphic broadcast image (720 pixels across) on Digi Beta or D1 and transferring to film using facilities such as those at Cinesite. They use software to interpolate the missing information when going from low (broadcast!) resolution to a higher (cinema) resolution. Allow for a small amount of extra cut-off at the top and bottom of picture; cinema shape is 16.65 x 9 rather than 16 x 9.
You should also prepare text and logos as separate elements that can be added at higher resolution. This means providing a version of the job without any lettering or logos. Better still, if budget will allow, have the whole job done at higher than broadcast resolution. (See the note on High Definition above). The actual number of pixels needed to produce cinema quality is a matter of some debate. The lowest number of pixels across the image generally accepted as being cinema quality is about 2,000, which equates to the 1920 x 1080 resolution. This is often referred to as "2K".

Things don't stand still, though and the future of digital cinema is most likely to be in 4K resolution which requires images approximately 4,000 pixels wide. Right now (Feb 2008), I would highly recommend you to use the 2K (1920 x 1080) pixel format for most work but to bear in mind that if ever you are to produce work for prestigious cinema release, you may well be asked to work at 4K. The overheads in rendering at higher resolutions are very great, particularly with 3D cgi, so you may well need to send your rendering to a "render farm" - a company who will take care of the rendering on your project using banks of dedicated computers.

Probably the most extreme high resolution requirement you are likely to encounter is for Imax cinema, where a 6K image is normal - oh and you will need to double the number of frames if the film is going to be in stereo!

 Comparative Resolutions

Ignoring Imax, the six most common resolutions you will encounter are given below in pixels. The more pixels you have to render, the slower your rendering times will be.

1 - S35 (Super 35mm) movie film    4,100 x 3,100
2 - 35mm movie film                     3,700 x 2,700
3 - S16 (Super 16mm) movie film    2,000 x 1,200
4 - HDTV                                    1,920 x 1,080
5 - 16mm movie film                      1,700 x 1,200
6 - SD (standard PAL TV)                 720 x 576  
 

Looking Ahead

Analogue transmissions will eventually cease and eventually the last 4 x 3 TV set will expire. 16 x 9 will then be the only shape to design for. An added bonus is that the newer plasma, LCD and OLED etc. flat-screens have little if any picture cut-off, so the problem of safe areas will be less of an issue.
HD (high definition) is now becoming widely adopted around the world and all "flavours" of HD use the same 16 x 9 shape. It means that one day designing for cinema, conventional and high definition TV will be a lot easier.
Until HD is universal, however, we're stuck with a messy compromise.


These notes were written by Jeff Goldner of Animation Post. Effort has been made to check the veracity of statements made and the accuracy of the illustrations, but no responsibility can be taken for the results of actions taken based on them.

Please email questions or comments to:

Jeff Goldner

Animation Post is a digital ink and paint service.

©  All material on this website copyright Jeff Goldner, except where otherwise attributed

Reproduction of the written and graphical elements on this web page without express permission for personal use only is fine.

Animation Post UK
Sep. 20 2002
High def. notes amended 5 Feb 2008
Ultra High def. notes added 18 June 2008