Cropping can be a result of incorrectly capturing the video image at the
wrong size, poor video processing done on your footage as you turn it into the
correct MPEG format, or something to do with your DVD player, or television.
Most domestic televisions are designed to perform some cropping of the received
images anyway.When you suspect cropping is occurring, it is very hard to determine the cause,
and exactly what effect it is having on the displayed image.
Professionals of course have a vast array of expensive equipment at their
disposal and training in how to use it. The amateur's only real option is to
look at the images shown on the television screen and compare them to what
the video processing tools claim they have produced, or view the MPEG file on
a PC and try and spot the differences between that, and what appears on the
TV.
It's notoriously difficult to compare to moving images on different displays,
not least because an exact frame for frame comparison can't be done easily.
A better option is to use a still image, and most Video CD burning
software ( such as Nero ) allows JPEG stills to be placed on a VCD or SVCD,
which makes the task a little easier.
What you really need though is a Video Testcard. This is a predefined image
which can be used as a reference, to compare how different DVD players and
televisions display the image.
Testcards have been used in the television industry for many years, firstly as
channel identification ( "Idents" ) before programme broadcasts, and for
making sure cameras and monitors are setup correctly, and "matched" in the
studio.
They are also broadcast over the air, to allow television sets to be adjusted,
but with the advent of 24 hour television, they are rarely seen these days.
The testcards here perform the same function as those which used to be
transmitted by the broadcasters; by viewing them on a television, you will be
able to tell what parts of the testcard image is being cropped, and you will
also be able to judge the output quality of the processing which has been
carried out on them to get them to the screen.
The testcards do not solve the problems of cropping or poor quality, nor
identify where such cropping or quality problems are occurring, but they do let
you see what is happening from a standard reference point. If your TV and DVD
combination seems to be extremely bad, you can take the disc round to a friend's
house and see how their playback system compares. You can even drag it off to
the local television retailer and see what it looks like on top of the range
equipment if you want to.
From my own experience, finding testcards is extremely hard. There are many
images which have been captured "off air" in galleries which celebrate the
history of television, but these are usually poor quality captures, and are
copyright protected material.
If you do find a high quality testcard image, it will probably be in the wrong
size, and reducing it to the correct size destroys the image, making it useless
as a quality testcard, or again, it will be a copyrighted image.
Having found myself in the situation that I wanted a testcard to check how much
cropping I was seeing on my television setup, and being unable to find any
suitable testcards, I set about designing my own.
These test cards are included here, and are made freely available for your
own use.
There are four testcards; "PALVCD-A", "PALVCD-B", "PALVCD-C" and "PALVCD-D".
All are 352 pixel wide by 288 pixel high images, which is the standard image
size for PAL VCD 2.0, which is what I use. I have labeled the testcards "A"
through "D" to indicate which cards they are and the number following the
identifying letter indicates the version number, should the cards change for
some reason.
The first two testcards ( "PALVCD-A" and "PALVCD-B" ) are for standard 4:3
television viewing. The "PALVCD-C" testcard is a full height, horizontally
compressed image which should display correctly on 16:9 televisions. The
"PALVCD-D" testcard is a full-width. vertically compressed image within a
letterbox which should also display correctly on 16:9 televisions, once the
letterboxing has been removed by in-television processing.
I have no intention of creating
further testcards for NTSC ( 352 x 240 ) nor other systems, as it is not
a simple task, and it would give me no benefit whatsoever. If you don't use
PAL; I'm sorry - You are welcome to take the images I do have and, crop, stretch
and modify them for your own use, but note that a change of aspect ratio will
affect the display of the circular parts of the image.
The images shown on this page are full-size GIF files, which are uncompressed
images of the testcards. Being uncompressed means that every pixel is where
it is meant to be and every colour is what it should be.
Unfortunately, most CD Burning software will not accept GIF files as being
usable for VCD or SVCD creation, but they usually accept JPEG files.
The best way to create a JPEG file is to take the GIF file and run it through
a GIF to JPEG converter program, without any compression. I have tried to do
this with the tools I have, but even at 100% quality, some compression has
been automatically applied.
A 100% quality, 1:1 compressed JPEG file can also be considerably larger than
the original GIF file, 100KB plus compared to the 8KB or so GIF file.
Having spent some time experimenting with the tools available, I have included
JPEG versions of the GIF files for those who do not want to have to create
their own.
For most people, these JPEG files will be suitable. If you want to create
better quality JPEG's, then download the GIF files, and find a suitable
Graphical Tool which will allow you to create what you need at the quality
desired.
In terms of using the testcard; all that is needed to create a reference disc
containing the testcards is to put all the downloaded ( or created ) JPEG
images onto the disc. In Nero version 5.0.2.4 ( and later ), this is a simple
drag and drop mechanism, once the creation process for a VCD or SVCD is
started. For users of other CD Burners, you may have to consult your manual.
The 352 x 288 image format is suitable for use with both VCD and SVCD under
Nero, and should work with other CD Burner software. The image will not be
replayed at the maximum quality that SVCD supports, but should be suitable
for most uses; it's better than nothing.
Once you have a disc burnt; you should be able to play it in your DVD or other
player ( assuming it supports VCD 2.0 or SVCD ), and the first testcard will be
shown. You will probably have to press "Next" or "Step", to move onto the next
testcard, but exactly what you need to do will depend upon your player - if
you are having problems; you may have to read the manual.
Image Resolution and Aspect Ratios
The standard PAL VCD frame resolution is 352 pixels wide by 288 pixels high,
and displays correctly on a 4:3 television.
One would expect the frame resolution aspect ratio ( 352:288 ) to match that
of a television ( 4:3 ), but that is not the case; pixels on a television are
not square.
The correct frame size ratio, in pixels, for a 4:3 display, is ...
Width Height
----- x 1.090909 = ------
4 3
To create a widescreen ( 16:9 ) testcard the correct frame size ratio is ...
Width Height
----- x 1.090909 = ------
16 9
A full height, 288 pixels high testcard should be 470 pixels wide. The full-size
widescreen testcard can be horizontally compressed to the normal 352 width, as
has been done with the "PALCVD-C" testcard.
A letterbox, full width, 352 pixels wide testcard should be 216 pixels high,
with equal padding of 36 pixel above and below the image to place it in a 288
pixel high frame, as has been done with the "PALVCD-D" testcard.
A Very Important Note
The GIF images here are as perfect as they can be ( ignoring the fact that
the colours and grey scale bars have not been officially calibrated ), but
further processing, turning them into JPEG images, and then burning them as
MPEG images to a disc may introduce artefacts which look like calibration
errors when viewed on a television.
This may lead you to perceive that there are calibration problems with your
system when none exists.
I ran into this problem with one image I created. One black-white contrast area
and a grey-black chequered stripe both showed a lot of ghosting and other
artefacts on playback, on two televisions I tried, although other similar
areas did not.
These artefacts were not present in either the GIF nor the compressed JPEG
images but did appear in the final television image.
It can only be assumed that the combined compression mechanisms used add the
artefacts to the final disc image.
I am looking at how to overcome this problem, and imagine that part of the
solution is to create a 100% quality, uncompressed JPEG image from the GIF
image before burning it to the disc. The other part of solution may be in
chosing a CD Burner which performs its JPEG to MPEG still image conversion in
a cleaner way.
The images here do not seem to contain any misleading artefacts when I have
processed, recorded and played them back on a television.
Using the Images
Users of Microsoft's Internet Explorer should use the following procedures
to access and download the images. Users of other browsers may have to
consult the documentation which comes with their product.
To save the 100% quality, uncompressed .GIF images - Place the mouse
cursor over the image, click the right-hand mouse button, and select
"Save Picture As ..."
To save the high quality .JPG images - Place the mouse cursor over the image,
click the right-hand mouse button, and select "Save Target
As ..."
To view the actual .JPG images - Place the mouse cursor over the image, and
click the left-hand mouse button.
