| |
Lesson 3
Step by step procedures to video capturing with the RealCap APAS
Module
| Make sure you have a video source connected to your capture card. In
this case it is the I-Omega Buz. This video source can be coming from a VCR, a camera, CD
or any video output. This video output should be connected by S-VHS cable to the I-Omega
Buz video input. You can use also RCA cables for Composite video input. However, the
resolution is not as good. The APAS digitizing system can enhance the markers or the
image. But, in all cases, S-VHS is a better video input.
|
| The following video file (2.3MB)
which you can download, will show you all the following procedures. It is recommended that
you download the file to your computer and then run it with the Microsoft Media Player
that can be found in the Windows directory. The name is exactly: Mplayer.exe I
made a link here to the Mplayer.exe so you can link to it directly. But this will work
only if your Windows directory is on your c:\windows\Mplayer.exe
sub-directory. This way you can advance frame at the time on the video file and see
the present text at the same time. Make sure you download the video Lesson3.avi (2.3MB) file first.
|
| Run the RealCap program:
|
| The first screen you see will look as following:
|
| The Menu bar on the top shows the: File ; View; Options;
Capture; and Help.
|
| Under the File options you have the following selections:
File:
| You can save the video in any location on your hard disks for latter
use.
|
| You can go to trimming so you can select the precise sequence and
synchronize the number of frames and starting and end fields to digitize. However, you can
do the trimming later on.
|
| You can trim an old video file that you save before.
|
View:
| In the View menu you should select all and the Overlay rather then the
Preview. Detailed about the difference between Preview and Overlay modes are explained in
previous lesson. Always select the Overlay Mode.
|
|
| Options:
| Under the options you can select Video Format and Video
Source. These are two important options to make sure that you receive the
correct video format whether it is PAL or NTSC and also selecting the proper size and
resolution.
|
Video format
The video format dialog is activated
using the menu item Video Format in the Option menu.
Using the
video format it is possible to setup the image size, format (compression type) and amount
of compression which is to be used under the capturing process.
To enable capturing of video images in full size and speed it is
absolutely necessary to capture in Motion JPEG format. If this is omitted your system will
not be able to do it.
Compression
(Quality)
Pressing
the Compress... button the amount of Motion JPEG can be adjusted. The amount of
compression is inversely proportional with the image quality. Meaning that the higher the
compression (less space for each image) the more information about the image you will have
to throw away. Also, it is important to keep in mind, that higher quality images,
requires higher throughput of the PC bus and hard drive. Just to see the differences
between the different amount of compression observe the video at the Lesson3.avi (2.3MB), which you download before at 33
kb/frame to the following 10 seconds of video capturing (600 fields) of the gait video data study at 10 kb/frame
(2MB). Notice the degradation in the video quality. This file was highly compressed to
save disk space and show the full raw data without trimming. However, normally the video
is compressed to 70 kb/frame which produce a very good image.
In order to compare the degradation in image quality
due to compression please download the following files.
1. Gait
data taken at 76KB per frame for 3 seconds, resulted in .7MB file size.
2. Gait
data taken at 10KB per frame for 10 seconds, resulted in 2MB file size.
3. Gait
data taken at 50KB per frame for 5 seconds, resulted in 2MB file size.
One can observe that the file taken at 50 KB per frame
resulted in a very efficient file size and the image is very clear and will meet the
automatic digitizing requirement. The file taken at 10KB per frame is not acceptable since
the image degraded too much. The best quality file is obtained at the 76 KB per frame.
However, the file size increase significantly to 26 MB for 5 second. But the image is
great. With large hard disks of today, there is no reason not to use this level of
compression.
| Video Format:
|
In the video format, always select the Motion JPEG and Image Size of 720 x 480. However,
in some conditions when you need a long time data collection, such as 5 minutes, you may
use the 352 x 240 window size which will reduce the file size but also the image size.
| Compress: For the Compress
Option select around 70 KB/Frame.
This depends on your computer speed. The new Pentium III
can go all the way to 200 KB/Frame. However, always remember that the greater
the KB/Frame, the larger is the file and the better is the resolution. We found that for
digitizing purpose, there is no significant difference between 60 to
200 Kb/Frame rate of data capturing.
|
Video Source
This part of the documentation is related
with the hardware setup of the Iomega Buz. We have not included the setup of other
hardware vendors but the basic idea by Video Source configuration is the same but the
front end to manipulating the setup is different.
Video source
The video source dialog is activated using
the menu item Video Source in the Option menu.
Using the Video Source dialog it is possible
to change various aspects of the video input signal.
Video input is used to setup the connection of
the video input i.e. where the video signal from either VCR or camcorder is provided.
The Iomega Buz is capable of detecting the
video standard supplied on the input. Clicking the Detect button adjusts the Video
standard for the appropriate standard. If the video input is not set correctly the
Detection process will not work.
The sliders in the bottom of the dialog
(Brightness, contrast, Saturation and Hue) is used to modify the image quality when
capturing. The settings of these settings is kept after the Software module ended. These
settings are important in setting the amount of contrast and brightness to enhance the
capability of the automatic digitizing and detection of markers. However, for manual
digitizing you should prefer the brightness to the contrast.
|
| Video Source:
| The video source is very significant. You must determine if you have
NTSC or PAL system and select the proper one. Also, you must select the proper cable
of video input if it is Composite (RGB, RCA cables) or S-Video which has a special cable.
Always prefer S-Video when it is possible. You can notice also that you can select the
Brightness; Contrast; Saturation and Hue to enhance the image. Always prefer the Contrast
to be higher and Brightness lower when using automatic digitizing with reflective passive
markers
|
| The following Video illustrates how
the various images controls are working:
(2.2 MB)
|
|
| Capture
:
Capturing Preferences
Capture options
Frame Rate: This parameter is the frame
rate of captured video sequence. The frame rate must be given in
frames/sec and not fields/sec. !!! However, the APAS software will separate the
frame into its two fields components. So, even though you capture at 30 frames per second
(25 for PAL), the APAS will allow you to analyze the data at 60 Hz (50hz for PAL).
Capture Time: The amount of sec to be
captured. When capturing the process can be halted by hitting the Esc key on the keyboard.
You should always select at least one to two seconds before the sequence of interest and
one or two seconds after the sequence of interest. This way you will not loose any part of
the sequence. When I was capturing the sequences from the Olympics, I set the
capturing time to 10 seconds before and 10 seconds after. In this way I never lost any
frames of interest. Just remember, that if you lost some frames you can always capture
again from the original video tape. This what is so great about video digitizing. You
never loose the original tape.
Enable Time Limit: Enables the automatic
time limit capturing. If this is unchecked the capturing process will continue until the
Esc key is pressed or the hard drive is full! You must do that or you will fill your hard
disk very quick. Why you would ever want it to work without checking? When you capture in
real time and do not know the end of the sequence, then you use the escape key to
terminate the capturing.
Delay: When starting the capturing process
a small delay can be invoked just before the actual capturing process is started. This
delay can be used in concert with VCR equipment which requires different time delays
because of the difference in mechanical construction. This delay should be used only with
a VCR where it takes time for the VCR to start. Make sure you add this time to the totally
time. So, if the delay is 1 second and the capture time is 5 second, you will loose the
first second. So, select 6 seconds to compensate for the delay. I do not use this option
since I always start the capturing ahead of the point of interest. In this case I do not
care if the first few frames are missing or the VCR did not get the full speed. Since the
sequence going to be trimmed latter on anyway, it is sometimes feel more comfortable and
time saving just to capture early and then trim.
Enable Delay: Enables the time delay
before capture start. In order for the delay to work you must enable it.
Temporary File Allocation
Filename: This is the filename of the pre
allocated video capturing file which is used during the capturing process. The default
name is c:\capture.avi. This file should be allocating at least 10MB of memory and
it will be overwritten every time you collect data to this file. It is really a temporary
file to allow contiguous disk space to optimize real time transfer of data to the hard
disk.
Allocation Size in Mb: This is the amount
of hard drive space which is to be used in the pre allocation file. The default is 10Mb
which in moderate compression is enough for 4-5 sec of video data. This translate to 300
frames. If you need more then 300 frames, then increase the default to 20 MB of disk
space for the file. You can increase the size and also if the file that you capture
will take more then this amount, the system automatically will allocate more space.
VCR Remote Control
Enable: Enable VCR remote control. Of
course, only if you have a compatible VCR with serial port interface. With the APAS
system, you can use any video source. So, it is not important anymore to control the VCR
from the computer. You just let the VCR or the camera run and you capture all the video
and then trim it to the point of interest.
Auto Start: Using this option the RealCap
will turn the VCR into play mode just before the capturing is started. This is used as
synchronization process between VCR equipment and Video capturing.
Type: Specify the type of VCR you have
connected to the computer. If this option is incorrectly setup the synchronization will
not work properly.
Serial Port: This option is to specify the
serial port to which the VCR is hooked.
Under the Capture option you have the Video Sequence and
the Preferences. The Video Sequence will immediately will start the
capturing of the video according to the Preferences options.
| The Preferences menu is shown here:
|
This Menu is extremely important since it determine the
way the RealCap program is going to capture the video. First you must select the Video
type whether it is PAL or NTSC. The PAL
mode allow capturing of 50 fields per second where the NTSC standard
allow 60 fields per second. Next is the Capture Time. The Capture
Time determine how long you are going to capture the sequence. As a rule of thumb, you
should always capture about one second before the view of interest and continue to about 1
second after the view of interest. This will allow you to have more frames than you really
need and then you can trim the sequence to exactly the frames to be selected for the
analysis. You should always Enable the time Limit.
The Delay represent the time in
Millisecond to delay the capturing. This is important factor to be used only when you use
a VCR. Since the VCR is a mechanical device, it take some time for the VCR to reach the
proper speed. Normally around 500 Millisecond. Therefore, the delay should be set to
around 500 to 800 MS. This means that you should take into consideration the time you want
to capture and add to it the mechanical delay.
Lets say that I want to capture 300 fields. Then I will
start capturing about 50 fields before and 50 fields after the activity. This will make it
5 seconds at 60 Hz make it 300 fields and to that I must add another second. So, the
capture time will be around 6 second. I would set the time to 8 seconds and be sure that I
have the sequence captured. Then I will trim it to the proper locations.
|
| Capturing Video
The actual capturing of the video is
very straight forward and does not require anything special. Use either the button in the toolbar
or the menu item Video Sequence under the Capture menu
and you are ready to go.
Attention: If there is no synchronization
between the computer (RealCap module) and the VCR/camcorder it is important to capture a
video sequence that is large enough to catch the whole sequence including some slack in
the beginning and the end. When storing the video sequence to disk the sequence will be
trimmed and only the images of interest will be stored.
Important: When capturing video to the
hard drive it was previous said that the through put to the hard drive has to be quite
high and if this is in some way slowed down the RealCap module will loose some
images/frames. If this is the case the RealCap will give a message that frames has been
dropped and the sequence can not be used. The first thing to do, is to increase the
compression. This will slow the frame data into the hard disk and will allow more time for
capturing.
The reasons for the drop of frames can be many, but one thing
that could cause it is the thermal calibration of the hard drive. A thermal
calibration of the hard drive is used to adjust the hard drive head for proper
positioning. Some hard drives do not do this calibration (special SCSI AV hard drives) but
they are pretty expensive. The use of a pre allocated file for the capturing process is
done to try to cover for the jumping of the hard drive head when capturing the video. In
some cases it is not always enough space to create a contiguous file and the jumping will
occur. To avoid this, try to defragment the hard drive using the Windows95/98 tool Disk
defragmenter. Also make sure that you have as much space available on the hard drive as
possible.
|
| VCR Synchronization When capturing a certain time sequence within a video it
is difficult to use the pause button and the PC keyboard to synchronize the capturing
process. Further it is important to ensure that most of the sequence of interest is part
of the captured sequence. Therefore a lager amount of video data has to be captured even
though in practice we can compensate for this approach.
Current VCR equipment using PC remote control is at the moment quite expensive and the
software provided with the equipment is also very bad.
At the moment the RealCap module supports the synchronization of the capturing process
using Panasonic VCR models AG-7350 and AG-5700. However, you do not need to purchase these
expensive VCRs. You can use a remote control of any VCR to capture any sequence. The
only difference is that you will have to give more attention to the trimming module to
synchronize the sequences.
The RealCap has just
been changed to be compatible with a modified Panasonic AG-2550C with a special
BCD-control interface. Support for other models are under development. In the future it
will be possible to synchronization the capturing using the playback
facilities of normal camcorder.
The basic idea with the VCR
synchronization is to automatically start the VCR before the capturing process from the
computer is initiated. To make this possible it is necessary to ensure that the VCR is in
proper speed before the images is put to disk. If the speed of the video tape is not
correct the first images will be incorrect/jittered. The time for the VCR to be in proper
speed is very individual. If for example. the VCR has been extensively used the friction
in the mechanical parts will be different from a brand new and the time for attain the
proper speed different.
To compensate for this time slack of
speed development it is possible to setup a delay from VCR play to capturing process
before the capturing start. Refer to the Capturing preferences for further details
on how to setup the Delay option.
Improve time for build up video
tape speed: When searching for the video sequence on the video that
needs to be captured it is wise to position the VCR in pause mode instead of stop mode.
The time for the VCR to build up the proper speed is much longer in stop mode than in
pause mode. This is due to the fact that the video tape is released from the video head in
the stop mode. The time for the VCR to grasp the video tape unto the VCR head is in many
cases more than 1 sec.
There is no way to check the time it
will take before the VCR has build up a proper speed where the images are OK. Try play
with the delay time and minimize it to about 500ms. This time is properly too small for
the VCR to build up the speed. Capture a sequence and use the trimming dialog to verify
the image quality.
After the Capturing process has been
configured with Auto start and the proper delay the capturing process is started in
exactly the same way as normal capturing process.
|
| Now you ready to capture video. If you
already have the capture card, go ahead and practice capturing. You have enough knowledge
to do it with no problems. However, if you do not have the capture card and want to
continue with the Trimming process, I already capture the files for you. I capture a
gait sequence from two cameras. The right camera and the left camera. The files names are:
Both of these files were compressed to save disk space. Therefore,
the resolution of the video and the markers reduced significantly, and therefore, it could
have a problem in the Automatic Digitizing mode since the contrast between the background
and the foreground is essential. Also, the video speed was reduce form 60 Hz to 15
Hz in order to save space. For demonstration of the digitizing process this is adequate.
Of course, in real life you would keep the 60 Hz and will not compress the files.
These two files are already trimmed and ready for analysis. However,
you should download them for latter use with the digitizing module. In the next
lesson you will have a download file that was not trimmed yet, and I will take you
step by step through the trimming process.
Make sure you taking a lunch or dinner or a nap while the computer download the files.
Both files will take around one hour to download. However, we will use these files to
digitize the data and synchronize the sequences in the next lesson.
|
|