[n/a]
I received an enquiry yesterday from Scott Lucadou-Wells in Australia:
>I'm looking at purchasing a digital video camera and software for gait
>analysis in a private clinic. Does you have any recommendations
on
current
>technologies? Your response is much appreciated.
By a strange coincidence, I have today been asked to purchase one for
our own department, so I wonder if anyone has any recommendations or
comments? Here's my reply to Scott, in which I describe the main
criteria I am looking for:
Dear Scott,
To be quite honest, I'm a bit out of touch.
There's been a lot of developments lately, and I've been more concerned
with 3D systems, so I must admit that I wouldn't know what to buy myself
now. I want to know, though, because I suspect there are now some nice
devices out there. I also suspect that there are a lot of cameras that
sound great but are in practice of no use whatsoever.
Note, for example, that when a camera says it is "digital" this does
not
necessarily (and usually doesn't) mean that it records to a computer
file. It still records to tape, and if you want to view and
(more
important for us) edit the video, you must perform a conventional
digitization step, which is very time-consuming. So you need to make
sure that it records to a MPG (MPEG), AVI or (for Macs) QuickTime
(MOV)
filetype. I used a Hitachi MPEGCam in Hong Kong, which was fine, except
that I had a lot of trouble editing the resulting files - so you must
be
careful that the filetype is standard (try editting it in QuickTime
or
Adobe Premiere).
After that, it is worth checking the frame-rate possible and
shutter
speed. For motion analysis work, you will need 30 frames-per-second
(30
fps) at at least 320x240 pixels picture size. More importantly - and
I'm
afraid this is difficult to check without recording a test video
- you
need to make sure that the frame rate is consistent. The best way to
do
this is to record a video of a sports stopwatch with hundredths
of a
second. Step through the recorded movie and note the time of
each
picture (over a second or so) - the times should be roughly equal
(within a hundredth of a second). Many of the cameras I have looked
at
fall down on this criteria.
Another factor is the shutter speed, which needs to be less than
1/500th
second to prevent blurring during swing phase (perhpas not critical
if
you only want stance phase). Some cameras have a seperate "Sports"
mode
for this, often denoted by a running man icon.
The last consideration is the time of the recording. Some still
digital
cameras (e.g. Sony Mavica) allow you to record short (e.e. 15 s) clips
of video, which (assuming the above criteria are met) would be
sufficient for gait work. In my opinion it is better to record short
clips than long ones. But if you are wnating to record lots of videos
and analyze them later, you will want to have a recording capacity
of an
hour or so.
That's basically it - after you've done your research I'd be interested
in what you find! Make sure that you get a test video (of a stopwatch
if
possible, or at least someone walking) before you part with any money
-
I'd be happy (and interested) to look at it if you can email it to
me,
by the way.
Good luck!
Chris
--
Dr. Chris Kirtley MD PhD
Associate Professor
HomeCare Technologies for the 21st Century (Whitaker Foundation)
NIDRR Rehabilitation Engineering Research Center on TeleRehabilitation
Dept. of Biomedical Engineering, Pangborn 105B
Catholic University of America
Consider if you need the zoom/pan&tilt function on the video camera
Synchronisation (and/or Time Base Correction) as you might want to
add a
piece of equipment or another camera
Recording archive type: be it on CD ROM, S-VHS, Digital Tape, DVD etc.
Output Format compatibility
Be careful in chosing the camera model: because you might pay extra
hundreds just for a medical device label, also be direct with the sale
reps. as you might think you got a good deal but instead you are buying
old
stock models and could have got the newer model for a cheaper price.
These are just some of my thoughts, if people are interested I can
post my
article to this list (providing that I find it first).
Regards,
Tam Nguyen
(Mr) Tam C Nguyen B.Sc, B.Eng(Biomedical)
GradIEAust, AACPSEM, MISB, MSMBE
Biomedical Engineer
Hugh Williamson Gait Laboratory
3rd Floor, Main Block
Royal Children's Hospital
Flemington Rd, Parkville, Victoria 3052
Tel: +61 (03) 9345 6792
Fax: +61 (03) 9345 5447
Mob: +61 0412 503158
Email: nguyent@cryptic.rch.unimelb.edu.au
URL: http://www.rch.unimelb.edu.au/gait
I spent a lot of time a year ago trying to find a
digital high speed video camera for my clinic. All of
the digital video cameras, when sampling in "progressive
mode" stored 30 frames/sec, with the two fields per
frame being duplicates of each other. The only camera I
found that was different was the JVC 9500 (www.jvc.com)
which can record at 60, 120 (& can be pushed to 240)
fields/sec. (The 9500 may have been replaced with the
9800.) As a "consumer" camera they are relatively
cheap. I got mine for around $1,500. Ariel Dynamics
(www.arielnet.com) discusses
this camera and has
software to extract the 240 frame rate. I do not know
anything about their software. Without getting really
expensive it was the fastest camera available.
Paul H.
Paul Hansen, PT, PhD
FircrestPT@att.net
Fircrest, Washington
The company SIMI only works with digital camera's for their
motion capture software. These are quite normal camera's who are directly
connected to a normal pc with Firewire connection (which is
in most modern
PC's). The company is developing the software but they also sell the
necessary equipment (cameras, computers, calibration devices) on request
and
possibly could tell you which cameras are best suitable. They know
a lot
about the newest products available.
Best Regards,
Herman Boiten
Otto Bock OI
Research/Development
Lic. Mario Cleva
Universidad Nacional del Nordeste
- Facultad de Ciencias Exactas y Naturales
Departamento de Ingeniería Eléctrica
9 de julio 1449
3400 - Corrientes - ARGENTINA
Te: 54-3783-423126
- Instituto de Ciencias Criminalísticas y
Criminología
Catanarca 375
3400 - Corrientes - ARGENTINA
Te: 54-3783-422096
Home: 54-3722-476379
E-mail: clevamario@hotmail.com
clevamario@arnet.com.ar
I would suggest two types for consideration depending
on the intend use.
The first one is a consumer model from Sony that uses digital
mini video
tapes (between US $1200 and 2000). Images are of excellent quality
(20
to 30 % better than superVHS and may be moved using the firewire
(IEEE)
transfer and then edited on a PC or MAC with one of the available
software.
The other is using comercial equipment, Sony has a digital video
recorder (DS30) that uses digital mini video tapes or full size digital
video tapes. The image capture is superb in this equipment and has
terrific slow motion and freeze frame. It also has an indexing feature
that I find very useful. You need a camera and Cannon and Sony make
very
good CCD cameras to connect to the recorder. Best of luck
--
Alberto Esquenazi, MD
Director Gait & Motion Analysis Laboratory and
Regional Amputee Center
MossRehab
1200 West Tabor Rd.
Philadelphia, PA 19141 USA
Voice: 215 456 9470 Fax: 215 456 9631
Email: Aesquena@aehn2.einstein.edu
http://www.einstein.edu/phl/1214p2.html
A Member of the Jefferson Health Network
I would suggest the new JVC 9500 series (I think it has evolved
into the
9800) and the APAS software. The camera works very well (we
have four of
them: two 9500 and two 9600). They have made significant improvements
with
the design of the latest version. The cameras can now go from
60-120-240
Hz. As you increase the speed the capture area is reduced but
for the
price, and most applications, it is an excellent value. It also
has the
'firewire' link for easy and fast uploading to the computer.
Note that to use the 'high-speed' capabilities you need to use
the APAS
software. We do and have been very happy with it so far.
The capture and
trimming software they provide really is fast and easy to use, allowing
the
synchronisation of multiple views in seconds.
Daniel Benoit
Direttore Ricerca Scientifica
/\
Laboratorio di Biomeccanica
|\/ \/|
Let People Move
__|\ | | /|__
via GB Pontani-9
\
/
Perugia, Italy
<---------------->
06128
#
Ph. +39 (0)75 500 3956
#
Fax +39 (0)75 501 0921
dbenoit@magma.ca
Chris
There was a discussion on digital cameras for gait analysis on the
Biomech-l list, which I think you contributed to. Bill
Sellers
(wis@holyrood.ed.ac.uk) posted a list of responses which may be in
the
archive. I found the following site particularly helpful in trying
to
undestand digital video
http://www.adaptec.com/technology/standards/1394formats.html#top
It made me realize that my wish for a universal standard for video has
not yet come. In my naivity I thought that being digital it could
be
used on any of the standards wiith the appropriate D to A. That
is not
the case. The camcorders and vcr's record for VHS or PAL
or whatever
standard is used in the country.
Dear Jim,
I think I'm still pretty well up on digital video based on my
experiences in Australia in 95-96 developing my Motion Toolbox package
on the Mac. Actually, what has suprised me most was that there hasn't
been much progress since I last looked.
The whole "digital camcorder" thing was always a con, I think. People
just assumed that since it said it was "digital" that meant the computer
could play it and you could put on web-sites etc. In fact, it's not
really any advance on the old analog (e.g. Hi-8) video, since you still
have to install a board into the computer and run the signal through
it
to record it to file. This very time-consuming, and means that you
have
to use a computer which has been so modified. The only slight advantage
of digital over analog is that the quality should be better - although
I
remain to be convinced.
What is really needed, I think, is a camera which records directly to
file so that gaits can be recorded outside the lab and subsequently
transferred to the computer. Here we are talking about the MPEG4
standard (hardware compression - software compression is too slow),
but
there are not many of these on the market. I had one of the early ones
(I think the first) in Hong Kong - the Hitachi MPEGCam. It recorded
up
to 260Mb (i.e. about half an hour) directly to MPEG file on a PCMCIA
hard disk. It worked quite well, but I had a lot of trouble editing
the
resulting files with the result that I had to keep the originals -
I
basically ran out of space trying to store it all! I suspect there
is
some way to edit the movies (QuickTime doesn't work and crashes) but
I
never got around to finding it.
The new JVC cameras (the two I mentioned) also seem to record
to file,
but note that they are slightly more expensive than similar cameras
-
about $1300, and only record 20 second bursts. Actually, I think that's
probably an optimal duration for gait work, since the cycle in even
a
very slow gait is not likely to last longer than a few seconds. I don't
know how many of these files the JVC cameras will record - it will,
of
course, be limited by the storage medium. I also wonder how easy they
are to edit.
If anyone out there has a JVC GR-DVM90U
<http://www.jvc.com/mainprod.jsp?modelId=MODL006276>
or GR-DVL9800U <http://www.jvc.com/mainprod.jsp?modelId=MODL006255>
I'd
be very interested to hear of their experiences.
The only other point I'd make is that a lot of motion analysis software
these days incorporates real-time video recording along with the 3D.
For
example, Vicon uses a Broadway card for this, which records
to MPEG4.
Once again, though - recording would, of course, be limited to the
laboratory with such a card.
Chris
--
Dr. Chris Kirtley MD PhD
Associate Professor
HomeCare Technologies for the 21st Century (Whitaker Foundation)
NIDRR Rehabilitation Engineering Research Center on TeleRehabilitation
Dept. of Biomedical Engineering, Pangborn 105B
Catholic University of America
I am trying to steer our lab towards storing data digitally (in the
UK
the transfer to Electronic Patient Records is a preferential
area in the
National Health Service).
At present we use 5 stationary VHS video cameras which are connected
to
a 2 channel mixer in pairs. The split screen images are recorded on
VHS
tape. Obviously we can't throw away the existing 5 analogue cameras,
but
I am looking for a way to store the video digitally.
I can capture the analogue video with an ATI card and then convert to
MPEG, but that's very time consuming.
I know of MPEG boards which capture a composite video signal (output
by
the mixer). Does anyone on the list have experience with such boards?
Would buying a video camera be of much use in this case?
(I have searched the CGA archives and Biomch-l, but I haven't found
a
direct answer to my query.)
Gabor
--
Dr Gabor Barton MD
Clinical Scientist
Gait Analysis Laboratory
tel: +44 (0)151 252 5949
Alder Hey Children's Hospital
fax1: +44 (0)870 052 1935
Eaton Road, Liverpool, L12 2AP, UK fax2: +44 (0)151 252
5846
I made some enquiries about digital cameras a few weeks ago. I was
looking for a camcorder that would record direct to MPEG, to avoid
the
time consuming digitizing process that you mention.
When I was in Hong Kong, I had a nice little machine made by Hitachi
called the MPEGCam <http://www.mpegcam.net/m2>
which recorded up to 30
minutes of MPEG videos (as well as JPEG stills and audio). I tried
to
buy one for my lab here only to find they have been discontinued. The
replacement is apparently the Hitachi DZ-MV100A
<http://www.hitachi.com/dvdcam>,
but this records to the DVD Video
Recording Format (new specification for real-time recording on DVD
discs) and provides 30 minutes recording time (1.4 GB). I'd love to
hear
from anyone who has used this to see if it is suitable for motion
analysis work.
The othe rpossibilities which seem to record direct to MPEG are the
Sony
DCR-PC110
<http://www.sonystyle.com/sonystyle/4784/4785/4792/7783.default.html>
and the Canon XL1 <http://canondv.com/xl1/index2.html>.
I must admit that I became completely bewildered by the exercise in
the
end, because the sales staff of all these companies don't seem to be
able to answer any questions about the recording format, frame rates
achieable, etc. Sign of the times, I'm afraid.
I'm sure there must be a nice camera out there somewhere which is
capable of recording good MPEG files, but as yet I haven't found it!
Chris
--
Dr. Chris Kirtley MD PhD
Associate Professor
HomeCare Technologies for the 21st Century (Whitaker Foundation)
NIDRR Rehabilitation Engineering Research Center on TeleRehabilitation
Dept. of Biomedical Engineering, Pangborn 105B
Catholic University of America
Want to know more? Email the CGA list!
[n/a]
