I second Luiz and Minho. The article by Minho shows a very interesting way
of measuring to get a useful and even optimal sampling rate. That article
is a must for anyone operating optical mocap. I should clarify what I said
earlier, basically boils down to this in the case of optical mocap (such as
Qualisys): Record raw at 100MHz (or 120, 160 or even a little bit higher)
and then once you have that raw you can experiment downsampling to speed up
your computations. If you record directly at 10Hz you will most probably
get unusable data or your holidays (or your research assistant's holidays)
manually correcting data and as Luiz says, potentially adding more garbage.
By the way, happy holidays and happy labelling :)
Best wishes,
Juan Ignacio Mendoza
Doctoral Student
Music Department
University of Jyväskylä
website: http://users.jyu.fi/~juigmend/
On 21 December 2016 at 19:42, Minho Song <minho.song@xxxxxxxxxx> wrote:
Hello,
Just a comment for selecting a sampling frequency. Of course human
movements are not that fast and can be covered by the frequency below
100Hz, but it really depends on the situation like you have mentioned. In
most of the cases(if occlusion is not the reason), then gaps and
discontinuity of the trajectories comes because the sampling frequency is
low. And if the system is passive system, the frequency sometimes needs far
higher than you expect. Here is the link of our recent work about how
to select the frequency of the mocap system.
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0150993
Happy Christmas for everyone!
Best,
Minho Song
Minho Song, Ph.D.
Post Doctoral Research Fellow
Department of Musicology, University of Oslo, Norway
http://www.hf.uio.no/imv/english/people/aca/minhos/index.html
------------------------------
*From:* mocaptoolbox-bounce@xxxxxxxxxxxxx <mocaptoolbox-bounce@
freelists.org> on behalf of Juan Ignacio Mendoza Garay <
juigmend@xxxxxxxxxxxxxx>
*Sent:* 21 December 2016 17:13
*To:* mocaptoolbox@xxxxxxxxxxxxx
*Subject:* [mocaptoolbox] Re: Request for the mocaptoolbox
Hello
Just a comment. Of course all depends on the data and what you are looking
for but you might want to consider the following:
Most of the time when you are working with motion of the human body over
100Hz is not necessary at all. I usually downsample to 10Hz without getting
significant artifacts. Also by downsampling (or directly recording in 100Hz
or less) you don't need to filter over that frequency, so you can just
discard applying a downpass filter and you can just downsample by selecting
every n-th sample. By doing this (the n-sample method of downsampling,
avoiding filtering and the downsampling itself) you can dramatically reduce
computational cost. The reduction of computational cost allows you to test
algorithms that can later be applied in an online (real-time) processing,
opposed to Matlab's generic offline computation.
Best wishes,
Juan Ignacio Mendoza
Doctoral Student
Music Department
University of Jyväskylä
website: http://users.jyu.fi/~juigmend/
On 21 December 2016 at 17:29, Phan Giahoang <giahoang108@xxxxxxxxx> wrote:
Dear Prof Goh,
Sorry for my late reply as I am in holiday.
May I know what type of motion capture you are using? Is it active or
passive system?
I am using Phoenix VZ4000 which is active system and can reach max 500hz.
Yes, the toolbox is free and very useful. However, you may need to build
some filter algorithm for the acquired signal from motion capture. All work
in matlab.
Pls feel free to contact me if you have any queries.
Thank you and Regards
Gia Hoang
PhD candidate,
Robotic Research Center
Nanyang Technological University, Singapore
On Dec 15, 2016 2:55 AM, "Chung Goh" <cgoh@xxxxxxxxxxx> wrote:
Thank you for sharing your information.
Hope to get successful results in your Ph.D. program soon.
Many thanks,
Chung
*From:* Juan Ignacio Mendoza Garay [mailto:juigmend@xxxxxxxxxxxxxx]
*Sent:* Wednesday, December 14, 2016 1:48 PM
*To:* luiznaveda@xxxxxxxxx
*Cc:* giulia.gaggero@xxxxxxxxxxxxxxxxx; Chung Goh <cgoh@xxxxxxxxxxx>;
mocaptoolbox@xxxxxxxxxxxxx
*Subject:* Re: [mocaptoolbox] Re: Request for the mocaptoolbox
The Mocap Toolbox is free, so you can open the functions, hack them,
improve them, share them and make more functions.
You can use this list to share the functions and scripts that you have
made and any knowledge related to the Mocap Toolbox and to motion analysis.
I am posting regularly in my website (URL below my signature) some
software (Matlab, Octave, Pd) for motion analysis, which you can download
and use for free (hack, improve, share), some of them ready to use with the
Mocap Toolbox. I would be grateful if you give feedback on them.
Best wishes,
Juan Ignacio Mendoza
Doctoral Student
Music Department
University of Jyväskylä
website: http://users.jyu.fi/~juigmend/
<https://urldefense.proofpoint.com/v2/url?u=http-3A__users.jyu.fi_-7Ejuigmend_&d=DgMFaQ&c=e7TYJBzRfB0YbjEn2u3vBA&r=vYgz6a_NlKZBVH-bpuHcow&m=jEed_1u_61R4pwcZ9x0RRhG5zVoDrIPViL6YvZpJMrA&s=Ri-UhNWyUeqjfV1jWyqAFmsy37GolhKfqzD_DsP0AsY&e=>
On 14 December 2016 at 18:36, Luiz Naveda <luiznaveda@xxxxxxxxx> wrote:
Dear Goh and Giulia,
The mocaptoolbox is freely available for Matlab platform at
https://www.jyu.fi/hum/laitokset/musiikki/en/research/coe
/materials/mocaptoolbox
<https://urldefense.proofpoint.com/v2/url?u=https-3A__www.jyu.fi_hum_laitokset_musiikki_en_research_coe_materials_mocaptoolbox&d=DgMFaQ&c=e7TYJBzRfB0YbjEn2u3vBA&r=vYgz6a_NlKZBVH-bpuHcow&m=jEed_1u_61R4pwcZ9x0RRhG5zVoDrIPViL6YvZpJMrA&s=AO0yf0XLN0979wKgjBUJeu1JYxH0vp-su_WGiJEnrhk&e=>
.
The developers provide a very nice tutorial inside the software package.
However, you will need some training in Matlab and motion processing.
If you want to capture mocap data you will need a mocap system
(hardware). There are many mocap systems available from 5k to many
thousands of USD. All mocap systems provide outputs that can be processed
using mocap toolbox.
I am sure people in the list will have many other interesting
suggestions.
All the best
Luiz Naveda
http://naveda.info
<https://urldefense.proofpoint.com/v2/url?u=http-3A__naveda.info&d=DgMFaQ&c=e7TYJBzRfB0YbjEn2u3vBA&r=vYgz6a_NlKZBVH-bpuHcow&m=jEed_1u_61R4pwcZ9x0RRhG5zVoDrIPViL6YvZpJMrA&s=55lTkn-o8r9NeRwQdJdovODd6Jx1VCAbv4M3jdiB21Q&e=>
Luiz Naveda
_____________________________________________________
http://www.naveda.info
<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.naveda.info&d=DgMFaQ&c=e7TYJBzRfB0YbjEn2u3vBA&r=vYgz6a_NlKZBVH-bpuHcow&m=jEed_1u_61R4pwcZ9x0RRhG5zVoDrIPViL6YvZpJMrA&s=6JBm3rL0WRwA16aA1qMVV5PCFjadYMBjQctHV5pnrBE&e=>
^v^
^v^
^v^
^~^~^~^~^~^~^~^~~^~~^~~^~^~^~~~^^~^~~~~
^~^~^~^~^~^~^~^~^~^~~^~~^~~^~^~^~~~^~~~
On Wed, Dec 14, 2016 at 2:23 PM, Giulia Gaggero <
giulia.gaggero@xxxxxxxxxxxxxxxxx> wrote:
Hello Prof. Goh,
I'm sorry but I don't have this toolbox. I'm actually working on data
collected with this but I never collected them by myself. If you tell me
the name of the person that gave you my email address, maybe I can
understand better the person who he/she was thinking about?
Best regards,
Giulia
2016-12-14 17:14 GMT+01:00 Chung Goh <cgoh@xxxxxxxxxxx>:
Hello,
I am an assistant professor at the University of Texas at Tyler.
I just started a collaborative research with my colleagues in the
department of Kinesiology at UT Tyler.
So I would like to learn the mocap toolbox to use a motion capture system.
Could you please share the mocap toolbox with me?
Thank you for your kind consideration.
Chung Goh
