1,276 Reply by joan

(20 replies, posted in General)

jon wrote:
yvesrpt wrote:

Kinovea compatible with ANDROID to be abble to use it on E- Tablette?

Hi yvesrpt, It's a good question you asked. Kinovea requires a substantial amount of CPU processing to perform its tasks (especially if you dual record). I doubt tablets at this stage would have enough processing power to achieve this.

When I worked on Android (at work) it was all done in Java and we had a limit of 16MB (possibly 24MB depending on devices) per application.
Kinovea can only be ported to platform supporting .NET and with support for C# and C++/CLI languages (currently only .NET for Windows can do this, Mono cannot do C++/CLI).

The Linux / Mono port would have precedence over anything else anyway, but even that is pretty hard due to some dependencies on Windows (and the C++/CLI part).
Personally I am more inclined to work on improving the Windows version. There is a million things to keep me busy already.

I think there are efforts by others to develop video analysis applications on Android.

Go to forum: General

1,277 Reply by joan

(1 replies, posted in Bug reports)

Hi,
Yes there are some known issues with the MTS files. You may need to restrict the working zone until it triggers the "analysis mode" (pre-extraction of frames into memory). Depending on image size / frame rate, this can be as low as 2.5 seconds.

Generally speaking, the difference between forward and backward frame by frame comes from the way the images are encoded inside the file. Some codecs use intra frame coding only (each image is compressed separately from the others) others use inter frame compression, which means for us that a given image can only be reconstructed after we also decompress another image. When we go backward, we actually need to go back a few frames first (back to the last full frame), and then decode all the frames inbetween.

There is also issues with some containers that do not index the images (VOB), or index them in a bizarre way (MTS).

Bottom line is :
- Some files are more suited to going backwards in the same ammount of time than forward.
- To even out the differences you can restrict the working zone to trigger the pre extraction of frames into memory.
- You can convert the files to a more suitable format using such tools as AviDemux or HandBrake.

Go to forum: Bug reports

1,278 Reply by joan

(3 replies, posted in Bug reports)

Ha ! I had forgotten about that one smile
Well I don't know where to find a side view but if you can find one in SVG or other image format, you may import it.
I did find a bunch of anatomy files here and there (Wikimedia Commons), but there is no side view skeleton.

Go to forum: Bug reports

1,279 Reply by joan

(3 replies, posted in Bug reports)

What do you mean with "skeleton" ?

Go to forum: Bug reports

1,280 Reply by joan

(14 replies, posted in Bug reports)

Great !
Yes, some stuff has been done regarding memory in the saving routine. Initially to work around an issue for capture with large image size, but it should help with dual screen saving and regular save as well.

Go to forum: Bug reports

1,281 Topic by joan

(0 replies, posted in Français)

Sortie de Kinovea 0.8.15, la nouvelle version de votre outil open source d'analyse vidéo !

Désormais disponible en 16 langues, cette version inclut de nouvelles fonctionnalités, des améliorations pratiques, et des correctifs de bugs.

L'installeur est disponible ici :  Kinovea.Setup.0.8.15.exe


Parmi les nouvelles fonctionnalités (par rapport à la version 0.8.7) :
- L'écran de capture : visualisez directement l'action live de votre caméra dans Kinovea; enregistrez images et vidéos; ou appliquez un délai au flux live pour l'entraînement autonome.
- Références d'observation : superposez des dessins complexes ou des images sur la vidéo pour servir de cadre de référence.
- Export combiné : enregistrez la vidéo combinée de vos analyses comparatives.

Voir aussi :
- Une liste plus complète de changements
- Le manuel en ligne (Existe aussi en Anglais, Italien).
- Le tableau des fonctions (PDF) (Existe aussi en Anglais, Italien).


Avec l'introduction de la capture live, cette version est une étape importante pour le projet, et un grand pas en avant.
Comme d'habitude, les retours des utilisateurs de la première heure ont constitués une aide cruciale pour orienter le développement : Merci !

Go to forum: Français

1,282 Topic by joan

(10 replies, posted in General)

We are excited to release Kinovea 0.8.15, the new version of your open source video analysis toolkit ! smile

Now in 16 languages, with several new features, usability and functional improvements, as well as bug fixes.

The installer is available here: Kinovea.Setup.0.8.15.exe

New features (over 0.8.7) include:
- The Capture screen : directly stream live action from your camcorder into Kinovea; record images and videos; or delay the live stream for self coaching.
- Observational references : overlay complex drawings or images as motion guides on top of the video.
- Dual Export: save composite output from your comparison analysis.

See also:
- A more complete list of changes (What's new in Kinovea 0.8.15)
- The Online Manual (Also in Italian, French).
- The Feature list (PDF) (Also in Italian, French).

With the introduction of live capture this version is an important milestone for the project, and another leap forward.
As always, early adopters feedback has been crucial to drive the developement: Thanks !

Go to forum: General

1,283 Reply by joan

(6 replies, posted in General)

This thread is now closed.
Thank you very much to all involved !

big_smile

Go to forum: General

1,284 Reply by joan

(20 replies, posted in General)

- markers = stuff to stick or attach to the person to improve automatic tracking.
- feature window = the smaller rectangle around the point during tracking. This is the actual patch of the image that is looked for in the next frame.
- search window = the large rectangle during tracking, this is the area where the smaller "feature" window will be looked for.

Tracking tennis/baseball balls is most probably out of reach without super high speed camera (can do it manually if needed). But tracking a joint is definitely doable, although it may need some manual adjustments. Using physical markers might lessen or even remove the need for manual adjustement.

Using ping pong balls might be a good idea smile Cheap, round, and come in various colors. But how to attach them ?
Or maybe using balls from these beach racket games with scratch material (or is it called "velcro" I don't know the English term))

Go to forum: General

1,285 Topic by joan

(20 replies, posted in General)

I was asked by mail about markers and the tracking algorithm.

The automatic tracking in Kinovea works by computing the cross correlation coefficient between a candidate window and the feature window of the previous image.
For each possible position in the search window, we get a score, and the best score is the match (unless it is under a specific threshold in which case we assume the target was lost).

So it is a measure of how much the candidate look like the original.
- Contrasting area will make the matching easier. (marker should have a different brightness than its background. Bright on bright is not so good.)
- It is not invariant to rotation, so if rotating the marker changes its look, it will be harder to match. (triangle, rectangle, square: not so good)
- It will look in the immediate surroundings, so having a target that does not resemble any other part of its vicinity is better to avoid mismatches. (background and other parts on the person should be clear of anything that look like the marker).

To sum it up (based on the theory, I haven't done extensive testing in real conditions yet)
- Circular marker.
- Marker that takes about half the feature window.
- A color and brightness contrast with background that is not present in the rest of the search window.

Please post your findings, experiments, which markers work best, etc.
Anyone to do a video showing how various markers perform ?

Go to forum: General

1,286 Reply by joan

(5 replies, posted in Français)

gilles wrote:

Pour finir j'ai la confirmation que le boîtier Intensity Shuttle ne fonctionne pas avec les Express Card USB 3.0. Il faut de l'USB 3.0 en "natif" sur le portable.

Merci pour l'info, c'est bon à savoir !

Go to forum: Français

1,287 Reply by joan

(2 replies, posted in Français)

Effectivement la vitesse affichée est la vitesse du segment de trajectoire précédent.

Pour faciliter le calcul de la vitesse moyenne, vous pouvez peut-être utiliser un outil chronomètre synchronisé avec la trajectoire, et configurer la trajectoire sur « distance » au lieu de « vitesse ».
« Distance » va correspondre à la distance totale jusqu'au point concerné, donc si le chrono est démarré au même moment que la trajectoire, il suffit de diviser par le temps pour avoir la vitesse moyenne.

Go to forum: Français

1,288 Reply by joan

(6 replies, posted in General)

Thomas wrote:

Hello
I have done 94% in swedish.
Its only the hard ones left, don´t have any good translations of someones.

Fantastic ! big_smile
Please send it by mail (joan at kinovea dot org) so I can integrate the strings, add the menu, update installer, etc.

I don't know if Alexander had started something already for Swedish ?

By the way, for anyone reading this and secretely working on a translation, the ultimate deadline is next friday (May 13th). The official release should be on May 15th or 16th.
After that you can of course still send your translations, but they will be integrated in a later version.
Thanks

Go to forum: General

1,289 Reply by joan

(1 replies, posted in Bug reports)

Hi,
Thanks for raising these points.

Winkelmandje wrote:

What I'm trying to accomplish is that a video captured at 120fps is synchronized with another video captured at 30 fps.

Hmm, so you have one video in slow motion and the other in regular speed, right ?
Synching videos with different playback frame rate is supported and should be transparent, but I guess your 120fps video is actually played back at 30fps as well.

I wanted to ask you to set the speed sliders at the same value (20% for both for example) but they have been locked together for ease of use in the most common scenario !
(Maybe a future enhancement to think about. Something like ALT + speed change could be specific to the screen it's applied to and not to both screens).

Nevertheless, I will open a bug for this, and think about possible ways to fix this internally. It will most certainly not make it to the next release though (only few days away, and risky change).
Maybe the "normal speed" video should be locked to the slow motion factor corresponding to the other…

Winkelmandje wrote:

When Kinovea is not running and I'm using right mouse click on a file and selecting "open with" > Kinovea, this works just fine.
But when Kinovea is running this option doesn't seem to work. I only see a loading icon for a few seconds and then nothing happens, Kinovea doesn't open the file.

Yes, I have noted this a few days ago. Explorer is probably trying to launch a second instance of Kinovea, which then detects that it is already running and aborts.

edit:
bug 245 - Synchronisation between a high speed clip and a normal one cannot be done
bug 246 - Windows Explorer "Open with" doesn't work if Kinovea is already open

Go to forum: Bug reports

1,290 Reply by joan

(6 replies, posted in General)

Well, it is in the pool of new features that will be scheduled for integration during the next series of experimental versions.

Go to forum: General