1,351

(1 replies, posted in General)

I have set up a page to list observational references here : ObsRef Repository.

The rendering of the images in the browser might not work for some (all ?) versions of Internet Explorer. (It's a shame, I know, but I don't have the time to work out a solution right now, especially considering that it's working in every other browser I tried… hmm edit: works in IE9).

(note: you need to use at least version 0.8.13 to fully leverage the feature)

For the time being, you can send your contributions by mail if you want.
You can also post suggestions on new reference in this thread, if you cannot create it, maybe someone else will.

Thanks

yvesrpt wrote:

the clock tool, a circle with hours on it

This sounds like this could be implemented as an observational reference.
This reminds me that I wanted to craft some more "protractor" like tools there. (And possibly remove the few anatomy reference for now as they don't seem so useful as they stand).

yvesrpt wrote:

also a multidots graphics

Can you expand a bit? I'm not sure I see what you mean.

daww wrote:

link the spotlight feature with the tracking feature

Yes, it would definitely be interesting to have that. I think similar idea has been suggested with the magnifier tool as well.
Maybe you would add a spotlight, and simply right-click it and choose a "Track this area" menu.
It would have to work slightly differently than the existing track tool though to simplify the usage.

(this is for mid-term evolution, please add your thoughts to the discussion smile)

I am not entirely satisfied with the existing set of drawing tools. Some important tools are missing, others could be improved in their functionality and usability.
I've been thinking for a while about what to do about it, with the goal of having the most comprehensive and useful set of tools possible. On the other hand, the interface must stay as simple as possible.

I like to think of drawing tools in two categories : tools to observe and tools to show. I also call them, respectively analysis tools and telestrator tools.
(Telestrators are devices used in TV broadcasts of team sports, to explain tactics to the audience)
Typically an arrow or a circle are tools to show the viewer where to look or what to look for. These are tools used during a debriefing, when creating a video explanation, or similar "teaching" actions.
The observation tools (measuring a line, checking an angle, stopwatch, etc.) are tools that would be used during the actual analysis of the performance.

Here are some of the intended changes :

Organisation
- Better organization of the tools on the toolbar to reflect the tool usage (just sorting them differently).
- Separation of the arrow tool from the line tool. Lines and arrows serve different purpose, the fact that they are implemented by the same underlying code shouldn't be reflected on the interface.

Possible new tools
1 - Spotlight : a tool that would dim the entire image except its own area.
2 - Rotation arrow : a tool to symbolize rotative motion.
3 - Auto numbering mark : a number inside a circle. Each time you add one, the number increments.
4 - Simple rectangle.
5 - Multi line - Should also support measure on each segment and/or angle value at each joint.

Improvements to existing tools
- Circle changed into ellipse ?

Considering the number of tools and the limited place some grouping of tools might be needed.

Mockups for some of the proposed new tools
(spotlight, rotation, auto-numbering)
http://www.kinovea.org/screencaps/mockups/tool-spotlight.jpghttp://www.kinovea.org/screencaps/mockups/tool-rotation.jpghttp://www.kinovea.org/screencaps/mockups/tool-autonb.jpg

Comments, ideas for new tools are much welcomed, as well as improvements suggestions on the functionality and usability of existing tools.

1,354

(1 replies, posted in Français)

Bonjour,
La vitesse affichée est la vitesse moyenne du segment de trajectoire précédent le point. C'est à dire la distance du segment sur la durée entre les deux images.
La distance est la distance euclidienne classique : racine de [(xi - xj)² + (yi - yj)²] (où i et j seraient deux images consécutives).
L'intervalle de temps est calculé en fonction de la fréquence d'image de la vidéo.

Si x correspond a une vitesse dans votre formule, vous calculez la variation de la vitesse sur l'intervalle, donc plutôt une accélération ?

1,355

(2 replies, posted in Ideas and feature requests)

I agree 100% smile
Actually, there was once a "PDF export" function to export key images and comments  as a report, but it was later removed for various reasons.
Since then, I've added a more ambitious sub-project to the todo-list, to adress the shortcomings of the defunct PDF export function.
Ideally, it would be based on templates for maximum flexibility and easy customization, and the data be exported to something easily modifiable afterwards (like ODF).

There's a page on the developpement wiki about this sub-project here.
I don't find the time to fully get into this for now. It will later compete for priority with things like multi-point tracking, chronophotography and other effects, etc.
Also, lately I've come to think that this templates approach is maybe too complex/ambitious and a more straightforward (read: less flexible) approach should be preferred.

Existing reports done by hand would help in designing the feature though.

1,356

(2 replies, posted in General)

Hi,
You could use :
Manufacturer/producer: Kinovea open source project.
or just www.kinovea.org

And due to this open nature, there is no "country" of origin, there is no company or formal organization. It is the result of the contributions of several individuals from all over the place.

1,357

(3 replies, posted in General)

I've always had mixed feelings about this tongue
Over the years I got used to Kinovea not having a main toolbar… but well, I guess they are popular for a reason.

This will probably be tested in next experimental version, don't hesitate to speak your mind whatever you think.

http://www.kinovea.org/screencaps/0.8.x/toolbar.png

Reminds me that "series" of keys have been suggested for other functions too:
- shortcuts to activate each drawing tool.
- shortcuts to automatically create a key image with a specific name (with a manager to add/remove instances depending on user application).
- shortcuts to automatically create a label in the image with a specific text. (similar as above for storing the list of possible text).

Can there be a keyboard shortcut to select/switch screen? Something like an "alt+tab" combination in windows.

Page Up/Page Down is supposed to do that currently. It only works for dual playback though. (And the key mapping is not intuitive).
CTRL+Tab is generally used to navigate through tabs in an application. That would probably be a better fit…

1,359

(2 replies, posted in Cameras and hardware)

Also check the which camera do you use? thread, it has interesting information. (And everyone is encouraged to post his/her experiments with any device there smile)

Live capture with USB is possible only with webcams unless you have a very specific camcorder that can do it (webcam mode?).
In general it would seem the new camcorders (storing to SD card or DVD) cannot stream live to the computer through USB or only at reduced resolution.

One solution suggested is to use the analog output, and then hook to an Analog/Digital converter card on the PC (see the other thread).
It will be interesting to see if they release camcorders with USB3 streaming though…

1,360

(6 replies, posted in General)

(Note that the first 2 messages of this discussion are rather old)

Regarding the animated GIF front, they are indeed supported by .NET natively so there's hope for support in Kinovea for the future.
It will need some architecture work to integrate with the existing, so this will be for mid-term future (not before the next official release).

1,361

(6 replies, posted in General)

Animated GIFs would certainly be a plus. (Baseball community seem to use them quite a lot)
The decoding library doesn't support them as far as I can tell, but now that I think of it, maybe there is a way directly in .NET or another library).
Problem with GIF is that they don't have a defined frame rate, each interval between frames may be different.

Raw sequence of images would be a nice input too.

Awesome ! Thanks for the feedback smile
The problem with the second screen is general and exists even for different models.
When the second screen is created, it doesn't realize the first camera of the list of devices is already streaming to the other screen and it tries to connect to it. It doesn't fail, we just get a black screen…
Apparently it's a bit tricky to know if a given device is already in use by another application. However detecting that it is used by the other screen should be doable… (The hard part is not to break the automatic reconnection mechanism)

Yes, it makes sense smile

So,
CTRL + Key = next 1% spot.
Shift + Key = next 10% spot.
Key alone = next 25% spot.

Respectively Left / Right for the decrease/increase of the delayed during live capture.

----------------
Regarding capture and left/right reversal.
The idea was that when you press pause and the buffer is frozen, you would want to use the same interface for playing back and step-by-steping the recent frames.
I think the most intuitive way for playback is to have a left-to-right navigation, left being the start of the video, right being the end. (as in the playback screen).
So in that case it would be better if you can go forward in the video by pressing RIGHT.

Let say the user didn't apply any delay. Slider is all the way left. Left end is most recent, right end is oldest.
As soon as you pause frame grabbing, we will want to turn the slider over so that left is now start and right is now end. Maybe even force the cursor back to left.

At that point sliding right (with mouse or keys) should make you go forward in time. This will be the opposite of when live capturing where sliding right makes you go back in time.

(Maybe the reversal of the function will not be so obvious, I don't know.)

Yes I think the PS3Eye driver is one example of a driver that doesn't support 2 cameras connected at once.
(Unless using the SDK that the people at CodeLaboratories propose, but it means developping code specifically around these devices).

Ah oui désolé je me suis trompé smile
Cette fonction n'est apparue que dans une version ultérieure.
La version expérimentale est accessible depuis le sujet attaché en haut de ce sous forum. Ici pour la 0.8.13.