OpenSesame runtime for Android

nsriram

Sorry for the delay. The Android runtime works on a 10" Galaxy tab 3 from Samsung. When it starts up the /sdcard folder maps on to /storage/emulated/0 which is fine. Using the "My Files" app the location of the 3 demo experiments (gaze, affordances, and attention capture) seems to be a mystery as I am not able to find it using the search in the file explorer app. It says that it is getting it from an "examples" folder that I cannot see. Also a search for tar.gz or gaze or any reasonable terms ends up empty. I do not wish to root the device but any clues as how the run time defaults are working?

When I have my own tar.gz studies I guess the thing to do would be to place them under /sdcard/mystudies (which would map to /storage/emulated/0/mystudies).

Other notes: the gaze and affordances are both 2 choice studies and they work well. The attention capture seems flakier (find the odd shape among circles and squares) and strange things happen (like trials auto-forwarding).

other questions: I assume only a number can be given as a subject ID. Non-numerics will not work correct (although one can edit the name of the CSV file)?

psychofs

It worked fine with HTC Evo 3D, with some Samsungs, of which I don't know the exact type. But with Samsung I'm having the problem, that the SD-Card often has a different path. I actually want my participants to run an experiment several times a day and it would be quiet annoying to always type in the path. Is there a way to change the default path?

sebastiaan

@nsiriram

Using the "My Files" app the location of the 3 demo experiments (gaze, affordances, and attention capture) seems to be a mystery as I am not able to find it using the search in the file explorer app. It says that it is getting it from an "examples" folder that I cannot see.

The example folder is part of the app itself. I don't know exactly how this works technically, but I suspect the folder is extracted to some temporary location when the app is launched. But long story short: You cannot access this folder. You can find the examples here, though:

• http://osdoc.cogsci.nl/examples/

When I have my own tar.gz studies I guess the thing to do would be to place them under /sdcard/mystudies (which would map to /storage/emulated/0/mystudies).

Sounds good.

other questions: I assume only a number can be given as a subject ID. Non-numerics will not work correct (although one can edit the name of the CSV file)?

Right, the subject number is always a number, but you can choose whatever name you prefer for the logfile.

@psychofs

Is there a way to change the default path?

No, the path defaults to /sdcard/Experiments/ if it exists, and /sdcard otherwise. It would be very atypical if /sdcard does not actually point towards the SD card though, although there are usually a few synonyms as well. What path is used for the SD card in your case?

If need be, you should be able to create a symlink from the SD card folder on your device to /sdcard/Experiments. I don't have an Android device to test, but if you install a terminal app, you should be able to do something like this:

ln -s /my/folder /sdcard/Experiments

Cheers!
Sebastiaan

psychofs

Thanks so much, Sebastiaan! Really saved me some work
I only realized that many Samusng devices use paths like "/storage/extSdCard" or "sdcard0" or similar, but you're right, at least on one device I tried, the default path worked fine, too.
Unfortunately I'm not good at programming and don't have an own Android device at the moment either, so I couldn't try out the symlink or a terminal app.

Next problem, I'm suffering from, is Android 4.4. Again, it's so far only based on one observation, but I cannot find the experiment even after changing the path to the path that shows up in a file explorer app like "my files" on Samsung. Most likely this is due to new restrictions under Android 4.4 that allow apps only to read and write in their own folder on the SD-Card that is created when installing the app. In other words the runtime environment most likely has no more read and write access to the SD-Card parent directory and to an "Experiments" folder. I told a friend to try to move the experiment to an "experiments" folder and to an "OpenSesame" folder but according to his report he coulnd't get any solution to work fine.

sebastiaan

I cannot find the experiment even after changing the path to the path that shows up in a file explorer app like "my files" on Samsung. Most likely this is due to new restrictions under Android 4.4 that allow apps only to read and write in their own folder on the SD-Card that is created when installing the app.

My test devices (Nexus 4 and 7 with Android 4.4.2) only have built-in storage, so I cannot really say what will happen with an external SD card. If this is indeed an issue (which would be annoying), then you could probably work around it by copying the experiment to the device's internal storage. I'd have to look into this to be able to say more about it though.

psychofs

Today I started my study and had 5 participants.
Which caused some trouble was a Samsung Galaxy running Android 4.3. Again putting the experiment on the external card lead to not finding the experiment, but the internal storage (we just used the download folder and she now has to set the path each time, 49 times :P) worked fine. So for a future update of the app, I guess giving users the option to change the default path permanently would be the best solution.

Other older devices and Android versions worked fine. So I'm really happy about getting my Ecological Momentary Assessment started.

(Just having another problem with OpenSesame on a PC, but I'll start a new thread for that)

Thank you so much for your support!

psychofs

Ok. I finished data collection for my thesis.
Summary concerning the path-problem: I just had to figure out, that the default path /sdcard points to the internal storage in Samsung devices (no matter if an external SD is put in or not), and to the external SD Card in other devices like htc. And actually in the end it seemed to work fine just on any device and with any OS version so far, once you've figured out these little dirty tricks.
To give some more information: I was running an ecological momentary assessment. 40 participants, studying a realtionship betweend mood and appetite or preference for certain foods and snacks. Since Open Sesame does not support a function to prompt the participants I imported automatically generated .ics-files into their calender using another app and participants just had to run open sesame app on their own.
On older devices the app was kind of laggy and using an "ok" button participants sometimes skipped some questions and produced several missings.
So in case you'd like to support a further usage in EMA-studies it would be nice to implement horizontal scales plus an option to force an answer plus an ok-button so that people still have the possibility to change their choice, especially because the radio buttons / check squares were really really small when scaled down for some devices with small screens and you could hardly see which one is selected.

Finally concerning results: nothing important so far, nothing supporting my hypothesis.

Anyway, again: Thank you for your support and thus enabling me to run an EMA as a Masters Thesis under a GNU licence and still having a great support in any case! THANKS!

sebastiaan

Anyway, again: Thank you for your support and thus enabling me to run an EMA as a Masters Thesis under a GNU licence and still having a great support in any case! THANKS!

You're welcome!

timo

Hi!

I am currently considering to use tablets with Android or Ubuntu for data collection.
As i am working with latencies, i am interested in your experience with the Android App on Nexus Tablets (4, 7, 10).

Can i assume the same behavior for tablets and PCs? I suppose there's not such a sophisticated benchmark study like in Behav Res Methods on tablets?

Thanks in advance for any helpful comment!

timo

sebastiaan

As i am working with latencies, i am interested in your experience with the Android App on Nexus Tablets (4, 7, 10)

I use the Nexus 4 and 7 (2012 version) for development, and they work pretty smoothly.

Can i assume the same behavior for tablets and PCs? I suppose there's not such a sophisticated benchmark study like in Behav Res Methods on tablets?

Aside from some minor limitations, described here, your experiment will run exactly the same on your tablet as on your computer. I don't have any specific benchmarks about the timing. However, I've used it successfully for some reaction-time experiments. And @jdenhartog, a technician at the VU University Amsterdam, has done some quick testing which showed pretty good performance. I'm not sure whether and how he followed up on this though.

Cheers!
Sebastiaan

Locke

Hey,
english is not my mother tounge, but i try to do my best;D

i do have some quite similar Problem,
i try to get my Experiment to run @ a Samsung galaxie Tablett 2 with Android.

Open sesame statrts his Examples,
but it does not start my own Experiment,
doesent matter where i put it;
I tried to put it into:

/storage/extsdcard/Experiment
/Storage/extsdcard/
/sdcard/

the above describtet file (/starage/emulated/0) i cant find on the Tablett

i got this, this Debbuging Report:

Launching resources/android/menu.opensesame
openexp.sampler._legacy.init_sound(): sampling freq = 48000, buffer size = 512
experiment.init_log(): using 'resources/android/defaultlog.csv' as logfile (utf-8)
experiment.run(): experiment started at Wed Nov 26 11:57:07 2014
opened /sdcard/opensesame-settings.txt for reading
experiment_folder = examples
logfile_folder = /sdcard

sebastiaan

Hi @Locke,

Based on your debug output, it looks like the /sdcard/ folder is accessible, so putting your experiment there should work. Are you sure that your experiment is there, i.e. by checking with an explorer app? And have you indicated the correct experiment folder (i.e /sdcard/) when OpenSesame starts up?

Cheers,
Sebastiaan

sebastiaan

I have a question for other Android users who are running Android Lollipop 5.0. Since I upgraded my Nexus 4 to Lollipop, the virtual keyboard in the OpenSesame runtime for Android (2.8.3) doesn't respond anymore. It's visible, but doesn't accept any text input. Does anyone else suffer from the same problem?

PunitShah

Hi Sebastian, Hi all,

OpenSesame experiment runtime runs quite nicely on the OnePlus One (Android CyanogenMod - 4.4.4). I will be messing around with getting mousetracking software running on the phone/other android tablets - so far it looks promising.

re your question about the virtual keyboard being unresponsive in Lollipop 5.0. I had a similar issue with mine, and found that installing 3rd party keyboard apps (e.g. Swiftkey) can help get it working - perhaps give that a try?

Sebastian, thanks for your work on the project - well done.

Punit Shah

sebastiaan

re your question about the virtual keyboard being unresponsive in Lollipop 5.0. I had a similar issue with mine, and found that installing 3rd party keyboard apps (e.g. Swiftkey) can help get it working - perhaps give that a try?

Thank you very much, I'm definitely going to try this out! If this works, it's a least a workaround until I find a proper solution.

sebastiaan

re your question about the virtual keyboard being unresponsive in Lollipop 5.0. I had a similar issue with mine, and found that installing 3rd party keyboard apps (e.g. Swiftkey) can help get it working - perhaps give that a try?

Indeed, I can confirm that this solves the problem. It must be due to an unfortunate interaction between OpenSesame and the new Android virtual keyboard then. Well, at least for now there's a workaround.

rino

Hi,

will there a update for Android 5?
I have trouble playing sound on Android 5.

Bests
Rino

sebastiaan

Hi Rino,

The latest Android package, while not up to date (still at 2.8.3) should work on Android 5. Sound should work as well, although the synth item is not supported. Just to ask the obvious: Do you hear sound when you run the experiment on your PC? And is the sound of your phone/ tablet turned on?

Cheers,
Sebastiaan

rino

Hi Sebastian,
thanks for the reply.
Yes, it works on my desktop computer.
And yes the sound is turned on.
I'm try to play a wav file.
On a xperia z1 with 4.4 the experiment is working.
Bests
Rino

sebastiaan

Hi Rino,

That does sound like a problem then. I'm afraid there's little chance of a fix in the near future, unless this is fixed in the PyGame subset for Android. This is the library used by the OpenSesame runtime for Android, but unfortunately it's not very actively developed anymore (although for the moment it does work quite well on most the devices).

Cheers,
Sebastiaan

panos_uol

Hi all,

here is an issue I have as I intend to use an open sesame experiment build for androids in an EMA study, were participants will complete the experiment in their mobile phones on their own time.

Before the experiment starts, participants have to identify some paths and also insert a subject number. My participants will have to do the experiment several times per day, so I have to instruct them to give a different subject_nr every time to avoid overwriting. However, this will be a bit of a hassle for them. Is there any way around it? Is it possible open sesame to automatically assign a different subject_nr every time without participants entering that manually?

Thanks a lot!

sebastiaan

Hi @panos_uol,

To do this, you would need to build your own .apk package from the source code. This is actually not that difficult, but it will require some knowledge of Linux. Basically, what you would need to do is change the Android script such that the menu is bypassed, and your experiment is called directly. You could take a look at the Android build script to see if you think it's worthwhile to do this:

• https://github.com/smathot/OpenSesame/blob/master/setup-android.py

Cheers!
Sebastiaan

panos_uol

Hi Sebastiaan!

Sounds a bit challenging for my programming abilities, but in any case I can have a look and see what I can do.

Many thanks for your time!

tkesnerus

Hi Sebastiaan,

the Android runtime worked fine on an "Archos 101g9"-Tablet (Android 4.0.3).

On an "Acer Iconia Tab 10" (A3-A20-HD, Android 4.4.2) it is working, but the soft navigation keys do not disappear during the experiment, the screen has black margins and is scaled down.
Here's a screenshot of our experiment:



I think it's because of the immersive mode of Android API19+.
Do you have any idea how to solve this issue?

sebastiaan

Hi @tkesnerus,

On an "Acer Iconia Tab 10" (A3-A20-HD, Android 4.4.2) it is working, but the soft navigation keys do not disappear during the experiment, the screen has black margins and is scaled down.

I've heard that you can remove these soft buttons with external apps. This is not specific to OpenSesame, so I would just search for general tips on how to hide soft buttons. If you find a way that works, please share!

Cheers,
Sebastiaan

tkesnerus

Hi Sebastiaan,

thanks for your quick answer.

Edit: I had the same problem with Hannspree Hannspad SN1AT71 (Android 4.1 and 4.2.2).
The best result for me gave "full!screen"-App which needs root access.
It modifies the navigation bar, so that the navigation keys appear in the bottom corners by default and can be configured.

Julia

Hi all!

I am trying to run a touchscreen picture selection study (participants see a picture, listen to the sentence and have to choose the right guy in the picture) on my Android 4.2.2. tablet pc. Open Sesame did manage to detect my experiment on the SD card (internal storage). However, when I open it and insert the participan't number and the name for the csv.file, nothing more happens, the programme just quits. (everything works just fine on my laptop). Has anyone had this problem? How can one go about it?

Julia

sebastiaan

Hi Julia,

You will find debugging output in the file /sdcard/opensesame-debug.txt. Based on the error message in that file, you should be able to tell what went wrong--an experiment can crash for any number of reasons.

• http://osdoc.cogsci.nl/getting-opensesame/android/#debugging

Cheers!
Sebastiaan

Julia

Hi Sebastian!

I have now figured out how I can get in here from my tablet PC to actually be able to post my debug output here. The debug-txt is saying that I actually have finished the experiment though I haven't even started it. Is it the problem with my script or with my device?
/sdcard/opensesame-autorun.yml
/mnt/sdcard/opensesame-autorun.yml
Launching resources/android/menu.opensesame
openexp.sampler._legacy.init_sound(): sampling freq = 48000, buffer size = 1024
experiment.init_log(): using 'resources/android/defaultlog.csv' as logfile (utf-8)
experiment.run(): experiment started at Wed Apr 15 16:45:08 2015
opened /sdcard/Experiments/opensesame-settings.txt for reading
experiment_folder = /sdcard/Experiments/
logfile_folder = /sdcard/Experiments/
opened /sdcard/Experiments/opensesame-settings.txt for writing
experiment.run(): experiment finished at Wed Apr 15 16:45:20 2015

Julia

Issue solved: So I really had the trouble with the script.
First, my tablet pc had some trouble reading .jpg and .wav files. That's why everything crashed in the first place. I had to convert them to .png and .ogg respectively.

A second issue that I had was that I actually indicated the file format twice in my script. I wrote filename.png into the block_loop and [object].png when defining my variables. When I got rid of that, everything worked just fine!