I am at the beginning stages of a project in which I will be trying to make a hearing aid application for Android. I have wrote a few patches in Pure Data,C sound, and the basic Android sound library which basical take the input from the microphone and play it through headphones. No filtering or amplification.
While Csound gave the best performance, the latency made the tools unusable. I know Android L is suppose to help, but my goal is to create a low cost device hearing assist device. So older phones probably won't get it.
The next idea is to see if I can access the adc and dac values directly, then use C to make my own versions of AudioTrack and Audio record by using the NDK. Basically pointing to the places in memory where these values are coming in.
Is this possible? Also what should I be researching? I can't find anything online about accesses the DAC and ADC directly.
Thank you for your time.
No. "Android" does not provide for direct access by apps to hardware at all.
The NDK does not change that, as you still lack permission to the audio hardware device nodes.
If you have a particular device on which you can install a customized build of Android, then you might be able to do something by adding new APIs or somehow giving your app or a special unix group access to the hardware nodes. But the details of how you might utilize that access would depend on the device chosen.
Related
I am working on an android project where I use a phone to power an external device. Through a USB-OTG connection, I can power the external device.
I want to know if I can write a program on android studio where I can press a button in the app to toggle the power to this external device (turn it on and off).
I've seen it done before on an older Samsung Galaxy 3 phone (on a much older version of android) by updating data in a file; specifically this file:
/sys/devices/virtual/host_notify/usb_otg/mode
But I'm trying to write the software so it is more generic and can work on any new android device.
Is there a way to do what I need?
I've tried finding a generic way to do this but it doesn't seem possible because the only way I've seen it possible is by using shell commands to toggle exposed kernel bits through root files.
I have concluded it's likely impossible to make generic code for this as most android phones root files are organised differently. So you need to discover the specific file each time you want to use a different phone for the software.
I don't think it would be feasible for myself to create a database of different brands of phones root file structures because new phones come out all the time with new structures which would require too much maintenance on my behalf.
1\ Regarding to Android Source Project hosted by source.android.com, Android supports multi-camera functionality. Also we have some information about how a camera driver really works in Android.
2\ As you know there is an old project called v4l2loopback that you can install and run on your Linux (it's simple if you have root access). For example you can also have fun with ffmpeg to stream a video as a dummy webcam device. (v4l2virtualdevice_android)
Qust\ What I'm really watching for is simple in logic and hard to develop. I want a solution that my Android application (that is provided as an APK) will be able to create a dummy camera and I can feed that from another source (not really important; maybe a video file). So user will be able to use original camera app, Instagram app (Live), etc ... to record the video.
I don't really know if it's necessary for device to be root or not. And if it's possible how can it be done using java or NDK?
You can do that if you replace the system camera HAL. It's quite possible if you build a custom ROM. Having root access on a known device is essentially same as having your custom ROM. But this doesn't help if your APK is installed (with root access) on a different device, even if the changes are minimal. Sure you can prepare logic that will work on many devices, but each will require separate consideration and testing.
Is it possible to get the user behavior on the phone (for example Alpesh has an Android phone and he uses multiple apps, browser YouTube etc). Whatever he is doing on the phone I want to get all those things from behind (which apps he has installed, which app he opens and what he search on the phone, All these data I want to get programmatically so what all can be get in android).
For now I am aware that installed apps list can be get easily but I want to get usage history and what he do all on mobile.
This is not a code solution, but an answer to your question, so you can get start some where.
In my opinion your question title are asking about two things.
(part 1) Getting User Behavior on the Android Phone (part 2)(App History, Browse
History etc)
1- First part Getting User Behavior on the Android Phone:
There is a concept called context awareness. Short described; it is about gathering different information from the phone, like light sensor, motion sensor, sound, location or even user behavior etc. and depending on your app requirement and the gathered information:
You could send these information over cloud data store for statically usage
You could make your phone doing (behavior) different things depending on location, motion or what ever.
etc.
For context awareness it is an open area for pervasive computing research. And it is not just few lines of code to write, it is typically a complete solution depending on requirement. Example I have built a context awareness application to gather noise collected by phones from different locations for research purpose inspired from this framework, but I am pretty sure you can find other frameworks or even build your own, as I did in my case.
The mentioned framework has some examples.
2- The second part is about App History, Browse History etc.:
This is possible, but you still need to build a peace of software (App) to collect all these information (logs) from the phone. Hereafter you can make phone act on different conditions and/or again send it over a RESTful API over cloud service data store, there is no limit for it.
The problem is, there is no thing out of the box for your requirement. Even if you find frameworks you still need to research it and further work on it.
You can find different examples for your requirement, like to collect browser history, you can find SO question here:
Get browser history and search result in android
Or get list of installed application:
How to get a list of installed android applications and pick one to run
My point here is you need to solve small goals at a time and put your knowledge together at the end.
Both 1 and 2 can also be related to each other, depending on your achievement.
Conclusion
Make a goal to your project.
Define the main requirements and tasks of your project.
Research your options (Technology, Cost, Target Audience, What data I can or I should not collect, what is possible to collect, what is the limits, Privacy issues etc.).
Split your project in small assets and try to solve small problems/goals.
Finally you would be able to put the puzzles together and build your final application
but i want to get usage history and what he do all on mobile
This is not possible and shouldn't ever be possible. Each app is sandboxed by Android so apps cannot inspect what other apps are doing. Think about it, you wouldn't want apps to be able to intercept private information such as banking details.
Every app is isolated from the other ones. Unless you develop a system signed app, you will not be able to gather all that data.
What you could do is to develop your own Android Rom where you then develop your data collection the exact way you want. Then you need to distribute your rom, which is another story...
I am newbie in android developing, I have searched for this question but i didn't find my answer.
I want to know is there any ability to edit the sound calls in android?
I mean i want to add noise or change sounds of caller, Is it possible to change the sound in calls or adding a new sound to it?
TL DR : The answer is not yet.
And it's not like we've been waiting. The first entry i can find is from July 31 of 2009, the issue #3434 and, as of today (May 13 of 2015) it's still has not been assigned.
It's really hard to actually work on low-latency project, audio recording and of course, voices changers when you can't do low latency.
Not to say there's hasn't been any workarounds, you could emulate yourself the call, and add the voice effect sure (build your own dialer, and work with that), but let me warn you : you probably won't have any good perfs when it comes to real-time appliucations. No low-latency means no efficiency when it comes to audio recording.
You'll have to wait then.
Your question can be resolved partially depending your using model. the premises are :
you just want to eject some noise into your outgoing audio stream,
not into the incoming audio stream.
you may use a third-party VoIP application to make the phone call.
or simply say, you just want the peer to hear some modified voice. it is feasible.
Normal a native phone application on Android platform uses "Android audio system module" in the framework, the vendor provided audio libraries and Linux ALSA audio libraries to transmit/receive the audio data. These .so and .a files are under the read-only mode normally and could not be overwritten by user, so you can not inject data into this data chain.
But you have more capability to manipulate the data if you use a VoIP application to make the phone call, some VoIP applications can give a real phone number, like Fongo, you can receive a phone call to that number, the caller does not know you are using a VoIP application to speak.
So if I was assigned to do this project, here are my steps:
find a usable and open sourced VoIP client on Android.
find the code to sample the audio data from microphone ,add the code
to manipulate the raw PCM data and send the result to audio encoder.
build and run it on Android
register or apply a phone number for this VoIP client.
done.
Hope it help
I have a great idea for an Android app, but as I'm only familiar with php/js, I'm uncertain of which approach I should choose for creating it. The app will be based on a google map with a lot of position markers. There won't be any fancy animations or other heavy resource-demanding activities.
As I see it there are three different options:
Read up on Java and program the whole thing in Java
Create the map activity in Java as a mapview and then use webviews for the other activities (which can easily be scripted as html5 webpages.)
Script everything as a webapp (not really an option, as this is not a real mobile app imho.
I'm most keen on using no. 2 as I'm quite familiar with html/php/js/mysql. Have to read up on the html5 specifics, though. Questions:
I need access to GPS and camera hardware. Is that acheivable in webviews?
How complicated is it to pass variables between js in webview activities and java in other activities?
How big a difference in performance can I expect if I use option 1 vs option 2?
Other thoughts?
Kind regards,
Anders
You can choose number 2, but as we are talking about an android phone, you might want to get really accurate coordinates for your map, and you can only achieve this by accessing your phone GPS, through webviews the best you can get is the location trought the device internet IP adress, wich doesnt lead to a very accurate geo position.
The best choice is a 100% java application in my opinion.
1) Yes it's possible, but as commented it will be less accurate and probably slow.
2) Not complicated. Painful if you need loads of interaction between a webview and native app. Using a Javascript Interface that can be set up from the native app. You can basically inject javascript in a webview's html.
3) Heterogeneity of performance depending on device. Because your implementation will be based on the device's browser you can expect to get really sluggish behavior for older devices. Anything to do with HTML events (Dragging, Tabbing...) will have a knock on most devices, from my experience.
4) As #vodich comments there are other party frameworks. My benchmarking on PhoneGap and other js-based options is that they're a waste of time if you are looking at developing a professional app. I haven't developed on Adobe AIR but find a pain the need to be installing plugins to get native functionality (access to sensors, camera, etc) Mobile is all about fast, responsive behaviour. HDI is your finger, user is fast, so app needs to be fast.
EDIT: So hell yeah! Java FTW!
Albert.
4.Other toughts?
Yes, if you really want to make a great Android app, you should be using only Android and specific Android UI components, and give it a native look and feel. And regarding 1,2 yes it is possible, I would say not so complicated to just integrate them, but I think you'll eventually get in big problems.
Learn Java and write your application natively.
Webviews might allow you to use your php skills to present something to the user, but it's entirely one-way - you'll not be able to interact with what's inside.
The Android developer site offers fantastic documentation and jumping from PHP to Java isn't greatly difficult, though you'll need to get used to strict typing and "real" OOP.
Other thoughts? Don't go down the PhoneGap/Cross platform toolkit road - it might allow you to write applications for multiple platforms and using your current skills, but in the end you get a subpar app that doesn't feel right on either platform and doesn't fair well as future versions of iOS and Android are released.