gvsumasl / echoprintforandroid Goto Github PK

http://www.mooma.sh/api.html

Unfortunately they will stop their service.
Good for us they released their fingerprints. Got some RPI2 laying around. Maybe I can use them as an EchoPrint server.

Rather than issue,it is a humble question for you? First of all great way of generating Code from sound files and great tutorial. I have an Android project where I need to detect a smoke alarm sound and for that I am planning to use your implementation. You are sending the generated code to the server to get music track files and information. But,instead how can I use the generated code to compare sound. Meaning ,I am planning to save the generated code of smoke alarm. So next time I will be supplying audio data to the Codegen and whenver the similar code (saved smoke alarm code ) is detected,I assume it will be an alarm sound. So my question is how can I effectively use the generated Code to find the similarity between audios.

P.S. I am not so good at C++ and it will take me more time to go through all the classes and its implementation.

Here is what is working for me

• I have fixed the app and I can build it in Android Studio.
• I see codegen generating the hash

But the app crashes..

• Initial suspect was Memory leak , but when i tested it with leak canary nothing came up .
• Next I looked through the app logs and saw it was crashing via some native function under libc causing a fault

A/libc: Fatal signal 11 (SIGSEGV), code 1, fault addr 0x10 in tid 3555 (edu.gvsu.masl)

Can you explain me why this would occur.

One more anomaly I could see was

But the app compiles well without any error and I can deploy it on the device.

Any help regarding this would be helpful.

Unfortunately it is not working for. Neither in Eclipse nor in Android Studio :(

Are there any plans to include this app in F-Droid? Would be nice to have an easy way to install it...

I cannot make that codegen link to the Java side, I mean the Native library doesn't link at all. Any help in that regard?

Hi,
I'm very interested in your EchoprintForAndroid. I'm tryng it using EchoPrintServer as server backend (official python version and nodejs porting) but I'm not able to match any songs...
Witch server implementation do you use? Can you help me?

Hello Alex,

Thanks for the code and the corresponding tutorial. However, there is an issue with the code, if I just build the project as it is I'm getting "Unknow error". It's the same issue that a user nicknamed @cem reported in one of the blog comments. I've taken a look at the AudioFingerprinter class and the problem starts at line 245.

The if(jobj.has("code")) always returns false, as well as every if statement below this line, so it always executes didFailWithException(new Exception("Unknown error"));, regardless of the presence of data in the EchoPrint response. I'm not an expert in JSON but it seems like Echo Nest has changed the structure of its response and your code fails to parse the data from the JSON object. Can you comment on this?

Everything up to this point works fine, although sometimes EchoPrint fails to find a match even when a minute ago it could, but that's beside the point of this comment.

/**
* Invoke the echoprint native library and generate the fingerprint code.

* Since echoprint requires the audio data to be an array of floats in the

* range [-1, 1] this method will normalize the data array transforming the

* 16 bit signed shorts into floats.
*
* @param data PCM encoded data as shorts
* @param numSamples number of PCM samples at 11025 KHz
* @return The generated fingerprint as a compressed - base64 string.
*/
public String generate(short data[], int numSamples){..}

I read your comment then I know you record audio at 11025 Khz. I've changed to 44100 Khz to see what happens then I know It doesn't recognize music anymore.

The mooma.sh api was shut down recently. So does the lib still work?