cmusphinx | CMU Sphinx - Speech Recognition Toolkit | Speech library

I had an old version of pocketsphinx. Make sure to have the latest installed.

I solved this problem by specifying the full path of sphinx_train.cfg.

This is the line I edit in slave_feat.pl at line 50.

It seems to be expected behavior for arm64 architecture, because of the way arm64 functions are invoked.

Invoking functions in a arm64 library that use ...) at the end, you have to take into account that the first 8 argument spots are for 'normal' arguments, then optionally the variable/params can start.

So, in my example I used 5 NULL values to fill the first 8 argument spots, then start the values for the ...)

See full answer: https://github.com/xamarin/xamarin-macios/issues/10285

I found some commands that worked for installation:

It is great you tried Vosk offline speech recognition on Android, here are some answers to your questions:

actually I read it still doesn't support my native language.

If you are about Hebrew, we might support it in the future, and you can build it yourself.

is that the reason?

You didn't provide enough information to answer this, please explain a bit more what is "it is not work"

I want to know how it works?

Extensive documentation on speech recognition is available on lectures, courses and books. You can find some introduction here for example: https://www.youtube.com/watch?v=q67z7PTGRi8

does it is doing speech to text and saved it in assets files?

It does speech to text, but it doesn't save results into assets, it just displays them. You can not modify assets, they are static.

does it is run in background?

Yes

does it is used AI models?

Sure

how it behaves when two apps need mic resource in parallel?

In android it is not possible to record audio from two apps in parallel, second one will be blocked.

noises?

It is robust to noises.

does it is work with Neural Networks API?

No, it is portable

how can I start developing such a thing?

Get some basic understanding and start writing the code. If you have further questions you can ask them in the Telegram chat

Check the source code for the FindSwig.cmake.

Unfortunately, if a find script does not work as expected and you do not see right away why that is the case, you usually have to dig into its source. In your case, it looks like CMake was able to find and run the SWIG executable, but then failed to obtain the swig directory.

Try manually running swig -swiglib and check that the printed directory indeed contains a swig.swg file. Also, be sure that the swig executable found by CMake is actually the correct one (you can verify this by inspecting the value of SWIG_EXECUTABLE in either the cmake-gui, the ccmake curses interface, or in the CMakeCache.txt file directly).

Note that CMake will not update the executable path once it has been found! So if you make changes to your system that influence the executable location, you will have to clear the cache (eg. by deleting the CMakeCache.txt) and re-run CMake for the changes to take effect.

Solved by associating perl files ( with pl extension) to perl.exe

Is it possible to feed the feature matrix instead of wav file into CMUSphinx for training?

It is better to train with more modern toolkit like Kaldi.

I have tried searching on google and the tutorial page and can't find any information about it. Am i missing something?

You just prepare mfc files yourself and skip first feature_extraction step. You have to write code for that like described here:

https://cmusphinx.github.io/wiki/mfcformat/

In kaldi you just prepare ark files yourself:

https://pypi.org/project/kaldiio/

If anyone else runs into this problem, check the Visual Studio Runtime Library property is set to Multi-threaded DLL (/MD), especially if you are running a newer version of Visual Studio. To check this, in Visual Studio go to:

Project -> Properties -> C/C++ -> Code Generation -> Runtime Library

to verify/change this setting. You may have to redownload sphinxbase and pocketsphinx before doing this. Build the .sln only after changing this setting.