readalongs / soundswallower Goto Github PK

An even smaller speech recognizer / force aligner

License: Other

CMake 0.74% C 90.39% Python 1.61% Shell 0.25% Perl 0.22% JavaScript 2.09% Yacc 0.27% ReScript 0.08% Cython 2.80% HTML 0.04% Raku 0.03% C++ 1.34% TypeScript 0.16%

soundswallower's People

Contributors

Stargazers

Watchers

Forkers

jimregan dhdaines twoxfh ming-hai

soundswallower's Issues

Node.js considered harmful, need alternate testing/ci solution

There is no good reason to compile to WebAssembly and run under Node.js aside from testing, and doing so makes everything more complicated and prevents us from easily packaging SoundSwallower as an ES6 module (though maybe that bug is fixed: emscripten-core/emscripten#17915)

Still, there isn't any good reason to do it. SoundSwallower is always used via Python for offline, command-line, etc, and this makes sense since all the code stuff for manipulating speech data and models is there anyway.

The problem is that we need to test the JavaScript code, and Node.js is pretty useful for that. There are I think like 15 different solutions for providing an automated test environment that looks like the browser, all of them have cutsey names and probably icons too, and I have no clue.

So probably I should figure this out.

Replace some parts of the C code with JavaScript or Python

There are most certainly large parts of the code that just shouldn't be written in C.

An obvious candidate is the JSGF parser, which can be reimplented with Jison http://zaa.ch/jison/ or PEG.js https://pegjs.org/ (yes, as far as I know, there isn't a JSGF parser written in JavaScript already, and yes, despite its name, you can't just parse JSGF as JavaScript)

It would be interesting to reimplement the FSG search as well. The HMM/GMM implementations would remain in C, obviously.

Integrate MLLR into the decoding API somehow

This requires linear algebra routines to be enabled everywhere (#10) and two-pass alignment to be implemented (#9).

Once we have that, we should think about how to do this. At the very least we should have an explicit "enrollment" mode that does force-alignment from a known text or finite-state grammar.

--align-text should set_search after set_alignment_fsg

Detail:

(.env) Hobbes:build kal$ soundswallower --model /Volumes/experiments/en-es/en/phonetic-1000/model_parameters/en-lss-02.cd_ptm_1000 --dict  /Volumes/experiments/en-es/en/etc/en-lss-02.dic --align-text "i am good thank you" ~/wav/i_am_good_thank_you.wav 
INFO:root:Setting sample rate to 24000
INFO:root:Increasing FFT size to 1024 for sample rate 24000
ERROR: "pocketsphinx.c", line 594: No search module is selected, did you forget to specify a language model or grammar?
Traceback (most recent call last):
  File "/Users/kal/dev/other/SoundSwallower/.env/bin/soundswallower", line 33, in <module>
    sys.exit(load_entry_point('soundswallower', 'console_scripts', 'soundswallower')())
  File "/Users/kal/dev/other/SoundSwallower/py/soundswallower/cli.py", line 152, in main
    _, file_align = decoder.decode_file(input_file)
  File "_soundswallower.pyx", line 732, in _soundswallower.Decoder.decode_file
    self.start_utt()
  File "_soundswallower.pyx", line 413, in _soundswallower.Decoder.start_utt
    raise RuntimeError, "Failed to start utterance processing"
RuntimeError: Failed to start utterance processing

However when using a JSGF, it does fine:

$ soundswallower --model /Volumes/experiments/en-es/en/phonetic-1000/model_parameters/en-lss-02.cd_ptm_1000 --dict  /Volumes/experiments/en-es/en/etc/en-lss-02.dic ~/wav/i_am_good_thank_you.wav -g i_am_good_thank_you.jsgf 
INFO:root:Setting sample rate to 24000
INFO:root:Increasing FFT size to 1024 for sample rate 24000
[[{"id": "i", "start": 0.0, "end": 0.2}, {"id": "am", "start": 0.2, "end": 0.37}, {"id": "good", "start": 0.37, "end": 0.75}, {"id": "thank", "start": 0.75, "end": 1.18}, {"id": "you", "start": 1.18, "end": 1.46}]]

Add pre-compiled wheels for Python 3.11

I got one e-mail about not being able to install ReadAlongs Studio with Python 3.11 on Windows 11 because the compiler was not installed. lxml and soundswallower were the issues.

It would be nice to include Python 3.12 at the same time if that's feasible already.

add TypeScript annotations

Let's add the files to get proper type hinting, etc, when using SoundSwallower from TypeScript. I guess these are those magical .d files :)

Rust binding

Somebody already did this for PocketSphinx, though it needs to be updated: https://github.com/cmusphinx/pocketsphinx-rs

Evidently it will not be "batteries included" like the Python, as that is not very rusty.

This is a good stepping stone to rewriting the whole library in Rust, which is a good idea (though it will probably make it less smol, at least at first).

Front-End is quite bad

First of all because it has a couple of layers of historical cruft around it ... you pass a config_t which then gets "parsed" into a separate parameters structure, which then gets "parsed" into actual parameters.

But more importantly because everything is directed from this top-level configuration, so once you have configured it, it will only give you one particular kind of features. This in particular makes the case where we'd like to do visualization (using power-spectrum, mel-spectrum, or smoothed spectrum) and also recognition (using MFCC) at the same time non-possible.

Also there are like 5 kinds of VTLN, none of which we use, and probably none of which we will ever use.

And finally because the code which does the buffering and windowing is just utter trash (and I should know because I wrote it).

SoundSwallower installer is not compatible with Python 3.12

Python 3.12.1
Ubuntu 22.04.3

$ pip install soundswallower
Collecting soundswallower
  Using cached soundswallower-0.6.0.tar.gz (10.4 MB)
  Installing build dependencies ... done
  Getting requirements to build wheel ... error
  error: subprocess-exited-with-error

  × Getting requirements to build wheel did not run successfully.
  │ exit code: 1
  ╰─> [33 lines of output]
      Traceback (most recent call last):
        File "/home/joa125/u/miniconda3/envs/ras-py312/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 353, in <module>
          main()
        File "/home/joa125/u/miniconda3/envs/ras-py312/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 335, in main
          json_out['return_val'] = hook(**hook_input['kwargs'])
                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "/home/joa125/u/miniconda3/envs/ras-py312/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 112, in get_requires_for_build_wheel
          backend = _build_backend()
                    ^^^^^^^^^^^^^^^^
        File "/home/joa125/u/miniconda3/envs/ras-py312/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 77, in _build_backend
          obj = import_module(mod_path)
                ^^^^^^^^^^^^^^^^^^^^^^^
        File "/home/joa125/u/miniconda3/envs/ras-py312/lib/python3.12/importlib/__init__.py", line 90, in import_module
          return _bootstrap._gcd_import(name[level:], package, level)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "<frozen importlib._bootstrap>", line 1387, in _gcd_import
        File "<frozen importlib._bootstrap>", line 1360, in _find_and_load
        File "<frozen importlib._bootstrap>", line 1310, in _find_and_load_unlocked
        File "<frozen importlib._bootstrap>", line 488, in _call_with_frames_removed
        File "<frozen importlib._bootstrap>", line 1387, in _gcd_import
        File "<frozen importlib._bootstrap>", line 1360, in _find_and_load
        File "<frozen importlib._bootstrap>", line 1331, in _find_and_load_unlocked
        File "<frozen importlib._bootstrap>", line 935, in _load_unlocked
        File "<frozen importlib._bootstrap_external>", line 994, in exec_module
        File "<frozen importlib._bootstrap>", line 488, in _call_with_frames_removed
        File "/gpfs/fs3c/nrc/ict/portage/u/joa125/tmp/pip-build-env-2oqu8f8y/overlay/lib/python3.12/site-packages/setuptools/__init__.py", line 16, in <module>
          import setuptools.version
        File "/gpfs/fs3c/nrc/ict/portage/u/joa125/tmp/pip-build-env-2oqu8f8y/overlay/lib/python3.12/site-packages/setuptools/version.py", line 1, in <module>
          import pkg_resources
        File "/gpfs/fs3c/nrc/ict/portage/u/joa125/tmp/pip-build-env-2oqu8f8y/overlay/lib/python3.12/site-packages/pkg_resources/__init__.py", line 2191, in <module>
          register_finder(pkgutil.ImpImporter, find_on_path)
                          ^^^^^^^^^^^^^^^^^^^
      AttributeError: module 'pkgutil' has no attribute 'ImpImporter'. Did you mean: 'zipimporter'?
      [end of output]

  note: This error originates from a subprocess, and is likely not a problem with pip.
error: subprocess-exited-with-error

× Getting requirements to build wheel did not run successfully.
│ exit code: 1
╰─> See above for output.

note: This error originates from a subprocess, and is likely not a problem with pip.

Incremental offline decoding/alignment API needed

For force alignment we are not actually doing live recognition, but the API inherited from PocketSphinx still assumes this. We need a better one, which:

Assumes that the entire input is available (though perhaps not all in memory)
Does CMN on a sufficient quantity of input to get good results
Emits progress events (which can be wrapped in an Observable)
Breaks the computation into small enough pieces to be run asynchronously in the main thread

Python build fails on ARM cpu

Hi,

I'm trying to build a Python package on an Arm board (Raspberry Pi 3b) but it keeps failing with the below error.
Any idea how I can debug it and try to solve it? Thanks

Looking in indexes: https://pypi.org/simple, https://www.piwheels.org/simple
Processing /home/zavalyshyn/SoundSwallower
  Installing build dependencies ... done
  Getting requirements to build wheel ... done
    Preparing wheel metadata ... done
Building wheels for collected packages: soundswallower
  Building wheel for soundswallower (PEP 517) ... error
  ERROR: Command errored out with exit status 1:
   command: /usr/bin/python3 /tmp/tmpwtugxr1s_in_process.py build_wheel /tmp/tmpnzbnnqdm
       cwd: /tmp/pip-req-build-1cursi7w
  Complete output (56 lines):
  
  
  --------------------------------------------------------------------------------
  -- Trying "Ninja" generator
  --------------------------------
  ---------------------------
  ----------------------
  -----------------
  ------------
  -------
  --
  Not searching for unused variables given on the command line.
  -- The C compiler identification is GNU 10.2.1
  -- Detecting C compiler ABI info
  -- Detecting C compiler ABI info - done
  -- Check for working C compiler: /usr/bin/cc - skipped
  -- Detecting C compile features
  -- Detecting C compile features - done
  -- The CXX compiler identification is GNU 10.2.1
  -- Detecting CXX compiler ABI info
  -- Detecting CXX compiler ABI info - done
  -- Check for working CXX compiler: /usr/bin/c++ - skipped
  -- Detecting CXX compile features
  -- Detecting CXX compile features - done
  -- Configuring done
  -- Generating done
  -- Build files have been written to: /tmp/pip-req-build-1cursi7w/_cmake_test_compile/build
  --
  -------
  ------------
  -----------------
  ----------------------
  ---------------------------
  --------------------------------
  -- Trying "Ninja" generator - success
  --------------------------------------------------------------------------------
  
  Traceback (most recent call last):
    File "/tmp/tmpwtugxr1s_in_process.py", line 280, in <module>
      main()
    File "/tmp/tmpwtugxr1s_in_process.py", line 263, in main
      json_out['return_val'] = hook(**hook_input['kwargs'])
    File "/tmp/tmpwtugxr1s_in_process.py", line 204, in build_wheel
      return _build_backend().build_wheel(wheel_directory, config_settings,
    File "/tmp/pip-build-env-c4y_864f/overlay/lib/python3.9/site-packages/setuptools/build_meta.py", line 244, in build_wheel
      return self._build_with_temp_dir(['bdist_wheel'], '.whl',
    File "/tmp/pip-build-env-c4y_864f/overlay/lib/python3.9/site-packages/setuptools/build_meta.py", line 229, in _build_with_temp_dir
      self.run_setup()
    File "/tmp/pip-build-env-c4y_864f/overlay/lib/python3.9/site-packages/setuptools/build_meta.py", line 174, in run_setup
      exec(code, locals())
    File "<string>", line 4, in <module>
    File "/tmp/pip-build-env-c4y_864f/overlay/lib/python3.9/site-packages/skbuild/setuptools_wrap.py", line 632, in setup
      env = cmkr.configure(
    File "/tmp/pip-build-env-c4y_864f/overlay/lib/python3.9/site-packages/skbuild/cmaker.py", line 302, in configure
      cmd.append(f"{prefix}_NumPy_INCLUDE_DIRS:PATH=" + np.get_include())
  AttributeError: module 'numpy' has no attribute 'get_include'
  ----------------------------------------
  ERROR: Failed building wheel for soundswallower
Failed to build soundswallower
ERROR: Could not build wheels for soundswallower which use PEP 517 and cannot be installed directly

Consider adding a requirements.txt containing scikit-build

A requirements.txt with scikit-build it in is all that's needed to populate the venv with everything it needs

Support at least some subset of the WebSpeech API

We should be able to just do this since we suport JSGF, etc: https://developer.mozilla.org/en-US/docs/Web/API/Web_Speech_API/Using_the_Web_Speech_API

Expose disabling logging to the Python interface

As stated above!

Build self-contained and portable wheels for PyPI

It's not totally clear how to do this from inside the source tree. Maybe we need to use scikit-build?

Alternately we can just compile libsoundswallower statically (but as PIC) and link it into the Python module. This seems like it might be the best idea.

Then we need to set up the whole thing to compile for manylinux: https://github.com/pypa/manylinux

This maybe has some hints https://thomastrapp.com/blog/building-a-pypi-package-for-a-modern-cpp-project/#packaging-a-precompiled-module-for-pypi

feature request: two pass phonetic alignment

It would be great to be able to easily get both word and phonetic timings/alignments -- perhaps as a two pass process, first to get which pronunciations and word timings, and then a second pass for phone timings. In general, it would be great to get phonetic alignments, but getting word AND phonetic alignments with some sort of sensible relationship would be superb.

We've done some of this by doing a second pass the pronunciation, using "phoneme words" with single-phone pronunciations, but that doesn't account for word-phone-position features in the clustering (b, i, m ,s) when using cd models.

Another issue that came up with that approach is that the word boundaries and the phoneme boundary timings didn't align, due to differences in the search. It would be possible to do a phonetic decode within the word boundary, or to adjust to word boundaries after the phonetic pass...

Package SoundSwallower as ES6 module

It seems that this is possible, so we should do it.

feature request: -cionly

For efficiency, and also for improved alignment in some situations, it is useful to use only the context-independent phones from an existing acoustic model. At a minimum this means simply disabling the CD phones in the mdef object. However, we should also be able to disable the cross-word triphone handling in the search.

WebRTC VAD code considered harmful (on the browser)

Considering that this code already exists somewhere in the guts of the browser, it is pretty silly to compile it separately into WebAssembly. Unfortunately, there isn't actually any API to access it from JavaScript, so we are stuck having to do our own VAD for endpointing.

The problem with the WebRTC code used in PocketSphinx5 is:

Computation is done in fixed-point, so we have to convert back and forth between Float32
WebAudio doesn't let us choose our buffer size, and neither does the VAD, so we have to implement a ring-buffer (we have to do this anyway, but...)
WebAudio can already do an FFT for us, more efficiently, but the AnalyzerNode API is utter garbage designed only for making pretty pictures, so never mind

For these reasons the ideal solution is, horror of horrors, something very much like the -remove_silence option in PocketSphinx that was the whole reason for creating SoundSwallower in the first place (because I was so seriously annoyed at it removing data from the input, making force-alignment useless). Of course, it has to be done in a way that makes endpointing optional and doesn't break the batch-mode API. So, specifically:

Encapsulate input features (MFCCs, but not necessarily) for the decoder
Create a fused feature extractor and endpointer which emits speech start/stop events and feature buffers, with timestamps

Internally we can either use the WebRTC method based on log-spectra or the PocketSphinx 5prealpha method.

Remove binary mdef code

Text-format model definitions are bigger on disk but they actually compress down much smaller (600k vs 900k for the default en-us model using default gzip settings) which is what matters when serving on the web.

We may wish to use JSON in the future but for now at least we can make the code smaller...

Enable linear algebra routines everywhere

Currently we have to punt on the linear algebra routines by default because not everybody has BLAS and LAPACK, and the "lite" versions are somewhat out of date and don't compile properly to JavaScript, among other things.

Solutions do exist! First we should update the "lite" ones and strip them down to the bare essentials. For JavaScript we can use emlapack. And we can optionally use a system LAPACK, but this should not be on by default as it is an external dependency.

Change the default branch name to main

We should follow this change we've done everywhere else, and rename the default branch main.

Logo

If we named this "SoundSwallow" we could make a cute bird logo for it :D

Get rid of public/private API distinctions

The PocketSphinx API uses incomplete types everywhere to do very strict data hiding. This is conceptually very nice but in reality we don't care, because we want this library to be small enough that everybody can just have their own copy of it and not bother with ABI compatibility. Also we're going to compile it to JavaScript anyway.

Basically this means that pocketsphinx_internal.h and friends should move to the public include directory.

Model converter needed

The acoustic model format has diverged somewhat from CMU Sphinx, and is expected to diverge further. Supporting multiple model formats is not consistent with the goal of making the smallest possible library, so we require a converter to be able to use publically available models. Currently this means:

Convert text to binary model definition
Convert mixture_weights to sendump
Rename text files to include ".txt" extension
Convert feat_params to JSON
Include default dictionary

In the future it may mean (but this is not in the scope of this issue):

Dictionary is an FST and may be a G2P model
Model definition is also an FST (i.e. the "HC" in "HCLG")
GMMs are also quantized
GMMs might be replaced by DNNs

Replace configuration API with JSON

Instead of command-line switches (there are too many of them) we should use JSON, or Python dictionaries, as the case may be, to configure the recognizer. Also, we need to stop parsing strings as floats internally to the code, this should only be done when we first read the configuration.

Disable VAD by default

The VAD implementation is poorly thought-out because it messes up the time alignments in the output. Requiring people to use the continuous API doesn't really fix this problem, and isn't what we want for a lot of use cases (force alignment, for instance).

It should be disabled pending some kind of fix to get the alignments right in the output.

Missing matrix.h and ps_search.h

Summary:

on 864bc85
macOS 12.3.1 (Monterey)

CMake Error at include/soundswallower/cmake_install.cmake:41 (file):
  file INSTALL cannot find
  "/Users/kal/dev/other/SoundSwallower/include/soundswallower/matrix.h": No
  such file or directory.
Call Stack (most recent call first):
  cmake_install.cmake:51 (include)

Detail:

Hobbes:other kal$ git clone https://github.com/ReadAlongs/SoundSwallower.git
Cloning into 'SoundSwallower'...
remote: Enumerating objects: 2564, done.
remote: Counting objects: 100% (19/19), done.
remote: Compressing objects: 100% (13/13), done.
remote: Total 2564 (delta 8), reused 17 (delta 6), pack-reused 2545
Receiving objects: 100% (2564/2564), 42.25 MiB | 8.51 MiB/s, done.
Resolving deltas: 100% (1664/1664), done.
Hobbes:SoundSwallower kal$ open README.md
Hobbes:SoundSwallower kal$ mkdir build
Hobbes:SoundSwallower kal$ cd build
Hobbes:build kal$ cmake ..
-- The C compiler identification is AppleClang 13.1.6.13160021
-- The CXX compiler identification is AppleClang 13.1.6.13160021
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for unistd.h
-- Looking for unistd.h - found
-- Looking for stdint.h
-- Looking for stdint.h - found
-- Looking for sys/types.h
-- Looking for sys/types.h - found
-- Looking for sys/stat.h
-- Looking for sys/stat.h - found
-- Looking for snprintf
-- Looking for snprintf - found
-- Check if the system is big endian
-- Searching 16 bit integer
-- Looking for stddef.h
-- Looking for stddef.h - found
-- Check size of unsigned short
-- Check size of unsigned short - done
-- Searching 16 bit integer - Using unsigned short
-- Check if the system is big endian - little endian
-- Configuring done
-- Generating done
-- Build files have been written to: /Users/kal/dev/other/SoundSwallower/build
Hobbes:build kal$ make
Scanning dependencies of target soundswallower
[  0%] Building C object src/CMakeFiles/soundswallower.dir/acmod.c.o
[  1%] Building C object src/CMakeFiles/soundswallower.dir/bin_mdef.c.o
[  2%] Building C object src/CMakeFiles/soundswallower.dir/bio.c.o
[  3%] Building C object src/CMakeFiles/soundswallower.dir/bitvec.c.o
[  4%] Building C object src/CMakeFiles/soundswallower.dir/blkarray_list.c.o
[  5%] Building C object src/CMakeFiles/soundswallower.dir/case.c.o
[  6%] Building C object src/CMakeFiles/soundswallower.dir/ckd_alloc.c.o
[  6%] Building C object src/CMakeFiles/soundswallower.dir/cmd_ln.c.o
[  7%] Building C object src/CMakeFiles/soundswallower.dir/cmn.c.o
[  8%] Building C object src/CMakeFiles/soundswallower.dir/cmn_live.c.o
[  9%] Building C object src/CMakeFiles/soundswallower.dir/dict2pid.c.o
[ 10%] Building C object src/CMakeFiles/soundswallower.dir/dict.c.o
[ 11%] Building C object src/CMakeFiles/soundswallower.dir/err.c.o
/Users/kal/dev/other/SoundSwallower/src/err.c:169:19: warning: variable 'oldfp' set but not used [-Wunused-but-set-variable]
    FILE *newfp, *oldfp;
                  ^
1 warning generated.
[ 12%] Building C object src/CMakeFiles/soundswallower.dir/feat.c.o
[ 12%] Building C object src/CMakeFiles/soundswallower.dir/fe_interface.c.o
[ 13%] Building C object src/CMakeFiles/soundswallower.dir/fe_sigproc.c.o
[ 14%] Building C object src/CMakeFiles/soundswallower.dir/fe_warp_affine.c.o
[ 15%] Building C object src/CMakeFiles/soundswallower.dir/fe_warp.c.o
[ 16%] Building C object src/CMakeFiles/soundswallower.dir/fe_warp_inverse_linear.c.o
[ 17%] Building C object src/CMakeFiles/soundswallower.dir/fe_warp_piecewise_linear.c.o
[ 18%] Building C object src/CMakeFiles/soundswallower.dir/filename.c.o
[ 18%] Building C object src/CMakeFiles/soundswallower.dir/fsg_history.c.o
[ 19%] Building C object src/CMakeFiles/soundswallower.dir/fsg_lextree.c.o
[ 20%] Building C object src/CMakeFiles/soundswallower.dir/fsg_model.c.o
[ 21%] Building C object src/CMakeFiles/soundswallower.dir/fsg_search.c.o
[ 22%] Building C object src/CMakeFiles/soundswallower.dir/genrand.c.o
[ 23%] Building C object src/CMakeFiles/soundswallower.dir/glist.c.o
[ 24%] Building C object src/CMakeFiles/soundswallower.dir/hash_table.c.o
[ 25%] Building C object src/CMakeFiles/soundswallower.dir/heap.c.o
[ 25%] Building C object src/CMakeFiles/soundswallower.dir/hmm.c.o
[ 26%] Building C object src/CMakeFiles/soundswallower.dir/jsgf.c.o
[ 27%] Building C object src/CMakeFiles/soundswallower.dir/jsgf_parser.c.o
[ 28%] Building C object src/CMakeFiles/soundswallower.dir/jsgf_scanner.c.o
[ 29%] Building C object src/CMakeFiles/soundswallower.dir/listelem_alloc.c.o
[ 30%] Building C object src/CMakeFiles/soundswallower.dir/logmath.c.o
[ 31%] Building C object src/CMakeFiles/soundswallower.dir/mdef.c.o
[ 31%] Building C object src/CMakeFiles/soundswallower.dir/mmio.c.o
[ 32%] Building C object src/CMakeFiles/soundswallower.dir/ms_gauden.c.o
[ 33%] Building C object src/CMakeFiles/soundswallower.dir/ms_mgau.c.o
[ 34%] Building C object src/CMakeFiles/soundswallower.dir/ms_senone.c.o
[ 35%] Building C object src/CMakeFiles/soundswallower.dir/pio.c.o
/Users/kal/dev/other/SoundSwallower/src/pio.c:77:1: warning: unused function 'guess_comptype' [-Wunused-function]
guess_comptype(char const *file, int32 *ispipe, int32 *isgz)
^
1 warning generated.
[ 36%] Building C object src/CMakeFiles/soundswallower.dir/pocketsphinx.c.o
[ 37%] Building C object src/CMakeFiles/soundswallower.dir/profile.c.o
[ 37%] Building C object src/CMakeFiles/soundswallower.dir/ps_lattice.c.o
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1181:15: warning: unused variable 'n_used' [-Wunused-variable]
        int32 n_used;
              ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1182:15: warning: variable 'to_is_fil' set but not used [-Wunused-but-set-variable]
        int16 to_is_fil;
              ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1247:19: warning: variable 'w1_is_fil' set but not used [-Wunused-but-set-variable]
            int16 w1_is_fil;
                  ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1196:22: warning: unused variable 'n_used' [-Wunused-variable]
        int32 bprob, n_used;
                     ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1274:22: warning: unused variable 'n_used' [-Wunused-variable]
        int32 bprob, n_used;
                     ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1322:17: warning: unused variable 'lback' [-Wunused-variable]
            int lback;
                ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1380:22: warning: unused variable 'n_used' [-Wunused-variable]
        int32 bprob, n_used;
                     ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1497:15: warning: unused variable 'n_used' [-Wunused-variable]
        int32 n_used;
              ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1568:15: warning: unused variable 'n_used' [-Wunused-variable]
        int32 n_used;
              ^
/Users/kal/dev/other/SoundSwallower/src/ps_lattice.c:1638:19: warning: unused variable 'n_used' [-Wunused-variable]
            int32 n_used;
                  ^
10 warnings generated.
[ 38%] Building C object src/CMakeFiles/soundswallower.dir/ps_mllr.c.o
[ 39%] Building C object src/CMakeFiles/soundswallower.dir/ptm_mgau.c.o
[ 40%] Building C object src/CMakeFiles/soundswallower.dir/s2_semi_mgau.c.o
[ 41%] Building C object src/CMakeFiles/soundswallower.dir/strfuncs.c.o
[ 42%] Building C object src/CMakeFiles/soundswallower.dir/tmat.c.o
[ 43%] Building C object src/CMakeFiles/soundswallower.dir/vector.c.o
[ 43%] Building C object src/CMakeFiles/soundswallower.dir/yin.c.o
[ 44%] Linking C static library libsoundswallower.a
[ 44%] Built target soundswallower
Scanning dependencies of target test_subvq
[ 45%] Building C object tests/CMakeFiles/test_subvq.dir/test_subvq.c.o
[ 46%] Linking C executable test_subvq
[ 46%] Built target test_subvq
Scanning dependencies of target test_strfuncs
[ 47%] Building C object tests/CMakeFiles/test_strfuncs.dir/test_strfuncs.c.o
[ 48%] Linking C executable test_strfuncs
[ 48%] Built target test_strfuncs
Scanning dependencies of target test_readfile
[ 49%] Building C object tests/CMakeFiles/test_readfile.dir/test_readfile.c.o
[ 50%] Linking C executable test_readfile
[ 50%] Built target test_readfile
Scanning dependencies of target test_ptm_mgau
[ 50%] Building C object tests/CMakeFiles/test_ptm_mgau.dir/test_ptm_mgau.c.o
[ 51%] Linking C executable test_ptm_mgau
[ 51%] Built target test_ptm_mgau
Scanning dependencies of target test_pitch
[ 52%] Building C object tests/CMakeFiles/test_pitch.dir/test_pitch.c.o
[ 53%] Linking C executable test_pitch
[ 53%] Built target test_pitch
Scanning dependencies of target test_log_int16
[ 54%] Building C object tests/CMakeFiles/test_log_int16.dir/test_log_int16.c.o
[ 54%] Linking C executable test_log_int16
[ 54%] Built target test_log_int16
Scanning dependencies of target test_listelem_alloc
[ 55%] Building C object tests/CMakeFiles/test_listelem_alloc.dir/test_listelem_alloc.c.o
[ 56%] Linking C executable test_listelem_alloc
[ 56%] Built target test_listelem_alloc
Scanning dependencies of target test_hash_iter
[ 57%] Building C object tests/CMakeFiles/test_hash_iter.dir/test_hash_iter.c.o
[ 58%] Linking C executable test_hash_iter
[ 58%] Built target test_hash_iter
Scanning dependencies of target test_log_shifted
[ 59%] Building C object tests/CMakeFiles/test_log_shifted.dir/test_log_shifted.c.o
[ 60%] Linking C executable test_log_shifted
[ 60%] Built target test_log_shifted
Scanning dependencies of target test_filename
[ 61%] Building C object tests/CMakeFiles/test_filename.dir/test_filename.c.o
[ 61%] Linking C executable test_filename
[ 61%] Built target test_filename
Scanning dependencies of target test_bitvec
[ 61%] Building C object tests/CMakeFiles/test_bitvec.dir/test_bitvec.c.o
[ 62%] Linking C executable test_bitvec
[ 62%] Built target test_bitvec
Scanning dependencies of target test_fsg
[ 63%] Building C object tests/CMakeFiles/test_fsg.dir/test_fsg.c.o
[ 64%] Linking C executable test_fsg
[ 64%] Built target test_fsg
Scanning dependencies of target test_fe
[ 65%] Building C object tests/CMakeFiles/test_fe.dir/test_fe.c.o
[ 65%] Linking C executable test_fe
[ 65%] Built target test_fe
Scanning dependencies of target test_lineiter
[ 66%] Building C object tests/CMakeFiles/test_lineiter.dir/test_lineiter.c.o
[ 67%] Linking C executable test_lineiter
[ 67%] Built target test_lineiter
Scanning dependencies of target test_case
[ 68%] Building C object tests/CMakeFiles/test_case.dir/test_case.c.o
[ 69%] Linking C executable test_case
[ 69%] Built target test_case
Scanning dependencies of target test_display_hash
[ 70%] Building C object tests/CMakeFiles/test_display_hash.dir/test_display_hash.c.o
[ 71%] Linking C executable test_display_hash
[ 71%] Built target test_display_hash
Scanning dependencies of target test_dict
[ 71%] Building C object tests/CMakeFiles/test_dict.dir/test_dict.c.o
[ 72%] Linking C executable test_dict
[ 72%] Built target test_dict
Scanning dependencies of target test_build_directory
[ 73%] Building C object tests/CMakeFiles/test_build_directory.dir/test_build_directory.c.o
[ 74%] Linking C executable test_build_directory
[ 74%] Built target test_build_directory
Scanning dependencies of target test_cmdln_parse_r
[ 75%] Building C object tests/CMakeFiles/test_cmdln_parse_r.dir/test_cmdln_parse_r.c.o
[ 76%] Linking C executable test_cmdln_parse_r
[ 76%] Built target test_cmdln_parse_r
Scanning dependencies of target test_feat_fe
[ 77%] Building C object tests/CMakeFiles/test_feat_fe.dir/test_feat_fe.c.o
[ 78%] Linking C executable test_feat_fe
[ 78%] Built target test_feat_fe
Scanning dependencies of target test_ckd_alloc_catch
[ 79%] Building C object tests/CMakeFiles/test_ckd_alloc_catch.dir/test_ckd_alloc_catch.c.o
/Users/kal/dev/other/SoundSwallower/tests/test_ckd_alloc_catch.c:12:7: warning: variable 'alloc1' set but not used [-Wunused-but-set-variable]
        int *alloc1;
             ^
1 warning generated.
[ 79%] Linking C executable test_ckd_alloc_catch
[ 79%] Built target test_ckd_alloc_catch
Scanning dependencies of target test_heap
[ 80%] Building C object tests/CMakeFiles/test_heap.dir/test_heap.c.o
[ 81%] Linking C executable test_heap
[ 81%] Built target test_heap
Scanning dependencies of target test_acmod_grow
[ 82%] Building C object tests/CMakeFiles/test_acmod_grow.dir/test_acmod_grow.c.o
[ 83%] Linking C executable test_acmod_grow
[ 83%] Built target test_acmod_grow
Scanning dependencies of target test_delete_hash
[ 84%] Building C object tests/CMakeFiles/test_delete_hash.dir/test_delete_hash.c.o
[ 85%] Linking C executable test_delete_hash
[ 85%] Built target test_delete_hash
Scanning dependencies of target test_bit_encode
[ 86%] Building C object tests/CMakeFiles/test_bit_encode.dir/test_bit_encode.c.o
[ 87%] Linking C executable test_bit_encode
[ 87%] Built target test_bit_encode
Scanning dependencies of target test_feat_live
[ 88%] Building C object tests/CMakeFiles/test_feat_live.dir/test_feat_live.c.o
[ 89%] Linking C executable test_feat_live
[ 89%] Built target test_feat_live
Scanning dependencies of target test_log_int8
[ 90%] Building C object tests/CMakeFiles/test_log_int8.dir/test_log_int8.c.o
[ 91%] Linking C executable test_log_int8
[ 91%] Built target test_log_int8
Scanning dependencies of target test_feat
[ 92%] Building C object tests/CMakeFiles/test_feat.dir/test_feat.c.o
[ 93%] Linking C executable test_feat
[ 93%] Built target test_feat
Scanning dependencies of target test_dict2pid
[ 94%] Building C object tests/CMakeFiles/test_dict2pid.dir/test_dict2pid.c.o
[ 95%] Linking C executable test_dict2pid
[ 95%] Built target test_dict2pid
Scanning dependencies of target test_ckd_alloc_fail
[ 96%] Building C object tests/CMakeFiles/test_ckd_alloc_fail.dir/test_ckd_alloc_fail.c.o
/Users/kal/dev/other/SoundSwallower/tests/test_ckd_alloc_fail.c:12:7: warning: variable 'alloc1' set but not used [-Wunused-but-set-variable]
        int *alloc1;
             ^
1 warning generated.
[ 97%] Linking C executable test_ckd_alloc_fail
[ 97%] Built target test_ckd_alloc_fail
Scanning dependencies of target test_jsgf
[ 97%] Building C object tests/CMakeFiles/test_jsgf.dir/test_jsgf.c.o
[ 98%] Linking C executable test_jsgf
[ 98%] Built target test_jsgf
Scanning dependencies of target test_acmod
[ 99%] Building C object tests/CMakeFiles/test_acmod.dir/test_acmod.c.o
[100%] Linking C executable test_acmod
[100%] Built target test_acmod
Hobbes:build kal$ make test
Running tests...
Test project /Users/kal/dev/other/SoundSwallower/build
      Start  1: lcase1
 1/39 Test  #1: lcase1 ...........................   Passed    0.08 sec
      Start  2: lcase2
 2/39 Test  #2: lcase2 ...........................   Passed    0.00 sec
      Start  3: lcase3
 3/39 Test  #3: lcase3 ...........................   Passed    0.00 sec
      Start  4: ucase1
 4/39 Test  #4: ucase1 ...........................   Passed    0.00 sec
      Start  5: ucase2
 5/39 Test  #5: ucase2 ...........................   Passed    0.00 sec
      Start  6: strcmp1
 6/39 Test  #6: strcmp1 ..........................   Passed    0.00 sec
      Start  7: strcmp2
 7/39 Test  #7: strcmp2 ..........................   Passed    0.00 sec
      Start  8: strcmp3
 8/39 Test  #8: strcmp3 ..........................   Passed    0.00 sec
      Start  9: nextword
 9/39 Test  #9: nextword .........................   Passed    0.05 sec
      Start 10: str2words
10/39 Test #10: str2words ........................   Passed    0.00 sec
      Start 11: string_join
11/39 Test #11: string_join ......................   Passed    0.00 sec
      Start 12: string_trim
12/39 Test #12: string_trim ......................   Passed    0.00 sec
      Start 13: fread_line
13/39 Test #13: fread_line .......................   Passed    0.00 sec
      Start 14: bit_encode
14/39 Test #14: bit_encode .......................   Passed    0.10 sec
      Start 15: bitvec
15/39 Test #15: bitvec ...........................   Passed    0.34 sec
      Start 16: ckd_alloc_catch
16/39 Test #16: ckd_alloc_catch ..................   Passed    0.10 sec
      Start 17: ckd_alloc_fail
17/39 Test #17: ckd_alloc_fail ...................   Passed    0.05 sec
      Start 18: build_directory
18/39 Test #18: build_directory ..................   Passed    0.05 sec
      Start 19: filename
19/39 Test #19: filename .........................   Passed    0.25 sec
      Start 20: hash_iter
20/39 Test #20: hash_iter ........................   Passed    0.09 sec
      Start 21: heap
21/39 Test #21: heap .............................   Passed    0.05 sec
      Start 22: lineiter
22/39 Test #22: lineiter .........................   Passed    0.05 sec
      Start 23: listelem_alloc
23/39 Test #23: listelem_alloc ...................   Passed    0.05 sec
      Start 24: log_int16
24/39 Test #24: log_int16 ........................   Passed    0.07 sec
      Start 25: log_int8
25/39 Test #25: log_int8 .........................   Passed    0.05 sec
      Start 26: log_shifted
26/39 Test #26: log_shifted ......................   Passed    0.05 sec
      Start 27: readfile
27/39 Test #27: readfile .........................   Passed    0.05 sec
      Start 28: acmod
28/39 Test #28: acmod ............................   Passed    1.89 sec
      Start 29: acmod_grow
29/39 Test #29: acmod_grow .......................   Passed    0.90 sec
      Start 30: dict2pid
30/39 Test #30: dict2pid .........................   Passed    0.22 sec
      Start 31: dict
31/39 Test #31: dict .............................   Passed    0.20 sec
      Start 32: fe
32/39 Test #32: fe ...............................   Passed    0.05 sec
      Start 33: feat_fe
33/39 Test #33: feat_fe ..........................   Passed    0.06 sec
      Start 34: feat_live
34/39 Test #34: feat_live ........................   Passed    0.05 sec
      Start 35: ptm_mgau
35/39 Test #35: ptm_mgau .........................   Passed    0.57 sec
      Start 36: subvq
36/39 Test #36: subvq ............................   Passed    0.05 sec
      Start 37: jsgf
37/39 Test #37: jsgf .............................   Passed    1.01 sec
      Start 38: fsg
38/39 Test #38: fsg ..............................   Passed    0.33 sec
      Start 39: pitch
39/39 Test #39: pitch ............................   Passed    0.07 sec

100% tests passed, 0 tests failed out of 39

Total Test time (real) =   6.92 sec
Hobbes:build kal$ make install
[ 44%] Built target soundswallower
[ 46%] Built target test_subvq
[ 48%] Built target test_strfuncs
[ 50%] Built target test_readfile
[ 51%] Built target test_ptm_mgau
[ 53%] Built target test_pitch
[ 54%] Built target test_log_int16
[ 56%] Built target test_listelem_alloc
[ 58%] Built target test_hash_iter
[ 60%] Built target test_log_shifted
[ 61%] Built target test_filename
[ 62%] Built target test_bitvec
[ 64%] Built target test_fsg
[ 65%] Built target test_fe
[ 67%] Built target test_lineiter
[ 69%] Built target test_case
[ 71%] Built target test_display_hash
[ 72%] Built target test_dict
[ 74%] Built target test_build_directory
[ 76%] Built target test_cmdln_parse_r
[ 78%] Built target test_feat_fe
[ 79%] Built target test_ckd_alloc_catch
[ 81%] Built target test_heap
[ 83%] Built target test_acmod_grow
[ 85%] Built target test_delete_hash
[ 87%] Built target test_bit_encode
[ 89%] Built target test_feat_live
[ 91%] Built target test_log_int8
[ 93%] Built target test_feat
[ 95%] Built target test_dict2pid
[ 97%] Built target test_ckd_alloc_fail
[ 98%] Built target test_jsgf
[100%] Built target test_acmod
Install the project...
-- Install configuration: ""
-- Installing: /usr/local/lib/libsoundswallower.a
-- Installing: /usr/local/include/acmod.h
-- Installing: /usr/local/include/bin_mdef.h
-- Installing: /usr/local/include/bio.h
-- Installing: /usr/local/include/bitvec.h
-- Installing: /usr/local/include/blkarray_list.h
-- Installing: /usr/local/include/byteorder.h
-- Installing: /usr/local/include/case.h
-- Installing: /usr/local/include/ckd_alloc.h
-- Installing: /usr/local/include/cmd_ln.h
-- Installing: /usr/local/include/cmdln_macro.h
-- Installing: /usr/local/include/cmn.h
-- Installing: /usr/local/include/dict2pid.h
-- Installing: /usr/local/include/dict.h
-- Installing: /usr/local/include/err.h
-- Installing: /usr/local/include/feat.h
-- Installing: /usr/local/include/fe.h
-- Installing: /usr/local/include/fe_internal.h
-- Installing: /usr/local/include/fe_type.h
-- Installing: /usr/local/include/fe_warp_affine.h
-- Installing: /usr/local/include/fe_warp.h
-- Installing: /usr/local/include/fe_warp_inverse_linear.h
-- Installing: /usr/local/include/fe_warp_piecewise_linear.h
-- Installing: /usr/local/include/filename.h
-- Installing: /usr/local/include/fsg_history.h
-- Installing: /usr/local/include/fsg_lextree.h
-- Installing: /usr/local/include/fsg_model.h
-- Installing: /usr/local/include/fsg_search_internal.h
-- Installing: /usr/local/include/genrand.h
-- Installing: /usr/local/include/glist.h
-- Installing: /usr/local/include/hash_table.h
-- Installing: /usr/local/include/heap.h
-- Installing: /usr/local/include/hmm.h
-- Installing: /usr/local/include/jsgf.h
-- Installing: /usr/local/include/jsgf_internal.h
-- Installing: /usr/local/include/listelem_alloc.h
-- Installing: /usr/local/include/logmath.h
CMake Error at include/soundswallower/cmake_install.cmake:41 (file):
  file INSTALL cannot find
  "/Users/kal/dev/other/SoundSwallower/include/soundswallower/matrix.h": No
  such file or directory.
Call Stack (most recent call first):
  cmake_install.cmake:51 (include)


make: *** [install] Error 1
Hobbes:build kal$

Set up compilation with Emscripten using existing CMakeLists.txt

Add whatever needs to be done to get this to compile directly with Emscripten

Implement two-pass state alignment and change the alignment search to use it

The alignment search currently doesn't really do the right thing, because it does a state-level alignment only (slow!) and because it doesn't accept silence/noise between words.

This API is fine but it needs to Do What I Mean, which is to do FSG-based "alignment" first then use the resulting coarse alignments to do phone and state level alignments.

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.