Node 03: convert wav file with sox

August 23rd, 2009 by producer

liberty@liberty-desktop:~/.kde/share/apps/simon/model/samtestwav.data$ sox de27-02.wav -r 16000 -c 1 -s de27-02-test.wav

found similar entry here

I hope to be able to convert with sox from 48100 and 22050 to 16.000
why was there a different file size when converting 16000 hertz file to 16000 hertz file?

Node 01: simonview.ui

August 1st, 2009 by producer

01 simonview.ui – looking at the file with Qt Creator, and Geany.
03 /usr/include/simon/simonrecognitionresult/recognitionresult.h
file:///usr/include/simon/simonrecognitionresult/simonrecognitionresult_export.h
– don’t read again
04 file:///usr/include/simon/simonactions/commandlistwidget.h
file:///usr/include/simon/simonactions/greedyreceiver.h
05 file:///usr/include/simon/simonactions/listcommand.h
file:///usr/include/simon/speechmodelmanagement/wordlisttype.h

06 file:///home/liberty/200908/speech2text/trunk/simonlib/speechgui/addserverconnection.cpp
file:///home/liberty/200908/speech2text/trunk/simonlib/speechgui/addserverconnectiondlg.ui – open with Qt Creator
file:///home/liberty/200908/speech2text/trunk/simonlib/speechgui/addserverconnection.h view with geany

07 file:///home/liberty/200908/speech2text/trunk/simonlib/speechgui/inlinewidget.cpp
file:///home/liberty/200908/speech2text/trunk/simonlib/speechgui/inlinewidget.h – could continue to read about classes in the c++ tutorial – c++ node 02

08 build_ubuntu.sh – reading the log http://pastebin.com/m4d269fb5 – continue with line 371

09 file:///home/liberty/200908/speech2text/trunk/simonlib/simonmodeltest/CMakeLists.txt – read two times

10 /usr/share/kde4/apps/simon/model/wav_config
file:///usr/share/kde4/apps/simon/model/tree1.hed – continue with line 132

Dates in configure simon

July 13th, 2009 by producer

I just looked into the folder /home/liberty/.kde/share/apps/simond/models/a/src:

synchronisation

You can compare the folders above with the dates in configure simon:

configure

Obviously, the previous speech models are stored in /home/liberty/.kde/share/apps/simond/models/a/src.

HHEd and the ModelCompilationManager

May 23rd, 2009 by producer

Here is a code snippet from the modelcompilationmanager.cpp:

bool ModelCompilationManager::buildHMM10()
{
return execute(‘”‘+hHEd+”\” -A -D -T 1 -H \”"+htkIfyPath(tempDir)+”/hmm9/macros\” -H \”"+htkIfyPath(tempDir)+”/hmm9/hmmdefs\” -M \”"+htkIfyPath(tempDir)+”/hmm10/\” \”"+htkIfyPath(tempDir)+”/mktri.hed\” \”"+htkIfyPath(tempDir)+”/monophones1\”");
}

I have marked some things in the code snippet. Take a look into HHEd/Use if you want to know what
- the -A -D – T 1 means (probably standard options)
- the -H switch means (load macro file)
- the -H switch means (again)
- the -M switch means (store output in this directory)
- mktri.hed (this is probably the edCmdFile containing some commands)
- monophones1 (probably defines a set of HMMs)

modelcompilationmanager.cpp and HERest

May 23rd, 2009 by producer

Let’s take a look into the file modelcompilationmanager.cpp:

herest

You can see that the tool HERest is beeing invoked. The passing of the arguments -A -D -V -T 1 obviously guarantees that sufficient information is available (registration needed).

Obviously, there are several tools that are accessed by simon: HDMan, HLEd, HCopy, HCompV, HERest, HHEd, and HVite. These tools are not part of simon, you have to install them separately.

Node 04: /simon/model/phonemes

May 14th, 2009 by producer

Left: SAMPA used by simon. Right: IPA used by PLS dictionary

a: aː
b b
@ ə
n n
ts ts
R ʀ
a a
x x
t t
l l
s s
m m
p p
f f
N ŋ
S ʃ
e: eː
v v
O ɔ
r r
aU
E
C
g
I ɪ
k
d
aI
u:
o: oː
U
h
i:
z
Y
OY
i
y:
o
e
oeh:
pf
j
E:
oeh œ
oe
oe:
sil
U:
u
@ r ɐ̯ ɐ
y
O:
Z
N=
n=
m=
A
OI
&