Menu

best dataset for medium acoustic model

Help
forsubhi
2012-04-28
2012-09-22
  • forsubhi

    forsubhi - 2012-04-28

    I have the following sentences used for training the acoustic model

    <s> ذَهَبَ أُسامَة إلى الكُلِّيَة </s>
    <s> ذَهَبَ إياد إلى الكُلِّيَة صَباحَاً </s>
    <s> ذَهَبَ أُسامَة إلى المَسجِد مَسَاءً </s>
    <s> ذَهَبَ أُسامَة إلى الكُلِّيَة فِي السَّاعَة العَاشِرَة مَسَاءً </s>
    <s> ذَهَبَ أُسامَة إلى السُّوق فِي السَّاعَة السَّادِسَة مَسَاءً </s>
    <s> دَخَلَ صُبحِي إلى المَسجِد </s>
    <s> دَخَلَ إياد إلى المَنزِل </s>
    <s> دَخَلَ صُبحِي إلى المَنزِل عِندَ السَّاعَة العَاشِرَة </s>
    <s> دَخَلَ أُسامَة إلى المَسجِد فِي السَّاعَة الثَالِثَة </s>
    <s> تَأخَّرَ إياد فِي الكُلِّيَة </s>
    <s> تَأخَّرَ أُسامَة فِي الكُلِّيَة كَثيرَاً </s>
    <s> تَأخَّرَ صُبحِي عَن الاِمتِحان </s>
    <s> تَأخَّرَ أُسامَة عَن المُحاضَرَة كَثيرَاً </s>
    <s> ذَهَبَ إياد إلى حَمَاه </s>
    <s> خَرَجَ إياد مِن حِمص السَّاعَة الثَالِثَة </s>
    <s> خَرَجَ صُبحِي مِن حَلَب </s>
    <s> كَانَت المُحاضَرَة سَهلَة فِي القَاعَة الثَانِيَة </s>
    <s> تَأخَّرَ إياد عَن المَوعِد </s>
    <s> تَأخَّرَ صُبحِي و إياد عَن المَوعِد </s>
    <s> كُلِّيَة الهَندَسَة المَعلُومَاتِيَّة فِي جَامِعَة حَلَب تُرَحِّب بِكُم </s>
    <s> كُلِّيَة الهَندَسَة المَعلُومَاتِيَّة تُرَحِّب بِكُم </s>
    <s> كُلِّيَة المَعلُومَاتِيَّة فِي جَامِعَة حِمص تُرَحِّب بِكُم و أهلَاً و سَهلَاً بِكُم </s>
    <s> جَامِعَة دِمَشق تُرَحِّب بِكُم </s>
    <s> أهلَاً و سَهلَاً بِكُم </s>
    <s> أهلَاً و سَهلَاً بِكُم فِي جَامِعَة حَلَب </s>
    <s> أهلَاً و سَهلَاً بِكُم فِي العَمَل الجَدِيد </s>
    <s> جَامِعَة حَلَب تُرَحِّب بِكُم </s>
    <s> جَامِعَة حِمص تُرَحِّب بِكُم </s>
    <s> أهلَاً بِكُم فِي جَامِعَة دِمَشق </s>
    <s> أهلَاً بِكُم فِي مَدِينَة حَمَاه </s>
    <s> اِشتَرَى أُسامَة جَوَّال فِي السَّاعَة التَّاسِعَة مَسَاءً </s>
    <s> اِشتَرَى إياد جَوَّال أُسامَة و ذَهَبَ إلى صُبحِي </s>
    <s> اِشتَرَى صُبحِي جَوَّال إياد </s>
    <s> أدرُسُ فِي كُلِّيَة الهَندَسَة المَعلُومَاتِيَّة </s>
    <s> أدرُسُ فِي الكُلِّيَة كَثيرَاً </s>
    <s> أعمَلُ فِي الكُلِّيَة كَثيرَاً </s>
    <s> ذَهَبَ أُسامَة إلى حِمص فِي العُطلَة </s>
    <s> خَرَجَ صُبحِي فِي العُطلَة إلى الحَفلَة </s>
    <s> خَرَجَ الدُّكتُور حَيَّان فِي العُطلَة إلى مَدِينَة دِمَشق </s>
    <s> دَخَلَ إياد إلى القَاعَة فِي التَّاسِعَة صَباحَاً </s>
    <s> خَرَجَ صُبحِي مِن القَاعَة عِندَمَا دَخَلَ الدُّكتُور </s>
    <s> عِندَمَا اِتَّصَلَ أُسامَة خَرَجَ إياد مِن الكُلِّيَة </s>
    <s> عِندَمَا اِتَّصَلَ إياد دَخَلَ صُبحِي إلى القَاعَة </s>
    <s> عِندَمَا اِتَّصَلَ صُبحِي دَخَلَ إياد إلى البَيت </s>
    <s> عِندَمَا تَأخَّرَ أُسامَة عَن المَوعِد ذَهَبَ إياد إلى البَيت </s>
    <s> تَأخَّرَ أُسامَة كَثيرَاً عَن المَوعِد </s>
    <s> ذَهَبَ أُسامَة إلى حَمَاه كَثيرَاً </s>
    <s> دَخَلَ إياد إلى الحَديقَة صَباحَاً </s>
    <s> تَفَوَّقَ أُسامَة فِي الدِّرَاسَة </s>
    <s> كَانَت الدِّرَاسَة عِندَ إياد صَعبَة </s>
    <s> أدرُسُ فِي كُلِّيَة الهَندَسَة فِي الجَامِعَة </s>
    <s> ذَهَبَ إياد إلى العَمَل </s>
    <s> اِنتَظِرنِي فِي الكُلِّيَة </s>
    <s> اِنتَظِرنِي فِي الكُلِّيَة فِي السَّاعَة الوَاحِدَة صَباحَاً </s>
    <s> اِنتَظِرنِي فِي المَسجِد فِي السَّاعَة الرَابِعَة مَسَاءً </s>
    <s> أدرُسُ فِي كُلِّيَة الهَندَسَة المَعلُومَاتِيَّة فِي هَندَسِة البَرمَجِيَّات </s>
    <s> أعُودُ إلى العَمَل فِي السَّاعَة الرَابِعَة </s>
    <s> نَجَحتُ فِي المَادَّة </s>
    <s> نَجَحتُ فِي المَادَّة عِندَ الدُّكتُور حَيَّان </s>
    <s> تَفَوَّقَ أُسامَة فِي هَندَسِة البَرمَجِيَّات </s>
    <s> تَفَوَّقَ صُبحِي فِي الجَامِعَة </s>
    <s> ذَهَبَ أُسامَة و إياد و صُبحِي إلى الحَفلَة </s>
    <s> نَسيتُ الحَاسُوب فِي المَنزِل </s>
    <s> نَسيتُ المُحاضَرَة فِي البَيت </s>
    <s> دَخَلَ صُبحِي إلى المُحاضَرَة عِندَ السَّاعَة التَّاسِعَة صَباحَاً </s>
    <s> أُحِبُّ العَمَل عِندَ إياد </s>
    <s> نَسيتُ الحَاسُوب عِندَ صُبحِي </s>
    <s> سوفَ أدرُسُ فِي الكُلِّيَة حَتَّى السَّاعَة الرَابِعَة </s>
    <s> اِشتَرَى صُبحِي المُحاضَرَة السَّادِسَة </s>
    <s> اِشتَرَى أُسامَة الحَاسُوب فِي السَّاعَة التَّاسِعَة </s>
    <s> اِشتَرَى صُبحِي جَوَّال جَدِيد </s>
    <s> اِشتَرَى إياد الأسئِلَة مِن الكُلِّيَة </s>
    <s> اِشتَرَى أُسامَة الأسئِلَة و ذَهَبَ إلى البَيت </s>
    <s> نَجَحتُ فِي هَندَسِة البَرمَجِيَّات </s>
    <s> سوفَ أعُودُ إلى المَنزِل بَعدَ سَاعَة </s>
    <s> كَانَت الأسئِلَة فِي الاِمتِحان سَهلَة </s>
    <s> كَانَت المُحاضَرَة الرَابِعَة سَهلَة </s>
    <s> كَانَت المُحاضَرَة فِي القَاعَة الثَالِثَة </s>
    <s> اِنتَظِرنِي عِندَ السَّاعَة العَاشِرَة </s>
    <s> اِنتَظِرنِي عِندَ السَّاعَة الثَامِنَة صَباحَاً فِي الكُلِّيَة </s>
    <s> أُحِبُّ جَوَّال أُسامَة كَثيرَاً </s>
    <s> أهلَاً بِكُم فِي المَادَّة الجَدِيدة </s>
    <s> أهلَاً بِكُم فِي حَلَب </s>
    <s> أهلَاً بِكُم فِي مَدِينَة حَلَب </s>
    <s> أهلَاً بِكُم فِي حِمص </s>
    <s> أُحِبُّ العَمَل مَسَاءً </s>
    <s> أُحِبُّ الدِّرَاسَة صَباحَاً فِي الكُلِّيَة </s>
    <s> أُحِبُّ العَمَل الجَدِيد </s>
    <s> أُحِبُّ الدُّكتُور الجَدِيد فِي جَامِعَة حَلَب </s>
    <s> أُحِبُّ الدُّكتُور حَيَّان كَثيرَاً </s>
    <s> دَخَلَ الدُّكتُور حَيَّان إلى السُّوق فِي السَّاعَة الرَابِعَة </s>
    <s> تَأخَّرَ الدُّكتُور حَيَّان عَن المُحاضَرَة السَّادِسَة </s>
    <s> ذَهَبَ صُبحِي إلى الحَديقَة صَباحَاً </s>
    <s> ذَهَبَ إياد و أُسامَة إلى الحَديقَة مَسَاءً </s>
    <s> تَأخَّرَ إياد حَتَّى السَّاعَة الرَابِعَة </s>
    <s> تَأخَّرَ الدُّكتُور عَن المُحاضَرَة حَتَّى السَّاعَة التَّاسِعَة صَباحَاً </s>
    <s> تَأخَّرَ الدُّكتُور عَن المُحاضَرَة حَتَّى السَّاعَة الوَاحِدَة </s>
    <s> أعمَلُ مَع الدُّكتُور حَيَّان فِي الكُلِّيَة </s>
    <s> أعمَلُ مَع إياد و صُبحِي فِي الكُلِّيَة </s>
    <s> أُحِبُّ المَعلُومَاتِيَّة كَثيرَاً </s>
    <s> دَخَلَ الدُّكتُور حَيَّان إلى القَاعَة الرَابِعَة </s>
    <s> نَسيتُ المَوعِد مَع الدُّكتُور حَيَّان </s>
    <s> نَسيتُ المَوعِد مَع إياد </s>
    <s> نَسيتُ الكِتَاب عِندَ إياد </s>
    <s> نَسيتُ المَوعِد مَع أُسامَة فِي السَّاعَة السَّابِعَة </s>
    <s> أُحِبُّ الدِّرَاسَة فِي السَّاعَة السَّابِعَة صَباحَاً </s>
    <s> أدرُسُ فِي الكُلِّيَة فِي السَّاعَة السَّادِسَة مَسَاءً </s>
    <s> ذَهَبَ أُسامَة إلى مَدِينَة حَلَب </s>
    <s> دَخَلَ إياد إلى مَدِينَة حَمَاه </s>
    <s> خَرَجَ صُبحِي و إياد مِن العَمَل الجَدِيد </s>
    <s> أُحِبُّ الحَاسُوب الجَدِيد كَثيرَاً </s>
    <s> أدرُسُ المُحاضَرَة مِن الحَاسُوب </s>
    <s> دَخَلَ صُبحِي إلى مَدِينَة دِمَشق صَباحَاً </s>
    <s> ذَهَبَ أُسامَة إلى المَنزِل بَعدَ الكُلِّيَة </s>
    <s> بَعدَ العَمَل أعُودُ إلى المَنزِل </s>
    <s> أعُودُ إلى البَيت بَعدَ الدِّرَاسَة عِندَ صُبحِي </s>
    <s> ذَهَبَ صُبحِي إلى الحَفلَة بَعدَ الدِّرَاسَة </s>
    <s> سوفَ أُحِبُّ هَندَسِة البَرمَجِيَّات بَعدَ الاِمتِحان </s>
    <s> سوفَ أعُودُ مِن حَلَب بَعدَ سَاعَة </s>
    <s> سوفَ أعُودُ مِن حَمَاه عِندَ السَّاعَة الثَامِنَة صَباحَاً </s>
    <s> سوفَ أعُودُ مِن مَدِينَة حَلَب بَعدَ الحَفلَة </s>
    <s> دَخَلَ أُسامَة إلى الكُلِّيَة الجَدِيدة </s>
    <s> ذَهَبَ أُسامَة إلى الحَديقَة الجَدِيدة </s>
    <s> سوفَ أدرُسُ مَع إياد و صُبحِي المُحاضَرَة التَّاسِعَة </s>
    <s> سوفَ أدرُسُ المُحاضَرَة الخَامِسَة عِندَ السَّاعَة الثَالِثَة </s>
    <s> سوفَ أدرُسُ مَع إياد المُحاضَرَة التَّاسِعَة و المُحاضَرَة الثَامِنَة </s>
    <s> دَخَلَ الدُّكتُور حَيَّان إلى الحَفلَة مَسَاءً </s>
    <s> كَانَت الأسئِلَة فِي الاِمتِحان صَعبَة </s>
    <s> كَانَت الدِّرَاسَة مَع أُسامَة صَعبَة </s>
    <s> كَانَت الدِّرَاسَة مَع إياد سَهلَة </s>
    <s> نَجَحتُ فِي المَادَّة الجَدِيدة بَعدَ الدِّرَاسَة </s>
    <s> نَجَحتُ فِي الاِمتِحان بَعدَ الدِّرَاسَة </s>
    <s> المَسجِد الجَدِيد فِي مَدِينَة حِمص </s>
    <s> نَجَحَ أُسامَة فِي الاِمتِحان </s>
    <s> نَجَحَ أُسامَة و إياد و صُبحِي فِي الاِمتِحان </s>
    <s> ذَهَبَ صُبحِي إلى المَسجِد فِي مَدِينَة حَمَاه </s>
    <s> نَجَحَ صُبحِي و أُسامَة و إياد </s>
    <s> نَجَحَ الدُّكتُور حَيَّان فِي الدِّرَاسَة </s>
    <s> أعُودُ مِن المَسجِد صَباحَاً </s>
    <s> اِتَّصَلَ أُسامَة مِن جَوَّال إياد </s>
    <s> جَوَّال صُبحِي جَدِيد </s>
    <s> جَوَّال أُسامَة جَدِيد </s>
    <s> تَأخَّرَ أُسامَة سَاعَة عَن المَوعِد </s>
    <s> دَخَلَ صُبحِي إلى الكُلِّيَة بَعدَ سَاعَة </s>
    <s> بَعدَ سَاعَة مِن المُحاضَرَة خَرَجَ إياد مِن القَاعَة </s>
    <s> كَانَت المُحاضَرَة الجَدِيدة صَعبَة </s>
    <s> أعمَلُ مَع إياد فِي الجَامِعَة </s>
    <s> أعمَلُ مَع إياد فِي السَّاعَة السَّادِسَة </s>
    <s> أعمَلُ مَع صُبحِي و إياد فِي السُّوق </s>
    <s> ذَهَبَ أُسامَة إلى السُّوق فِي السَّاعَة الثَانِيَة </s>
    <s> دَخَلَ صُبحِي إلى السُّوق عِندَ السَّاعَة السَّابِعَة </s>
    <s> أُحِبُّ العُطلَة بَعدَ الاِمتِحان </s>
    <s> كَانَت الأسئِلَة سَهلَة </s>
    <s> خَرَجَ إياد فِي العُطلَة إلى حَمَاه </s>
    <s> خَرَجَ إياد فِي العُطلَة عِندَ السَّاعَة الثَانِيَة </s>
    <s> خَرَجَ صُبحِي فِي العُطلَة عِندَ السَّاعَة الثَامِنَة صَباحَاً </s>
    <s> دَخَلَ الدُّكتُور إلى الكُلِّيَة عِندَ السَّاعَة الخَامِسَة </s>
    <s> خَرَجَ أُسامَة مِن البَيت بَعدَ السَّاعَة الوَاحِدَة </s>
    <s> خَرَجَ صُبحِي مِن البَيت بَعدَ السَّاعَة العَاشِرَة مَسَاءً </s>
    <s> خَرَجَ إياد مِن البَيت بَعدَ السَّاعَة الخَامِسَة مَسَاءً </s>
    

    I repeated every word by average 10 and I recorded these sentences by 4 people
    and I get 3% word error rate and 10 % sentence error rate
    is it sufficient to do that or should I record more sounds ?
    can I send you the result to take overall look on it ( I mean the model
    parameter and model architecture and wav file ) all the database file ?

     
  • Nickolay V. Shmyrev

    is it sufficient to do that or should I record more sounds ?

    Suffucient for what. Your requirements are not clear.

    can I send you the result to take overall look on it ( I mean the model
    parameter and model architecture and wav file ) all the database file ?

    Before sending something to look on you need to describe the problem.

     
  • forsubhi

    forsubhi - 2012-04-28

    I mean I have 5 speaker and each speaker record all the 160 sentences then the
    error rate become 3% word and 10%sent
    but in documentation I see that we should record 20 hours , do you thing that
    this is sufficient ?

     
  • Nickolay V. Shmyrev

    I have 5 speaker and each speaker record all the 160 sentences then the
    error rate become 3% word and 10%sent

    It seems you didn't follow proper instructions about the test set, and you
    test on the same audio you train. If you test properly error rate will be
    bigger

    documentation I see that we should record 20 hours

    Documentation is correct if you are asking about that.

    do you thing that this is sufficient

    Sufficiency is only defined by requirements you have which I do not know.

     
  • forsubhi

    forsubhi - 2012-04-30

    my requirements is that I have graduate project about speech recognition and I
    want to make 86 word available for multi speaker , I notice that rm1 database
    has about 1200 words with only 1600 sentence and only 80 speaker and this is
    not like documentation , can you explain why they do that ?

     
  • forsubhi

    forsubhi - 2012-04-30

    can I get samples of rm1 wav files to know what is the quality of recording
    that they use ?

     
  • Nickolay V. Shmyrev

    can you explain why they do that ?

    In 1989 when RM1 was designed it was not really easy to record more than they
    recorded. RM1 is not sufficient for real dictation though.

    can I get samples of rm1 wav files to know what is the quality of recording
    that they use ?

    The quality is similar to the quality in an4 database.

     

Log in to post a comment.