Hi!
Im using the ITSM spanish acousticand language models and i get great results in tems of accuracy. However it seems that these models restrict the number of words recognized per line (one word per line), thus in order to get several words recognized i have to wait an udesirable amout of time between words when speaking.
From the arquitecture of sphinx i imagine the problem is on the searchgraph provided by the linguist when using the spanish models, but i found no way of fixing this.
Any help will be really apreciated!!
Thanks and excuse me for my english!
Santiago
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Well i finally found the problem, you wont believe it, there was an "*" missing at the end of the line in the .gram file. Now it display several words per line in spanish too.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
And how should I start your test then? Did you try to start it yourself before uploading? Or should I spend a half of an hour trying to setup everything properly.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Don´t know what you mean by test, i just modified the config.xml, trained the models and included the jar files, after that i started using the application. If there´s something else to be done in order to simplify things for you please let me know.
I realise you are helping me and i really apreciate it, it is not my intention to make you do things i could do myself just to save time. Forgive me for my ignorance, im just a beginner at this.
Thanks.
Santiago
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
> else to be done in order to simplify things for you please let me know
You use some application, WavFile.jar or something like that. Upload it and give me a way to run it and reproduce your problem. In last archive you only uploaded library jars, there is no application to run.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I´m considering using other spanish models, since the concatenation of words is up to the languaje model. Do you know if theres some software to create personal spanish acoustic and languaje model?? or complete models already created other from ITSM?
Thanks
Santiago.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi!
Im using the ITSM spanish acousticand language models and i get great results in tems of accuracy. However it seems that these models restrict the number of words recognized per line (one word per line), thus in order to get several words recognized i have to wait an udesirable amout of time between words when speaking.
From the arquitecture of sphinx i imagine the problem is on the searchgraph provided by the linguist when using the spanish models, but i found no way of fixing this.
Any help will be really apreciated!!
Thanks and excuse me for my english!
Santiago
Well i finally found the problem, you wont believe it, there was an "*" missing at the end of the line in the .gram file. Now it display several words per line in spanish too.
I'd recommend you to put your files in archive so we can reproduce your problem, upload them somewhere and give us a link.
My files are at http://www.files.to/get/356959/t5ta3vxwtf
I modified the hellodigits demo for my application and i kept the name.
I uploaded the demo, language and acoustic models, if there´s any other file needed please let me know.
Thanks!
Santiago
Hm, can you please upload jar files I'll be able to run and reproduce your problem with multiple words.
Jar files at http://www.files.to/get/358098/mpkdkln1yi
Thanks.
Santiago
And how should I start your test then? Did you try to start it yourself before uploading? Or should I spend a half of an hour trying to setup everything properly.
Don´t know what you mean by test, i just modified the config.xml, trained the models and included the jar files, after that i started using the application. If there´s something else to be done in order to simplify things for you please let me know.
I realise you are helping me and i really apreciate it, it is not my intention to make you do things i could do myself just to save time. Forgive me for my ignorance, im just a beginner at this.
Thanks.
Santiago
> else to be done in order to simplify things for you please let me know
You use some application, WavFile.jar or something like that. Upload it and give me a way to run it and reproduce your problem. In last archive you only uploaded library jars, there is no application to run.
I uploaded the jar files in the bin folder, i thinks these are the ones you need.
http://www.files.to/get/359811/rhm9skinwa
Thanks!
Santiago
hi, did you make the application run??
I´m considering using other spanish models, since the concatenation of words is up to the languaje model. Do you know if theres some software to create personal spanish acoustic and languaje model?? or complete models already created other from ITSM?
Thanks
Santiago.