+- Hack Forums (https://hackforums.net)
+-- Forum: Tech (https://hackforums.net/forumdisplay.php?fid=444)
+--- Forum: Coding (https://hackforums.net/forumdisplay.php?fid=151)
+---- Forum: Software Development (https://hackforums.net/forumdisplay.php?fid=118)
+---- Thread: [.NET] Good Speech to Text for numbers? (/showthread.php?tid=5717347)
Good Speech to Text for numbers? - SIM-3327349 - Sep 14, 2017
I'm trying to find a good Speech to Text for numbers. Here is example .mp3 file http://www.filedropper.com/payload_8 which I need converted into text. I need 100% correct numbers converted.
Please does anyone know of something what can convert it and I can use it in VB.NET?
Thank you in advance
* - Ninja - Sep 14, 2017
This sounds familiar. This Hotmail or Google?
You can use http://bypassvoice.com/ which is reliable, or you can use the SpeechRecognitionEngine:
https://msdn.microsoft.com/en-us/library/system.speech.recognition.speechrecognitionengine(v=vs.110).aspx
* - SIM-3327349 - Sep 14, 2017
It is Google.. I already tried Microsoft SpeechRecognitionEngine, but it has so many faults, when I tried convert it.
* - Ninja - Sep 15, 2017
(Sep 14, 2017 06:48 AM)_Grow_ Wrote: It is Google.. I already tried Microsoft SpeechRecognitionEngine, but it has so many faults, when I tried convert it.
Try and find a third party library that does speech recognition. I am sure there are some.
* - SIM-3327349 - Sep 15, 2017
(Sep 15, 2017 03:47 PM)Ninja Wrote: Try and find a third party library that does speech recognition. I am sure there are some.
I tried to find one which is good, but without success.. I'm thinking of using Google Speech to Text, but it's not cheap
* - byte[] - Sep 15, 2017
(Sep 15, 2017 04:40 PM)_Grow_ Wrote: I tried to find one which is good, but without success.. I'm thinking of using Google Speech to Text, but it's not cheap
how much does it cost?
* - SIM-3327349 - Sep 15, 2017
(Sep 15, 2017 04:46 PM)byte[] Wrote: how much does it cost?
0.006$ for every 15 seconds
* - Ninja - Sep 16, 2017
(Sep 15, 2017 06:13 PM)_Grow_ Wrote: 0.006$ for every 15 seconds
This seems easily solvable. All you gotta do is remove the background noise. Try finding patterns in the speech, like when the people start speaking, and at what frequency or something, and remove the rest.
* - SIM-3327349 - Sep 17, 2017
(Sep 16, 2017 02:09 PM)Ninja Wrote: This seems easily solvable. All you gotta do is remove the background noise. Try finding patterns in the speech, like when the people start speaking, and at what frequency or something, and remove the rest.
But if I want to solve hours and hours, then it will still cost so much
* - deiski - Sep 17, 2017
(Sep 17, 2017 12:52 AM)_Grow_ Wrote: But if I want to solve hours and hours, then it will still cost so much
I was going to say this sounds promising, but you are limited to 50 a day.
* - SIM-3327349 - Sep 17, 2017
(Sep 17, 2017 05:14 AM)deiski Wrote: Something like what they did here would have been good, but it's limited to 50 a day via the free option.
Thanks for the tip, but unfortunately there is the limitation :/
* - Ninja - Sep 17, 2017
(Sep 17, 2017 12:52 AM)_Grow_ Wrote: But if I want to solve hours and hours, then it will still cost so much
What I meant is that if you remove all the background noise and such then you can just use a normal speech to text library.
* - SIM-3327349 - Sep 17, 2017
(Sep 17, 2017 07:30 AM)Ninja Wrote: What I meant is that if you remove all the background noise and such then you can just use a normal speech to text library.
Is that really possible? :-O I have no idea how to do things like that
* - Ninja - Sep 17, 2017
(Sep 17, 2017 01:23 PM)_Grow_ Wrote: Is that really possible? :-O I have no idea how to do things like that
You wont find any library or automated service (except with real humans listening like bypassvoice.com) that can do this without audio processing and training. I suggest you start Googling.
* - SIM-3327349 - Sep 17, 2017
(Sep 17, 2017 03:35 PM)Ninja Wrote: You wont find any library or automated service (except with real humans listening like bypassvoice.com) that can do this without audio processing and training. I suggest you start Googling.
I'm on it! Thank you very much