Let me start by saying, yes, I've attempted to read all the documentation. So...
I want to develop a device that recognizes a total 19 words, spoken three at a time as a command, and then acts on the command. It appears to me that I could use a cheap Android phone as the microphone and processor to do all heavy lifting of the speach recognition, and then send a command message via BlueTooth to a second device, thereby saving a lot of hardware development time. Also, the controlled device needn't need anywhere near the processing power, RAM, battery, etc. So, some questions:
1. Does this sound reasonable?
2. Any thoughts on what CPU level in an Android phone has enough power to do this level of speech recognition?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes, but overall Android phones are pretty closed-source platform and it is not easy to properly program them. For example, proper protection from sleep is not easy if you want to listen continuously. Some Linux system might be a better option probably.
Any thoughts on what CPU level in an Android phone has enough power to do this level of speech recognition?
400Mhz single core ARM CPU should be enough for that, most of the phones have higher computing power.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi, and thanks for the comments. You bring up some interesting points.
Some Linux system might be a better option probably.
Do you mean scrapping the OS on a phone and replacing it with a Linux version?
Or are you suggesting I use a completely different physical platform?
Remember, I'm trying to shortcut the hardware development. Mobile phones are, for the most part, expertly designed embedded systems. Battery, charging system, Bluetooth, screen, etc., all integrated for a bargain price. Makes sense to build from there.
So now I'm wondering about the entire reason for porting Sphinx to Android. Was the intent to have a voice-controlled interface to an application that you start, use and then stop?
Last edit: Focus Research 2016-12-23
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
So now I'm wondering about the entire reason for porting Sphinx to Android. Was the intent to have a voice-controlled interface to an application that you start, use and then stop?
There was no reason
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Let me start by saying, yes, I've attempted to read all the documentation. So...
I want to develop a device that recognizes a total 19 words, spoken three at a time as a command, and then acts on the command. It appears to me that I could use a cheap Android phone as the microphone and processor to do all heavy lifting of the speach recognition, and then send a command message via BlueTooth to a second device, thereby saving a lot of hardware development time. Also, the controlled device needn't need anywhere near the processing power, RAM, battery, etc. So, some questions:
1. Does this sound reasonable?
2. Any thoughts on what CPU level in an Android phone has enough power to do this level of speech recognition?
Yes, but overall Android phones are pretty closed-source platform and it is not easy to properly program them. For example, proper protection from sleep is not easy if you want to listen continuously. Some Linux system might be a better option probably.
400Mhz single core ARM CPU should be enough for that, most of the phones have higher computing power.
Hi, and thanks for the comments. You bring up some interesting points.
Do you mean scrapping the OS on a phone and replacing it with a Linux version?
Or are you suggesting I use a completely different physical platform?
Remember, I'm trying to shortcut the hardware development. Mobile phones are, for the most part, expertly designed embedded systems. Battery, charging system, Bluetooth, screen, etc., all integrated for a bargain price. Makes sense to build from there.
So now I'm wondering about the entire reason for porting Sphinx to Android. Was the intent to have a voice-controlled interface to an application that you start, use and then stop?
Last edit: Focus Research 2016-12-23
There was no reason