I am trying to use pocketsphinx but I get about 50% accuracy with it's default continuous.c program. But when I try google voice recognizer, or cortana I get about 100% accuracy!
It's strange for me that why this happens? Do google or microsoft use better language models? better acoustic models? better dictionaries? or they are better engines than CMU toolkit?
So I am looking for a way to use google or microsoft voice recognizers in offline mode! is it possible?
Are there googles language model or acoustic model to use with CMU pocketsphinx?
Related
We are trying to use Project Tango in our prototype/future product. We are open to use either Asus or Lenovo.
Does anyone know what the licensing is? Or is it now in the hands of Asus/Lenovo?
you would need to contact google as tango is their product and lenovo/asus license it from google. FYI, the future of tango is somewhat up in the air currently as google has said little about the program and have seriously reduced the updates to the handsets and code. you would be better off if you tried ARCore by google which seems to be the way forward with AR as far a google is concerned.
Being that Im a C++ Developer, Curious to know if its possible to develop Voice UI and Speech recoginition Software like Amazon alexa but using c# speech libraries to open a specific Software through voice commands?
Can someone give some insight or einlighten me ? I used Speech.Recognition from C# and wrote a sample Code, where in it Can take my Input, but the Problem is it coudnt understand my Input that perfectly.
Is it a good idea to continue to work on this ? I feel like totally lost. Any Inputs will be apprerciated please.
It's possible - but depends on if you want to go through the trouble of recreating all the Voice based assistants & technology already available for you to leverage.
For example, you could easily create a Skill using the Cortana Skills Kit, and enable Cortana to be the digital assistant that does the speech to text, provide context, enable deep-linking to your software etc. while you focus on the business logic/functionality. Does that help?
I'm trying to figure out a way to create a real-time chatting app with Xamarin and some kind of backend technology. How can I do it?
I have a better solution. The above-mentioned technologies(buddy SDK & SignalR) will require you to integrate and handle everything by your self. Instead, you can go for CometChat SDK, which has real-time chat along with Audio/Video chat functionalities. Also, you can use its ready-made UI inside your application. It is the easiest & quickest way for you to make a real-time chatting system. For more information on CometChat you can refer to the following link:
https://developer.cometchat.com/docs/xamarin-forms-quick-start
As mentioned by ChristiaanV, SignalR is the way to go.
The Performance is great with Xamarin, we delivered more than 150k messages within 2 hours without any problems regarding the connection.
With Azure Mobile App and ServiceBus, you can easily scale out your backend, as described here http://www.asp.net/signalr/overview/performance/scaleout-with-windows-azure-service-bus
Have a look at SignalR, it's a good and easy to use framework to do realtime 'things'. They also have a Xamarin Component which is free to use.
https://github.com/SignalR/SignalR
http://components.xamarin.com/view/signalr
Please note that I personally don't have experience with using it on mobile devices, so I can't tell you anything about how it's performing battery / performance wise.
Hi guys i googled about voicexml and google speech recognition but i didn't understand if google speech recognition model is based on voicexml model or not.
If possible, can you tell me the differences please?
Thanks
Google Voice Search (as it is sometimes called) is closed source, thus it's hard to say what it's based on or how their algorithms differ. If you click this link you'll find a list and comparison of many voice recognition implementations, including Google's and VoiceXML.
If it does have anything to do with Verbio's VoiceXML then Verbio is likely under a contractual agreement not to disclose that as there's no mention of Google anywhere on their website.
the iPhone SDK provides with MapKit a simple solution for maps und reverse geocoding without any additional costs. Does anyone know if there is a similar Framework for Cocoa on OSX? I know that google maps is capable of that but as far as I know and understand I have to pay while using google maps in a commercial application?
Thanks for your answers
twickl
You won't have to pay as long as you don't use the quota.
The api documentation is here :
http://code.google.com/apis/maps/documentation/geocoding/
If you plan to use it reasonably, this will remain free (no api key is required anymore by google)