BBCode Helper

Job Description

An artificial speech - generating device is featured in Microsofts latest patent - Neowin A VoIP (voice over Internet Protocol) is generally established in systems to enable users to communicate through audio messages over a network. There could be many reasons why such call sessions suffer from quality issues , for example , packet loss and network jitters. As these issues are usually network - related , it is quite difficult to improve the quality of these sessions.Microsoft believes that the current methods in place for the mitigation of such problems are insufficient. For this reason , the company has conceptualized a device which would detect instances where network issues might be affecting the quality of voice calls. After detection , it would then begin converting transmitted voice information into packets of text , sending these to a receiving device. In this way , the text could be viewed on a visual display , separate from the speakers being used to output audio. As all this would be done in real time , the receiver would be able to immediately comprehend the message despite poor networking conditions.According to another implementation , the transmitted text packets could also be converted back to audio output at the receiving end , a process which is known as speech synthesis. A visual cue could be used to notify users when this procedure is occurring , or the user - specific text - to - speech voice could be played on a separate channel as well. Interestingly , these implementations could also , theoretically , be extended to video content. For example , videos captured at the senders device may be encoded and sent over the network to the receiving terminal. There , they could be decoded by the device on the receiving end , before being viewed.The notion of such a device does seem to be rather attractive , particularly for those constantly plagued with network issues. However , there is no guessing if the presented idea will ever actually materialize , as is generally the case with hundreds of patents that are published daily.How can you patent something that has already existed for 20+ years Speech synthesis software existed for window 3.1 , utterly ridiculous you can patent such an idea just because you are repurposing it for a slightly different use. Thank goodness this patent is only valid in americaIts not speech synthesis. The algorithm analysis data packets to determine what the person said and convert it to text. It really just an advance speech to text system that is designed to be more reliable. I think patents like this arent sought to implement right away. Theyre designed to be used in the future to go after companies that might have implemented the same idea in a similar way , to make them pay a license fee.Not about speech synthesis - this is sending recognized textual information along with sound data.Also , speech synthesis is much older than Windows 3.1. Microsoft had the WSS in the Win 3.1 timeframe with both speech synthesis and voice recognition , and neither technology was new or unique to Microsoft beyond how the APIs and hardware worked. Im pretty sure there is mechanical speech synthesis that goes back to the 19th century and possibly sooner.However , even if this was just voice synthesis , Microsoft does have patents and specific innovations in how to do speech synthesis that is unique. For example , there is a reason Cortana sounds more like a person than even Microsofts own generic Speech API and voice technologies. So there are ways to do speech synthesis with new innovations that are different enough to deserve patents.Well speech to text isnt new either , apple had it with siri , google had it with whatever they call it , and Im sure other companies used similar technologies years ago. Nothing in this patent is new or was created by microsoft , its a joke that they can be awarded this.I dont think this article elaborates what this patent really is. but I believe what this patent is not to convert speech to text only but to predict and rebuild the speech it self while is in realtime meaning rebuilding missing part of audio using synthesized speech probably using machine learning techniques. that is definitely newPlease enter your reason for reporting this comment.The following codes can be used in comments.,

Keyskills :
windows codes voip api it wss design output udiomastering speechsynthesis machinelearning algorithmanalysis