I need to get the twilio voip audio signals from a phone number and convert the audio to text in realtime without disconnecting the call to take specific actions.
[1. call voice] -> [2. convert to text #realtime] -> [3. detect voice end] -> [4. get final string]
How can this be done without disconnecting the call?
I looked into translator.js, any suggestions?
As you've probably deciphered based upon your translator.js suggestion above, Twilio Recordings do not handle this. Even with a way to transcribe calls it will not do so in real time as you are hoping for.
With a quick search, one thing that looks pretty cool and might be helpful in your use case is the Web Speech API.
If you check it out, do let me know if it helps!
Related
I would like to use an API that checks pronunciation. Its input is audio. I would like to achieve the following: the user speaks to the microphone - an audio file is generated - and it is sent to the API. The API sends back the answer - the evaluation of the pronunciation. How can achieve this?
I would be also interested in how to display the microphone for the users.
My main aim is to make it work in a browser.
Thank you very much for your answer.
I couldn't set up the audio recording yet.
If I try to load back 2 back clips, it appears that they interupt each other. The documentation seems to suggest that by using audioClipStatus you get an array of audioClips, suggesting they can be queued.
Anyone has any info on this?
Secondly, I like the "Chime" audioclip and would love to be able to concatenate an audioClip preceeded by a Chime to announce an announcement. It would be great if there was an option like that.
One suggestion to the Sonos Dev team, perhaps allow the loadAudioClip to have an array of clips as opposed to a single clip.
Any reaction from the Sonos Dev team?
Windows Volume Mixer shows audio output for individual applications.
Using NAudio, what is the right way for me to tap into this information? I essentially want to be able to make my application say:
Always record all audio input/output. Unless otherwise specified, only keep a buffer of the last 30 seconds. Throw the test away. (I know how to do this)
When Skype, Vonage, or Ring Central plays audio for more than 5 seconds, ask the user if they want to start saving the audio. (How would I do this?)
If so, save the 30-second buffer to a file and then start recording live. (I know how to do this)
Thanks for the help!
Windows won't let you capture audio from individual applications. You can use NAudio's WasapiLoopbackCapture to capture audio from all applications.
If you just want to see audio output levels for all apps, that can be achieved with the IMMDevice APIs which NAudio has wrappers for. It doesn't come with a specific demo showing that, but there's another open source project, EarTrumpet that you could explore to see how its done.
Is it possible for an app which records the voice of the user while a phone call?
I searched for it and the AVFoundation Framework seems to stop recording when a phone call is incoming. I don't want to record the whole call, I am just interested in record what the user of the iPhone is saying.
I know that this possibility might be a safety risk, but it would be nice for e.g an interview.
In case of no possibility to do it like i want, i have to create my own VoIP connection and then i can record it, am i right?
There is no direct way or API from Apple to achieve your need of recording voice while phone call. AVAudioRecorder wont support it as well. Because it will be stopped once the call begin.
But there are work around like hook AudioUnitProcess in order to access phone call's audio streams and save it as m4a. This will work without jailbreak but not sure this will be accepted by Apple app store review team.
For more info on the work around check this SO accepted answer.
Also creating your own VoIP connection method work for recording during phone call.
There are Cydia tweaks which are able to do that. For example CallRecorder.
I am trying to get audio from the microphone and stream it across as data through socket to a node server. Since I haven't really done this before, I am confused on how to do this properly.
First of all, how do you actually stream data from a specific microphone? I need the audio to be streamed, not recorded then sent. In other words, it needs to be like a call, where the word you say is automatically sent to to the recipient rather than recorded, then sent as some sort of recording.
Second of all, how do I specify where the audio comes in? I have seen some questions about this but I couldn't find a good solution to how to do this, especially for my case, where the audio input is from the lightening audio.
I couldn't find a good example of how to do this using AVAudioSession. Is there any good resources (examples, tutorials) that I can use to help me?
Thanks!