We are using twilio video call. We need to create real time transcription for this call. How shall I do that? We tried to use twilio composition api but it's taking too long.
Tried twilio composition api to compose the separate audio files from each participant. This takes forever to create the composition. We need it near real time.
Related
I have a need to play LIVE audio to enqueued twilio calls (instead of playing hold music from an mp3 file for instance).
I've tried pointing the hold music to a live mp3 stream (icecast), which didn't work.
The only thing I can think of is to start a conference, put a call into it that is sending the audio I need to be played on hold, and then on-hold calls are placed (muted) into that conference.
Doesn't seem like best way, and I'd like to avoid conference costs though (there are millions of minutes per month of on-hold time).
Is there a more elegant solution for this problem?
Twilio developer evangelist here.
As far as I am aware this is not possible with <Play>. When you give Twilio an mp3 file to play it first downloads and caches the file (if the headers allow for it). When working with mp3 files Twilio expects an existing file and a finite file size.
I think using a conference, or a series of direct one to one calls, to play the stream as you suggested is likely the best solution. If you do have millions of minutes per month then I recommend you get in touch with the Twilio sales team who might be able to make those minutes more affordable.
My application uses Google WebRTC framework to make audio calls and that part work. However I would like to find a way to stream an audio file during a call.
Scenario :
A calls B
B answer and play a music
A hear this music
I've downloaded entire source code of WebRTC and trying to understand how it works. On the iOS part it seems that it is using Audio Unit. I can see a voice_processing_audio_unit file. I would (maybe wrongly) assume that I need to create a custom audio_unit that is reading its data from a file?
Does anyone has an idea in which direction to go?
After fighting an entire week with this issue. I finally manage to find a solution for this problem.
By editing WebRTC Code, I was able to get to the level of AudioUnits and in the AudioRenderingCallback, catch the io_data buffer.
This callback is called every 10ms to get the data from the mic. Therefor in this precise callback I was able to change this io_data buffer to put my own audio data.
I believe the answer is no, but does Twilio provide ability to pause/resume a recording? Use case is recording a call, but pausing recording when collecting sensitive information. From the REST documentation, it doesn't appear to be a supported capability. Thought someone might have found some options for this requirement.
This is possible, though it's not wholly obvious how from the documentation.
You can modify call state using the REST API, as per https://www.twilio.com/docs/api/rest/change-call-state , and we basically use it to tell the call to re-dial to the same agent (presumably this is a call centre?) but with no-record, and then again with record re-activated once we're done.
You end up with two separate recordings for the call, which in our case we download, stitch together, and store back to our storage platform.
Edit:
Having discussed this issue with Twilio support, there's another possibility which allows you to just have a single recording.
Instead of dialling the two ends of the call together, you instead put them both into a conference that's recorded when you initially connect the call. When you want to pause it, using the REST API, you add a new "hold" leg into the conference, then move the two real ends of the call onto a new conference that isn't recorded. When you're done, you move them back again and it's "unpaused". You then only have one recording from the original conference.
None of these is ideal, and apparently they are working on a proper support setup for this (fairly obvious!) requirement, but this should solve it for now.
The Recording Pause & Resume feature is now supported in the Twilio API. Here's a link that gets you started:
https://support.twilio.com/hc/en-us/articles/360010199074-Getting-Started-with-Call-Recording-Controls#pause_resume
Is it possible to record output audio in an app using Swift? So, for example, say I'm listening to a podcast, and I want to, within a separate app, record a small segment of the podcast's audio. Is there any way to do that?
I've looked around but have only been able to find information on recording microphone recording and such.
It depends on how you are producing the audio. If the production of the audio is within your control, you can put a tap on the output and record to a file as it plays. The easiest way is with the new AVAudioEngine feature (there are other ways, but AVAudioEngine is basically an easy front end for them).
Of course, if the real problem is to take a copy of a podcast, then obviously all you have to do is download the podcast as opposed to listening to it. Similarly, you could buffer and save streaming audio to a file. There are many apps that do this. But this is not because the device's output is being hijacked; it is, again, because we have control of the sound data itself.
I believe you'll have to write a kernel extension to do that
https://developer.apple.com/library/mac/documentation/Darwin/Conceptual/KEXTConcept/KEXTConceptIOKit/iokit_tutorial.html
You'd have to make your own audio driver to record it
It appears as though
That is how softonic made soundflowerbed.
http://features.en.softonic.com/how-to-record-internal-sound-on-a-mac
I want to get four related pieces of information from a caller, and get recordings of each. Do I need to implement this as four separate calls (Say/Record pairs in the XML file) to the Twilio API and my web endpoint (the Record 'action' or StatusCallback)? I'm using Python and Flask, but examples in other languages and frameworks would be helpful too.
Twilo evangelist here.
Four separate recordings is probably the easiest way to do what you want.
<Say>Begin Recording 1</Say>
<Record>
<Say>Begin Recording 2</Say>
<Record>
<Say>Begin Recording 3</Say>
<Record>
<Say>Begin Recording 4</Say>
<Record>
<Say>Thanks for recording four things</Say>
The hard part is going to be knowing that the user actually gave you the right information, so you might want to think about adding a step at the end of your call flow that allows your caller to hear their recording and choose to re-record the info if they want.
Hope that helps.