similar to: Streaming for ASR

Matt Riddell wrote: > >> On 17/10/2016, at 3:43 PM, Luca Pradovera <luca.pradovera at gmail.com >> <mailto:luca.pradovera at gmail.com>> wrote: >> >> I have been working on designs for two different projects, where both >> of them would need to use the IBM Watson streaming ASR service. >> >> Would it be possible to write out the audio frames

Hello, (sorry for not continuing the thread, I had set the list to digest). Would UnicastRTP be able to output u-law frames directly? If so, I think that is all I need. Does anyone know what the EAGI output is? Raw RTP? Best regards, Luca -------------- next part -------------- An HTML attachment was scrubbed... URL:

Hello Luca, Thank you for your response. I?m familiar with speech recognition and TTS, but new to MRCP. Yes, the 100k options is used for names in a directory listing. In the pre-MRCP support, Nuance ASR used API events/methods for the application to tell ASR when the prompt was playing and when it stopped. If ASR detected speech, it would signal an event so we would stop playing the prompt.

Is there any program Asterisk users use to calculate ASR and ACD ?? Thanks for any comments. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20090910/af1f9656/attachment.htm

Is it correct that the unimrcp is the best approach for Asterisk and ASR/TTS? Could anyone provide pros/cons for the various ASR options for Asterisk? We need the ability for very large grammars (over 100,000 options). Because of this, my initial thought is Nuance or Lumenvox. Does this sound correct? Have a great day! Dan -------------- next part -------------- An HTML attachment was

Ah, no, you misunderstand. Asterisk wouldn't care one little bit what is on the page - Chromevox would do all that. A screenreader usually tabs or arrows their way about, selecting headings to read content. Thus, Asterisk ONLY needs to be able to hear content FROM the browser and pipe it to the channel, and pass keypresses back TO the browser. The human is the parser, if that makes sense?

Hello, All of the Opus quality studies that I've seen focused on human-perceived quality. I'm interested to know of any experience with machined "perceived" quality, particularly related to speech recognition or biometrics. I'm also interested in folks thoughts on optimizing Opus for ASR. For example, removing certain classes of comfort noise, filtering non-speech bands,

Hello, I'm in the early stages of designing an Emergency calling service IVR application. The IVR application asks simple one or two questions like "which is the postal code of the area you are currently calling from ?" "Is the correct ?". The expected values are a 5-digits number like "twenty-five-thousand-two-hundreds-twelve" or

Hi all Arnaud, Some newly purchased Ellipse ASR-model UPS find their way to my office for a first evaluation test. It concerns the Ellipse 1500 and Ellipse 1000 models. As you know we are still working with hidups. Peripheral recognition, start of NUT, mains short interruptions management, all that work fine... as expected ;-) But changes occured on the delayed shutdown procedure after the

Hello, Question for developers: what is the asterisk way to integrate with ASR (speech recognition)? Question to the community: has someone done anything in this direction? On the first glance that shouldn't be too hard. One part is delivering audio to the engine (for example, main ASR players Nuance and Speechworks will be happy with RTP) - can be done via RTP forking. The other part is

Is there a ready add-on to asterisk that will display the ACD/ASR per channel, source & destination? Thanks. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20091014/48076c6b/attachment.htm

I'm having a wierd problem. Somewhere around 1-2% of the time, the first DTMF digit dialed gets dropped. This is occurring during a SpeechBackground application call. If the caller reenters the digits when given a second chance, all is OK. Any suggestions how to debug this intermittent problem?

Dear list, I am looking for an open source SIP client(or any SDK) that can work on a browser. It may be based html5, javascript, flash, adobe air. I have done some research myself and I would like to ask the community if they have any further hints for me. Real life experience would be awesome. Thanks, Regards, Arstan Jusupov -------------- next part -------------- An HTML attachment was

For the last couple months, Nuance has performed extensive testing on how the Opus codec performs in the speech recognition task. I'm hoping to publish a full report in the coming months, but until then all I have is a teaser. Opus performed within about 1% of the WER (Word Error Rate) of unencoded audio. This is compared to about 5% for Speex, which was the previous codec of choice. Well

>>I'm not sure that your question is clear. You'll probably want to be more specific. >> What is pulse? You mention "as a user", are you talking about voicepulse.com ? >> What are you trying to do with pulse? >> What problem are you running into? Sorry Rusty... I am trying to get Asterisk 11 to co-exist with a centos 7 box that has pulse audio running as

Hello everybody, Our supoprt team is facing a problem which froze the CPU. They suspect that this happens after an automatic battery test, which lets the battery goes down to 30% and then stop the test. At this level of charge, our kernel also triggers a shutdown (CPU and UPS). I saw (in the doc found here : tools.*mge*ops.net/download/intl/products/*ellipse*-asr/28-umeasr_08.pdf) that this

I'm having a lot of problem with it recognizing "oh" for zero. I've tried both "o" and "oh". In one case, I just tried: $digit = o { out = "0"; } | fundamental {out = "2"; }; So I gave it a choice that was VERY far away from what I said. But when I said "o o o o o", more than 75% of the time, it had a bunch of

Hello, I have Ellipse ASR 1500 UPS and now I'm trying to monutor it with NUT from FreeBSD. It seems that there are some problems with this UPS: =================================== ugen0: MGE UPS SYSTEMS ELLIPSE, rev 1.10/42.41, addr 2 > usbdevs -v Controller /dev/usb0: addr 1: full speed, self powered, config 1, OHCI root hub(0x0000), NEC(0x0000), rev 1.00 port 1 powered port 2

Hello guys, I have to manage an Eaton Ellipse ASR 750 VA UPS under a FreeBSD 8.2 server. I used nut from ports collection but I'm unable to detect the UPS by USB cable. Here is the output of 'usbconfig -u 4 -a 2 dump_device_desc' : ugen4.2: <ELLIPSE EATON> at usbus4, cfg=0 md=HOST spd=LOW (1.5Mbps) pwr=ON bLength = 0x0012 bDescriptorType = 0x0001 bcdUSB = 0x0110

Is there a simple way to get call volume from a particular trunk within the dialplan (for conditional branching)? I suspect we will have to build an AGI script but I'm hoping something new in Asterisk 13 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20150627/6774c750/attachment.html>