Hello, I've been lurking around the mailing list and browsing around on Asterisk-related links while I wait for my X100P to come in the mail. So far I haven't found very much information related to what I want to do with Asterisk. I was wondering if someone could point me in the direction of any work that may already have been done on a project similar to the one I'm trying to do. I'm interested in creating an outbound dialer application that will leave voice alerts with our customers. I want it to select a list of phone numbers and accompanying text messages from a PostgreSQL database, use a TTS engine to convert the text messages into audio files, and use all available lines to send out the messages as quickly as possible. I also want to make sure that it works with an answering machine. After looking into a number of cheap dialer apps that only support half-duplex voice modems, I found Asterisk and the accompanying hardware at the Asterisk store, which I'm told supports full-duplex audio. As far as I understand it, full-duplex audio is necessary to detect answering machines well. This is what I have in mind for the answering machine detection algorithm: 1) Dial the number. 2) Wait until line is picked up. 3) Wait until 1-2 seconds of relative silence (silence threshold will require some calibration). 4) Begin leaving message. 5) If during message, noise is heard coming from the other end, stop sending message and loop back to step number 3. 6) After leaving message successfully, hang up. Anyway, if anyone could point out any work that has already been done in this regard, I would really appreciate it. Thanks, Carl Youngblood
On Thu, 2003-11-27 at 13:14, Carl Youngblood wrote:> Hello, > > I've been lurking around the mailing list and browsing around on > Asterisk-related links while I wait for my X100P to come in the mail. > So far I haven't found very much information related to what I want to > do with Asterisk. I was wondering if someone could point me in the > direction of any work that may already have been done on a project > similar to the one I'm trying to do. > > I'm interested in creating an outbound dialer application that will > leave voice alerts with our customers. I want it to select a list of > phone numbers and accompanying text messages from a PostgreSQL database, > use a TTS engine to convert the text messages into audio files, and use > all available lines to send out the messages as quickly as possible. I > also want to make sure that it works with an answering machine. After > looking into a number of cheap dialer apps that only support half-duplex > voice modems, I found Asterisk and the accompanying hardware at the > Asterisk store, which I'm told supports full-duplex audio. As far as I > understand it, full-duplex audio is necessary to detect answering > machines well. This is what I have in mind for the answering machine > detection algorithm: > > 1) Dial the number. > 2) Wait until line is picked up.Just a warning that you will not know when the line is picked up with a X100P. Later when you upgrade to T1/E1 service you will know when it is picked up.> 3) Wait until 1-2 seconds of relative silence (silence threshold will > require some calibration).Silence detection is already built into asterisk, but I think you may want to shorten that timeout.> 4) Begin leaving message. > 5) If during message, noise is heard coming from the other end, stop > sending message and loop back to step number 3.You may want to rethink this just a bit, wouldn't want to restart the message because someone coughed on the line. Maybe a limit of how far in to the message before you ignore noise.> 6) After leaving message successfully, hang up.Shouldn't be a problem. If you write your own DSP stuff, you could write this using EAGI, otherwise you will need to write it into an asterisk app to get at the already written silence detection. Either way, you could use festival for the TTS, or for certain messages, you could go ahead a record prompts to be played. This way you are sure your message will be understood.> Anyway, if anyone could point out any work that has already been done in > this regard, I would really appreciate it. > > Thanks, > > Carl Youngblood > > _______________________________________________ > Asterisk-Users mailing list > Asterisk-Users@lists.digium.com > http://lists.digium.com/mailman/listinfo/asterisk-users-- Steven Critchfield <critch@basesys.com>
We are working with realspeak and it is a wonderfull product (even in product) it supports up to 20 languages and has aquired a really good prod. stability ! ________________________________ Van: Steve Underwood [mailto:steveu@coppice.org] Verzonden: vr 28/11/2003 4:41 Aan: asterisk-users@lists.digium.com Onderwerp: Re: [Asterisk-Users] Multi-line TTS Outbound Dialer Carl Youngblood wrote:> What is EAGI? I will probably use festival for the time being, but I > thing that I would eventually like to use ScanSoft's RealSpeak SDK > because it is so life-like. Unfortunately our text alerts are fully > customizeable, so we can't pre-record them.Beware the likelike TTS, that sucks up thousands of dollars and gets thrown away. RealSpeak is great for demos :-) Regards, Steve _______________________________________________ Asterisk-Users mailing list Asterisk-Users@lists.digium.com http://lists.digium.com/mailman/listinfo/asterisk-users -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/ms-tnef Size: 4352 bytes Desc: not available Url : http://lists.digium.com/pipermail/asterisk-users/attachments/20031128/85a91f5e/attachment.bin
Michael Devenijn wrote:> We are working with realspeak and it is a wonderfull product (even in product) it supports up to 20 languages and has aquired a really good prod. stability ! > ________________________________What kind of money we talking for that product? Thx. B.
Brian Capouch wrote:> Michael Devenijn wrote: > >> We are working with realspeak and it is a wonderfull product (even in >> product) it supports up to 20 languages and has aquired a really good >> prod. stability ! >> ________________________________ > > > What kind of money we talking for that product?I think it sucks, and it is about $1000 per port. They count ports in a rather "maximal" way, too. If you have, say, 1000 ports, but a maximum of, say, 10 will be used for TTS at any time they want you to licence 1000 ports of Realspeak. Crazy, but true :-) Regards, Steve