There are a bunch of cloud-based transcription services these days (Nuance, google, watson etc) that provide voice to text, and I am working on developing a freeswitch integration. It seems to me that they don’t fit quite precisely into the mod_asr framework as there is, for instance, no concept of a speech grammar. Just wondering from the freeswitch devs, whether they would recommend that this be implemented as an asr module anyways (i.e stub out and ignore grammar) or should it really be a new type of module.
We faced a similar issue with TTS while developing a IVR based self care service. We ended up customizing mod_say for our requirement and we are currently working on an ASR on the same lines. Interfacing with cloud service would certainly be a good idea. On 22 September 2016 at 19:06, Dave Horton <daveh@beachdognet.com> wrote:
There are a bunch of cloud-based transcription services these days (Nuance, google, watson etc) that provide voice to text, and I am working on developing a freeswitch integration. It seems to me that they don’t fit quite precisely into the mod_asr framework as there is, for instance, no concept of a speech grammar. Just wondering from the freeswitch devs, whether they would recommend that this be implemented as an asr module anyways (i.e stub out and ignore grammar) or should it really be a new type of module. _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org http://www.freeswitchsolutions.com
Official FreeSWITCH Sites http://www.freeswitch.org http://confluence.freeswitch.org http://www.cluecon.com
FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
-- Ashwin Kumar Rath
and on later stages you could make something like relatime translator :) app/module that would intercept audio from A side, Voice2Text it and play it in B side chosen language :D On Thu, Sep 22, 2016 at 3:44 PM, Ashwin Rath <ashwinrath@gmail.com> wrote:
We faced a similar issue with TTS while developing a IVR based self care service. We ended up customizing mod_say for our requirement and we are currently working on an ASR on the same lines. Interfacing with cloud service would certainly be a good idea.
On 22 September 2016 at 19:06, Dave Horton <daveh@beachdognet.com> wrote:
There are a bunch of cloud-based transcription services these days (Nuance, google, watson etc) that provide voice to text, and I am working on developing a freeswitch integration. It seems to me that they don’t fit quite precisely into the mod_asr framework as there is, for instance, no concept of a speech grammar. Just wondering from the freeswitch devs, whether they would recommend that this be implemented as an asr module anyways (i.e stub out and ignore grammar) or should it really be a new type of module. _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org http://www.freeswitchsolutions.com
Official FreeSWITCH Sites http://www.freeswitch.org http://confluence.freeswitch.org http://www.cluecon.com
FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
-- Ashwin Kumar Rath
_________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org http://www.freeswitchsolutions.com
Official FreeSWITCH Sites http://www.freeswitch.org http://confluence.freeswitch.org http://www.cluecon.com
FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
-- Regards, Mirko ¯\_(ツ)_/¯
If this is something you guys are interested in sponsoring let the FS team know! From: freeswitch-users-bounces@lists.freeswitch.org [mailto:freeswitch-users-bounces@lists.freeswitch.org] On Behalf Of Mirko Brankovic Sent: Thursday, September 22, 2016 9:00 AM To: FreeSWITCH Users Help <freeswitch-users@lists.freeswitch.org> Subject: Re: [Freeswitch-users] transcription module development and on later stages you could make something like relatime translator :) app/module that would intercept audio from A side, Voice2Text it and play it in B side chosen language :D On Thu, Sep 22, 2016 at 3:44 PM, Ashwin Rath <ashwinrath@gmail.com <mailto:ashwinrath@gmail.com> > wrote: We faced a similar issue with TTS while developing a IVR based self care service. We ended up customizing mod_say for our requirement and we are currently working on an ASR on the same lines. Interfacing with cloud service would certainly be a good idea. On 22 September 2016 at 19:06, Dave Horton <daveh@beachdognet.com <mailto:daveh@beachdognet.com> > wrote: There are a bunch of cloud-based transcription services these days (Nuance, google, watson etc) that provide voice to text, and I am working on developing a freeswitch integration. It seems to me that they don’t fit quite precisely into the mod_asr framework as there is, for instance, no concept of a speech grammar. Just wondering from the freeswitch devs, whether they would recommend that this be implemented as an asr module anyways (i.e stub out and ignore grammar) or should it really be a new type of module. _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org <mailto:consulting@freeswitch.org> http://www.freeswitchsolutions.com Official FreeSWITCH Sites http://www.freeswitch.org http://confluence.freeswitch.org http://www.cluecon.com FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org <mailto:FreeSWITCH-users@lists.freeswitch.org> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org -- Ashwin Kumar Rath _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org <mailto:consulting@freeswitch.org> http://www.freeswitchsolutions.com Official FreeSWITCH Sites http://www.freeswitch.org http://confluence.freeswitch.org http://www.cluecon.com FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org <mailto:FreeSWITCH-users@lists.freeswitch.org> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org -- Regards, Mirko ¯\_(ツ)_/¯
As far as TTS coud-based services, I have implemented for instance an integration to Nuance NCS, and that does not present the same problem — it fits nicely into the freeswitch asr_tts framework conceptual model. On the ASR side, though, these cloud-based services seem to be dispensing with sophisticated speech grammars in favor of keyword spotting or just simple transcription with nothing resembling a grammar. On Sep 22, 2016, at 9:44 AM, Ashwin Rath <ashwinrath@gmail.com> wrote: We faced a similar issue with TTS while developing a IVR based self care service. We ended up customizing mod_say for our requirement and we are currently working on an ASR on the same lines. Interfacing with cloud service would certainly be a good idea. On 22 September 2016 at 19:06, Dave Horton <daveh@beachdognet.com <mailto:daveh@beachdognet.com>> wrote: There are a bunch of cloud-based transcription services these days (Nuance, google, watson etc) that provide voice to text, and I am working on developing a freeswitch integration. It seems to me that they don’t fit quite precisely into the mod_asr framework as there is, for instance, no concept of a speech grammar. Just wondering from the freeswitch devs, whether they would recommend that this be implemented as an asr module anyways (i.e stub out and ignore grammar) or should it really be a new type of module. _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org <mailto:consulting@freeswitch.org> http://www.freeswitchsolutions.com <http://www.freeswitchsolutions.com/> Official FreeSWITCH Sites http://www.freeswitch.org <http://www.freeswitch.org/> http://confluence.freeswitch.org <http://confluence.freeswitch.org/> http://www.cluecon.com <http://www.cluecon.com/> FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org <mailto:FreeSWITCH-users@lists.freeswitch.org> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users <http://lists.freeswitch.org/mailman/listinfo/freeswitch-users> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users <http://lists.freeswitch.org/mailman/options/freeswitch-users> http://www.freeswitch.org <http://www.freeswitch.org/> -- Ashwin Kumar Rath _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org http://www.freeswitchsolutions.com Official FreeSWITCH Sites http://www.freeswitch.org http://confluence.freeswitch.org http://www.cluecon.com FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
participants (4)
-
Ashwin Rath
-
Dave Horton
-
Ken Rice
-
Mirko Brankovic