Hi all, Thank you to all contributing to FreeSWITCH ! I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding). *mod_httapi* looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts. *mod_shout* does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT). Is there a way to use these modules to achieve our goal ? If not, we are willing to author a specific module or rather contribute to the existing ones. We've identified two approaches: 1. From *mod_httapi* Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers! 2. From *mod_shout* 1. Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method 2. Modify mod_shout to support wav files 3. Implement our IVR in script (javascript/lua) What do you think ? Thank you for your help. Cheers, Vincent
Please look at webdav mounted filesystem. On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> wrote:
Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
*mod_httapi* looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
*mod_shout* does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches:
1. From *mod_httapi* Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
2. From *mod_shout* 1. Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method 2. Modify mod_shout to support wav files 3. Implement our IVR in script (javascript/lua)
What do you think ?
Thank you for your help.
Cheers, Vincent _________________________________________________________________________ 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
Hello Sergey, Thank you for your answer. I've looked into webdav mounted filesystem. Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3. Any other idea ? Thank you ! Best regards Vincent On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> wrote:
Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
*mod_httapi* looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
*mod_shout* does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches:
1. From *mod_httapi* Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
2. From *mod_shout* 1. Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method 2. Modify mod_shout to support wav files 3. Implement our IVR in script (javascript/lua)
What do you think ?
Thank you for your help.
Cheers, Vincent _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache? On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> wrote:
Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> wrote:
Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
*mod_httapi* looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
*mod_shout* does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches:
1. From *mod_httapi* Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
2. From *mod_shout* 1. Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method 2. Modify mod_shout to support wav files 3. Implement our IVR in script (javascript/lua)
What do you think ?
Thank you for your help.
Cheers, Vincent _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
webdav, mod_http_cache or mod_httapi all results in sending the recording only *after* it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP. I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout. On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> wrote:
Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> wrote:
Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
*mod_httapi* looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
*mod_shout* does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches:
1. From *mod_httapi* Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
2. From *mod_shout* 1. Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method 2. Modify mod_shout to support wav files 3. Implement our IVR in script (javascript/lua)
What do you think ?
Thank you for your help.
Cheers, Vincent
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server. For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> . After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch. Sergey вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>:
webdav, mod_http_cache or mod_httapi all results in sending the recording only *after* it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> wrote:
Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> wrote:
Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
*mod_httapi* looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
*mod_shout* does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches:
1. From *mod_httapi* Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
2. From *mod_shout* 1. Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method 2. Modify mod_shout to support wav files 3. Implement our IVR in script (javascript/lua)
What do you think ?
Thank you for your help.
Cheers, Vincent
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
Ok thanks. It looks promising ! I'll dig into mod_vlc. Best Vincent On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com> wrote:
Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> . After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>:
webdav, mod_http_cache or mod_httapi all results in sending the recording only *after* it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> wrote:
Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> wrote:
Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
*mod_httapi* looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
*mod_shout* does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches:
1. From *mod_httapi* Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
2. From *mod_shout* 1. Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method 2. Modify mod_shout to support wav files 3. Implement our IVR in script (javascript/lua)
What do you think ?
Thank you for your help.
Cheers, Vincent
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com <mailto:s.safarov@gmail.com>> wrote: Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary>. After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com <mailto:vincent.gire@gmail.com>>: webdav, mod_http_cache or mod_httapi all results in sending the recording only after it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com <mailto:s.safarov@gmail.com>> wrote: Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com <mailto:vincent.gire@gmail.com>> wrote: Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com <mailto:s.safarov@gmail.com>> wrote: Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com <mailto:vincent.gire@gmail.com>> wrote: Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
mod_httapi looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
mod_shout does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches: From mod_httapi Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
From mod_shout Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method Modify mod_shout to support wav files Implement our IVR in script (javascript/lua) What do you think ?
Thank you for your help.
Cheers,
We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ? On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com> wrote:
I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com> wrote:
Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> . After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>:
webdav, mod_http_cache or mod_httapi all results in sending the recording only *after* it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> wrote:
Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> wrote:
> Hi all, > > Thank you to all contributing to FreeSWITCH ! > > I'm working on a IVR project where logic is implemented on a HTTP > server. > We are leaving Twilio because we now need to record and live stream > the session to the HTTP server in WAV format (chunked transfer encoding). > > *mod_httapi* looks great (HT TAPI very similar to Twilio's) but it > seems that the records are first saved to disk before there are sent to the > server as chunked data. > We need the transfer to start as soon as the recording starts. > > *mod_shout* does start the request almost as the records starts but > it does not support WAV file and shout:// is not exactly a HTTP request > (SOURCE method instead of PUT). > > Is there a way to use these modules to achieve our goal ? > > If not, we are willing to author a specific module or rather > contribute to the existing ones. > > We've identified two approaches: > > 1. From *mod_httapi* Modify mod_httapi to directly stream the > record instead of completely saving it to disk before the HTTP chunked > transfer starts. > This seems the most logical but with more than 3000 lines, > mod_httapi does not seem to be the easiest module to build upon for > newcomers! > > 2. From *mod_shout* > 1. Modify libshoot to replace the custom SOURCE method with > standard HTTP PUT method > 2. Modify mod_shout to support wav files > 3. Implement our IVR in script (javascript/lua) > > What do you think ? > > Thank you for your help. > > Cheers, >
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
-- Vincent Gire
Look at http://www.unimrcp.org/ and https://wiki.freeswitch.org/wiki/Mod_unimrcp On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com> wrote:
We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ?
On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com> wrote:
I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com> wrote:
Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> . After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>:
webdav, mod_http_cache or mod_httapi all results in sending the recording only *after* it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> wrote:
Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
> Please look at webdav mounted filesystem. > > On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> > wrote: > >> Hi all, >> >> Thank you to all contributing to FreeSWITCH ! >> >> I'm working on a IVR project where logic is implemented on a HTTP >> server. >> We are leaving Twilio because we now need to record and live stream >> the session to the HTTP server in WAV format (chunked transfer encoding). >> >> *mod_httapi* looks great (HT TAPI very similar to Twilio's) but it >> seems that the records are first saved to disk before there are sent to the >> server as chunked data. >> We need the transfer to start as soon as the recording starts. >> >> *mod_shout* does start the request almost as the records starts >> but it does not support WAV file and shout:// is not exactly a HTTP request >> (SOURCE method instead of PUT). >> >> Is there a way to use these modules to achieve our goal ? >> >> If not, we are willing to author a specific module or rather >> contribute to the existing ones. >> >> We've identified two approaches: >> >> 1. From *mod_httapi* Modify mod_httapi to directly stream the >> record instead of completely saving it to disk before the HTTP chunked >> transfer starts. >> This seems the most logical but with more than 3000 lines, >> mod_httapi does not seem to be the easiest module to build upon for >> newcomers! >> >> 2. From *mod_shout* >> 1. Modify libshoot to replace the custom SOURCE method with >> standard HTTP PUT method >> 2. Modify mod_shout to support wav files >> 3. Implement our IVR in script (javascript/lua) >> >> What do you think ? >> >> Thank you for your help. >> >> Cheers, >> >
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
-- Vincent Gire _________________________________________________________________________ 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
Unimrcp isn't going to provide the interface he's talking about. Can I ask which engine this is? On Tuesday, April 5, 2016, Sergey Safarov <s.safarov@gmail.com> wrote:
Look at http://www.unimrcp.org/ and https://wiki.freeswitch.org/wiki/Mod_unimrcp
On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com <javascript:_e(%7B%7D,'cvml','vincent.gire@gmail.com');>> wrote:
We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ?
On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com <javascript:_e(%7B%7D,'cvml','mike@jerris.com');>> wrote:
I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com <javascript:_e(%7B%7D,'cvml','vincent.gire@gmail.com');>> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com <javascript:_e(%7B%7D,'cvml','s.safarov@gmail.com');>> wrote:
Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> . After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com <javascript:_e(%7B%7D,'cvml','vincent.gire@gmail.com');>>:
webdav, mod_http_cache or mod_httapi all results in sending the recording only *after* it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com <javascript:_e(%7B%7D,'cvml','s.safarov@gmail.com');>> wrote:
Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com <javascript:_e(%7B%7D,'cvml','vincent.gire@gmail.com');>> wrote:
> Hello Sergey, > > Thank you for your answer. > I've looked into webdav mounted filesystem. > > Unfortunately, most WebDav clients (especially davfs2 on debian) do > a lot of buffering, caching and even lock-null requests (lock a non > existent resource before writing to it). I also suspect that they wait for > the end of the write operation. > The result is a latency of a few seconds witch is not much better > than what I achieve with mod_shout if I transcode the MP3. > > Any other idea ? > > Thank you ! > > Best regards > > Vincent > > On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com > <javascript:_e(%7B%7D,'cvml','s.safarov@gmail.com');>> wrote: > >> Please look at webdav mounted filesystem. >> >> On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com >> <javascript:_e(%7B%7D,'cvml','vincent.gire@gmail.com');>> wrote: >> >>> Hi all, >>> >>> Thank you to all contributing to FreeSWITCH ! >>> >>> I'm working on a IVR project where logic is implemented on a HTTP >>> server. >>> We are leaving Twilio because we now need to record and live >>> stream the session to the HTTP server in WAV format (chunked transfer >>> encoding). >>> >>> *mod_httapi* looks great (HT TAPI very similar to Twilio's) but >>> it seems that the records are first saved to disk before there are sent to >>> the server as chunked data. >>> We need the transfer to start as soon as the recording starts. >>> >>> *mod_shout* does start the request almost as the records starts >>> but it does not support WAV file and shout:// is not exactly a HTTP request >>> (SOURCE method instead of PUT). >>> >>> Is there a way to use these modules to achieve our goal ? >>> >>> If not, we are willing to author a specific module or rather >>> contribute to the existing ones. >>> >>> We've identified two approaches: >>> >>> 1. From *mod_httapi* Modify mod_httapi to directly stream the >>> record instead of completely saving it to disk before the HTTP chunked >>> transfer starts. >>> This seems the most logical but with more than 3000 lines, >>> mod_httapi does not seem to be the easiest module to build upon for >>> newcomers! >>> >>> 2. From *mod_shout* >>> 1. Modify libshoot to replace the custom SOURCE method with >>> standard HTTP PUT method >>> 2. Modify mod_shout to support wav files >>> 3. Implement our IVR in script (javascript/lua) >>> >>> What do you think ? >>> >>> Thank you for your help. >>> >>> Cheers, >>> >>
Professional FreeSWITCH Consulting Services: consulting@freeswitch.org <javascript:_e(%7B%7D,'cvml','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 <javascript:_e(%7B%7D,'cvml','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
-- Vincent Gire _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting@freeswitch.org <javascript:_e(%7B%7D,'cvml','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 <javascript:_e(%7B%7D,'cvml','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
We use multiple ones to compare confidence. They all support chunk transfer and provide a HTTP transaction similar to : http://developer.att.com/apis/speech/docs/v3 Look at 3/ Make API Calls On Tue, Apr 5, 2016 at 7:14 PM, Michael Jerris <mike@jerris.com> wrote:
Unimrcp isn't going to provide the interface he's talking about. Can I ask which engine this is?
On Tuesday, April 5, 2016, Sergey Safarov <s.safarov@gmail.com> wrote:
Look at http://www.unimrcp.org/ and https://wiki.freeswitch.org/wiki/Mod_unimrcp
On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com> wrote:
We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ?
On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com> wrote:
I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com> wrote:
Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> . After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>:
webdav, mod_http_cache or mod_httapi all results in sending the recording only *after* it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com> wrote:
> Input/output latency is not problem. I use Kazoo on my servers and > call recording is stored to database during 5 seconds after hangup. > What is broken in your case if save file using webdav or http_cache? > > On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> > wrote: > >> Hello Sergey, >> >> Thank you for your answer. >> I've looked into webdav mounted filesystem. >> >> Unfortunately, most WebDav clients (especially davfs2 on debian) do >> a lot of buffering, caching and even lock-null requests (lock a non >> existent resource before writing to it). I also suspect that they wait for >> the end of the write operation. >> The result is a latency of a few seconds witch is not much better >> than what I achieve with mod_shout if I transcode the MP3. >> >> Any other idea ? >> >> Thank you ! >> >> Best regards >> >> Vincent >> >> On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com >> > wrote: >> >>> Please look at webdav mounted filesystem. >>> >>> On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> >>> wrote: >>> >>>> Hi all, >>>> >>>> Thank you to all contributing to FreeSWITCH ! >>>> >>>> I'm working on a IVR project where logic is implemented on a HTTP >>>> server. >>>> We are leaving Twilio because we now need to record and live >>>> stream the session to the HTTP server in WAV format (chunked transfer >>>> encoding). >>>> >>>> *mod_httapi* looks great (HT TAPI very similar to Twilio's) but >>>> it seems that the records are first saved to disk before there are sent to >>>> the server as chunked data. >>>> We need the transfer to start as soon as the recording starts. >>>> >>>> *mod_shout* does start the request almost as the records starts >>>> but it does not support WAV file and shout:// is not exactly a HTTP request >>>> (SOURCE method instead of PUT). >>>> >>>> Is there a way to use these modules to achieve our goal ? >>>> >>>> If not, we are willing to author a specific module or rather >>>> contribute to the existing ones. >>>> >>>> We've identified two approaches: >>>> >>>> 1. From *mod_httapi* Modify mod_httapi to directly stream the >>>> record instead of completely saving it to disk before the HTTP chunked >>>> transfer starts. >>>> This seems the most logical but with more than 3000 lines, >>>> mod_httapi does not seem to be the easiest module to build upon for >>>> newcomers! >>>> >>>> 2. From *mod_shout* >>>> 1. Modify libshoot to replace the custom SOURCE method >>>> with standard HTTP PUT method >>>> 2. Modify mod_shout to support wav files >>>> 3. Implement our IVR in script (javascript/lua) >>>> >>>> What do you think ? >>>> >>>> Thank you for your help. >>>> >>>> Cheers, >>>> >>>
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
The best way to handle this is probably to write a custom module against the speech interface that sends the streams like you are describing, and supports multiple providers. How exactly do you go about sending to multiple at the same time and combining the results, as this might actually be the trickiest part as youll need to integrate that into a module that uses the speech interface to be able to at all sanely handle that in freeswitch
On Apr 5, 2016, at 1:24 PM, Vincent Gire <vincent.gire@gmail.com> wrote:
We use multiple ones to compare confidence. They all support chunk transfer and provide a HTTP transaction similar to : http://developer.att.com/apis/speech/docs/v3 <http://developer.att.com/apis/speech/docs/v3> Look at 3/ Make API Calls
On Tue, Apr 5, 2016 at 7:14 PM, Michael Jerris <mike@jerris.com <mailto:mike@jerris.com>> wrote: Unimrcp isn't going to provide the interface he's talking about. Can I ask which engine this is?
On Tuesday, April 5, 2016, Sergey Safarov <s.safarov@gmail.com <mailto:s.safarov@gmail.com>> wrote: Look at http://www.unimrcp.org/ <http://www.unimrcp.org/> and https://wiki.freeswitch.org/wiki/Mod_unimrcp <https://wiki.freeswitch.org/wiki/Mod_unimrcp> On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com <>> wrote: We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ?
On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com <>> wrote: I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com <>> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com <>> wrote: Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary>. After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com <>>: webdav, mod_http_cache or mod_httapi all results in sending the recording only after it is complete. They all write the recording to a file, wait for the recording to complete and the file to close and then send it over HTTP.
I would like to start sending the recording to the remove server as soon as it starts (max 1 sec latency). mod_http_cache or mod_httapi would be perfect if they were streaming the recording like mod_shout.
On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com <>> wrote: Input/output latency is not problem. I use Kazoo on my servers and call recording is stored to database during 5 seconds after hangup. What is broken in your case if save file using webdav or http_cache?
On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com <>> wrote: Hello Sergey,
Thank you for your answer. I've looked into webdav mounted filesystem.
Unfortunately, most WebDav clients (especially davfs2 on debian) do a lot of buffering, caching and even lock-null requests (lock a non existent resource before writing to it). I also suspect that they wait for the end of the write operation. The result is a latency of a few seconds witch is not much better than what I achieve with mod_shout if I transcode the MP3.
Any other idea ?
Thank you !
Best regards
Vincent
On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov <s.safarov@gmail.com <>> wrote: Please look at webdav mounted filesystem.
On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com <>> wrote: Hi all,
Thank you to all contributing to FreeSWITCH !
I'm working on a IVR project where logic is implemented on a HTTP server. We are leaving Twilio because we now need to record and live stream the session to the HTTP server in WAV format (chunked transfer encoding).
mod_httapi looks great (HT TAPI very similar to Twilio's) but it seems that the records are first saved to disk before there are sent to the server as chunked data. We need the transfer to start as soon as the recording starts.
mod_shout does start the request almost as the records starts but it does not support WAV file and shout:// is not exactly a HTTP request (SOURCE method instead of PUT).
Is there a way to use these modules to achieve our goal ?
If not, we are willing to author a specific module or rather contribute to the existing ones.
We've identified two approaches: From mod_httapi Modify mod_httapi to directly stream the record instead of completely saving it to disk before the HTTP chunked transfer starts. This seems the most logical but with more than 3000 lines, mod_httapi does not seem to be the easiest module to build upon for newcomers!
From mod_shout Modify libshoot to replace the custom SOURCE method with standard HTTP PUT method Modify mod_shout to support wav files Implement our IVR in script (javascript/lua) What do you think ?
Thank you for your help.
Cheers,
_________________________________________________________________________ Professional FreeSWITCH Consulting Services: 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 <> 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/>
-- Vincent Gire _________________________________________________________________________ Professional FreeSWITCH Consulting Services: 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 <> 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/> _________________________________________________________________________ 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/>
-- Vincent Gire _________________________________________________________________________ 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
We already handle the multiple providers on the HTTP server. So you would recommend to implement a module against the speech interface streaming to the HTTP server ? Wouldn't it be easier (and less redundant) to add streaming capabilities to mod_httpapi ? On Wed, Apr 6, 2016 at 12:22 AM, Michael Jerris <mike@jerris.com> wrote:
The best way to handle this is probably to write a custom module against the speech interface that sends the streams like you are describing, and supports multiple providers. How exactly do you go about sending to multiple at the same time and combining the results, as this might actually be the trickiest part as youll need to integrate that into a module that uses the speech interface to be able to at all sanely handle that in freeswitch
On Apr 5, 2016, at 1:24 PM, Vincent Gire <vincent.gire@gmail.com> wrote:
We use multiple ones to compare confidence. They all support chunk transfer and provide a HTTP transaction similar to : http://developer.att.com/apis/speech/docs/v3 Look at 3/ Make API Calls
On Tue, Apr 5, 2016 at 7:14 PM, Michael Jerris <mike@jerris.com> wrote:
Unimrcp isn't going to provide the interface he's talking about. Can I ask which engine this is?
On Tuesday, April 5, 2016, Sergey Safarov <s.safarov@gmail.com> wrote:
Look at http://www.unimrcp.org/ and https://wiki.freeswitch.org/wiki/Mod_unimrcp
On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com> wrote:
We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ?
On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com> wrote:
I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com> wrote:
Think is requred streaming feature of freeswitch. Look at mod_esf <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> Instruction to compile mod_vlc on provided link is to old but helpfull to undestand how to stream media to http server.
For compiling mod_vlc please use vlc repo <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos instruction <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> . After you intall vlc, then you can enable mod_vlc module in freeswitch sources(SPEC file) and compile freeswitch.
Sergey
вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>:
> webdav, mod_http_cache or mod_httapi all results in sending the > recording only *after* it is complete. > They all write the recording to a file, wait for the recording to > complete and the file to close and then send it over HTTP. > > I would like to start sending the recording to the remove server as > soon as it starts (max 1 sec latency). > mod_http_cache or mod_httapi would be perfect if they were streaming > the recording like mod_shout. > > > On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com> > wrote: > >> Input/output latency is not problem. I use Kazoo on my servers and >> call recording is stored to database during 5 seconds after hangup. >> What is broken in your case if save file using webdav or http_cache? >> >> On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> >> wrote: >> >>> Hello Sergey, >>> >>> Thank you for your answer. >>> I've looked into webdav mounted filesystem. >>> >>> Unfortunately, most WebDav clients (especially davfs2 on debian) >>> do a lot of buffering, caching and even lock-null requests (lock a non >>> existent resource before writing to it). I also suspect that they wait for >>> the end of the write operation. >>> The result is a latency of a few seconds witch is not much better >>> than what I achieve with mod_shout if I transcode the MP3. >>> >>> Any other idea ? >>> >>> Thank you ! >>> >>> Best regards >>> >>> Vincent >>> >>> On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov < >>> s.safarov@gmail.com> wrote: >>> >>>> Please look at webdav mounted filesystem. >>>> >>>> On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> >>>> wrote: >>>> >>>>> Hi all, >>>>> >>>>> Thank you to all contributing to FreeSWITCH ! >>>>> >>>>> I'm working on a IVR project where logic is implemented on a >>>>> HTTP server. >>>>> We are leaving Twilio because we now need to record and live >>>>> stream the session to the HTTP server in WAV format (chunked transfer >>>>> encoding). >>>>> >>>>> *mod_httapi* looks great (HT TAPI very similar to Twilio's) but >>>>> it seems that the records are first saved to disk before there are sent to >>>>> the server as chunked data. >>>>> We need the transfer to start as soon as the recording starts. >>>>> >>>>> *mod_shout* does start the request almost as the records starts >>>>> but it does not support WAV file and shout:// is not exactly a HTTP request >>>>> (SOURCE method instead of PUT). >>>>> >>>>> Is there a way to use these modules to achieve our goal ? >>>>> >>>>> If not, we are willing to author a specific module or rather >>>>> contribute to the existing ones. >>>>> >>>>> We've identified two approaches: >>>>> >>>>> 1. From *mod_httapi* Modify mod_httapi to directly stream >>>>> the record instead of completely saving it to disk before the HTTP chunked >>>>> transfer starts. >>>>> This seems the most logical but with more than 3000 lines, >>>>> mod_httapi does not seem to be the easiest module to build upon for >>>>> newcomers! >>>>> >>>>> 2. From *mod_shout* >>>>> 1. Modify libshoot to replace the custom SOURCE method >>>>> with standard HTTP PUT method >>>>> 2. Modify mod_shout to support wav files >>>>> 3. Implement our IVR in script (javascript/lua) >>>>> >>>>> What do you think ? >>>>> >>>>> Thank you for your help. >>>>> >>>>> Cheers, >>>>> >>>>
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
Firstly, You cannot stream WAV. Its not possible. WAV is not a streaming format, its a container. Its basically a raw audio data stream with a header explaining the characteristics of the stream. Next, You need to compare your specific needs to the goal of the software to be flexible and usable to all. The reason the speech interface was suggested is that the point of the speech interface is precisely to tap into live audio streams and perform asynchronous ASR operations while the original call does whatever it wants like talk to another caller or listen to a file playing etc. What you envision is more of a synchronous single instruction to record a file in realtime to the remote server over httapi. The considerations to implement this would be what format to send the audio in. The source audio could be any combination of sample rate, channel count and encoding format. To record we would translate the encoded format into signed linear at the original input sample rate and channels. These variances must be communicated before you start sending the data so it would be a matter of figuring out an existing standard or implement a proprietary one. On Wed, Apr 6, 2016 at 10:42 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
We already handle the multiple providers on the HTTP server. So you would recommend to implement a module against the speech interface streaming to the HTTP server ? Wouldn't it be easier (and less redundant) to add streaming capabilities to mod_httpapi ?
On Wed, Apr 6, 2016 at 12:22 AM, Michael Jerris <mike@jerris.com> wrote:
The best way to handle this is probably to write a custom module against the speech interface that sends the streams like you are describing, and supports multiple providers. How exactly do you go about sending to multiple at the same time and combining the results, as this might actually be the trickiest part as youll need to integrate that into a module that uses the speech interface to be able to at all sanely handle that in freeswitch
On Apr 5, 2016, at 1:24 PM, Vincent Gire <vincent.gire@gmail.com> wrote:
We use multiple ones to compare confidence. They all support chunk transfer and provide a HTTP transaction similar to : http://developer.att.com/apis/speech/docs/v3 Look at 3/ Make API Calls
On Tue, Apr 5, 2016 at 7:14 PM, Michael Jerris <mike@jerris.com> wrote:
Unimrcp isn't going to provide the interface he's talking about. Can I ask which engine this is?
On Tuesday, April 5, 2016, Sergey Safarov <s.safarov@gmail.com> wrote:
Look at http://www.unimrcp.org/ and https://wiki.freeswitch.org/wiki/Mod_unimrcp
On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com> wrote:
We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ?
On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com> wrote:
I would stay away from mod_vlc. Its audio portions with recording have known issues. We do use the rtmp streaming in mod_av heavily but thats obviously not wav. Can you explain a bit more why you have this requirement?
On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
Ok thanks. It looks promising ! I'll dig into mod_vlc.
Best
Vincent
On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com> wrote:
> Think is requred streaming feature of freeswitch. > Look at mod_esf > <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and > mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> > Instruction to compile mod_vlc on provided link is to old but > helpfull to undestand how to stream media to http server. > > For compiling mod_vlc please use vlc repo > <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos > instruction > <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> > . > After you intall vlc, then you can enable mod_vlc module in > freeswitch sources(SPEC file) and compile freeswitch. > > Sergey > > > вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>: > >> webdav, mod_http_cache or mod_httapi all results in sending the >> recording only *after* it is complete. >> They all write the recording to a file, wait for the recording to >> complete and the file to close and then send it over HTTP. >> >> I would like to start sending the recording to the remove server as >> soon as it starts (max 1 sec latency). >> mod_http_cache or mod_httapi would be perfect if they were >> streaming the recording like mod_shout. >> >> >> On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov <s.safarov@gmail.com >> > wrote: >> >>> Input/output latency is not problem. I use Kazoo on my servers >>> and call recording is stored to database during 5 seconds after hangup. >>> What is broken in your case if save file using webdav or >>> http_cache? >>> >>> On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> >>> wrote: >>> >>>> Hello Sergey, >>>> >>>> Thank you for your answer. >>>> I've looked into webdav mounted filesystem. >>>> >>>> Unfortunately, most WebDav clients (especially davfs2 on debian) >>>> do a lot of buffering, caching and even lock-null requests (lock a non >>>> existent resource before writing to it). I also suspect that they wait for >>>> the end of the write operation. >>>> The result is a latency of a few seconds witch is not much better >>>> than what I achieve with mod_shout if I transcode the MP3. >>>> >>>> Any other idea ? >>>> >>>> Thank you ! >>>> >>>> Best regards >>>> >>>> Vincent >>>> >>>> On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov < >>>> s.safarov@gmail.com> wrote: >>>> >>>>> Please look at webdav mounted filesystem. >>>>> >>>>> On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> >>>>> wrote: >>>>> >>>>>> Hi all, >>>>>> >>>>>> Thank you to all contributing to FreeSWITCH ! >>>>>> >>>>>> I'm working on a IVR project where logic is implemented on a >>>>>> HTTP server. >>>>>> We are leaving Twilio because we now need to record and live >>>>>> stream the session to the HTTP server in WAV format (chunked transfer >>>>>> encoding). >>>>>> >>>>>> *mod_httapi* looks great (HT TAPI very similar to Twilio's) >>>>>> but it seems that the records are first saved to disk before there are sent >>>>>> to the server as chunked data. >>>>>> We need the transfer to start as soon as the recording starts. >>>>>> >>>>>> *mod_shout* does start the request almost as the records >>>>>> starts but it does not support WAV file and shout:// is not exactly a HTTP >>>>>> request (SOURCE method instead of PUT). >>>>>> >>>>>> Is there a way to use these modules to achieve our goal ? >>>>>> >>>>>> If not, we are willing to author a specific module or rather >>>>>> contribute to the existing ones. >>>>>> >>>>>> We've identified two approaches: >>>>>> >>>>>> 1. From *mod_httapi* Modify mod_httapi to directly stream >>>>>> the record instead of completely saving it to disk before the HTTP chunked >>>>>> transfer starts. >>>>>> This seems the most logical but with more than 3000 lines, >>>>>> mod_httapi does not seem to be the easiest module to build upon for >>>>>> newcomers! >>>>>> >>>>>> 2. From *mod_shout* >>>>>> 1. Modify libshoot to replace the custom SOURCE method >>>>>> with standard HTTP PUT method >>>>>> 2. Modify mod_shout to support wav files >>>>>> 3. Implement our IVR in script (javascript/lua) >>>>>> >>>>>> What do you think ? >>>>>> >>>>>> Thank you for your help. >>>>>> >>>>>> Cheers, >>>>>> >>>>>
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
-- Anthony Minessale II ♬ @anthmfs ♬ @FreeSWITCH ♬ ☞ http://freeswitch.org/ ☞ http://cluecon.com/ ☞ http://twitter.com/FreeSWITCH ☞ irc.freenode.net #freeswitch ☞ *http://freeswitch.org/g+ <http://freeswitch.org/g+>* ClueCon Weekly Development Call ☎ sip:888@conference.freeswitch.org ☎ +19193869900 https://www.youtube.com/watch?v=9XXgW34t40s https://www.youtube.com/watch?v=NLaDpGQuZDA
On Wed, Apr 6, 2016 at 6:10 PM, Anthony Minessale < anthony.minessale@gmail.com> wrote:
Firstly,
You cannot stream WAV. Its not possible. WAV is not a streaming format, its a container. Its basically a raw audio data stream with a header explaining the characteristics of the stream.
Yes. Sorry for the language abuse and thank you for clearing this out.
Next,
You need to compare your specific needs to the goal of the software to be flexible and usable to all.
The reason the speech interface was suggested is that the point of the speech interface is precisely to tap into live audio streams and perform asynchronous ASR operations while the original call does whatever it wants like talk to another caller or listen to a file playing etc.
What you envision is more of a synchronous single instruction to record a file in realtime to the remote server over httapi.
Exactly.
The considerations to implement this would be what format to send the audio in. The source audio could be any combination of sample rate, channel count and encoding format. To record we would translate the encoded format into signed linear at the original input sample rate and channels. These variances must be communicated before you start sending the data so it would be a matter of figuring out an existing standard or implement a proprietary one.
I am not sure I am following here. I might again have misused the word streaming. I was thinking that when we specify a http target for a record in mod_httapi, it could start sending the corresponding file (WAV or other) to the HTTP server as chunked data as soon as the record starts. In the case of a WAV target, the WAV header would be sent first as part of the first HTTP data chunks and the raw audio data stream would follow in the next data chunks. Does it make sense ?
On Wed, Apr 6, 2016 at 10:42 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
We already handle the multiple providers on the HTTP server. So you would recommend to implement a module against the speech interface streaming to the HTTP server ? Wouldn't it be easier (and less redundant) to add streaming capabilities to mod_httpapi ?
On Wed, Apr 6, 2016 at 12:22 AM, Michael Jerris <mike@jerris.com> wrote:
The best way to handle this is probably to write a custom module against the speech interface that sends the streams like you are describing, and supports multiple providers. How exactly do you go about sending to multiple at the same time and combining the results, as this might actually be the trickiest part as youll need to integrate that into a module that uses the speech interface to be able to at all sanely handle that in freeswitch
On Apr 5, 2016, at 1:24 PM, Vincent Gire <vincent.gire@gmail.com> wrote:
We use multiple ones to compare confidence. They all support chunk transfer and provide a HTTP transaction similar to : http://developer.att.com/apis/speech/docs/v3 Look at 3/ Make API Calls
On Tue, Apr 5, 2016 at 7:14 PM, Michael Jerris <mike@jerris.com> wrote:
Unimrcp isn't going to provide the interface he's talking about. Can I ask which engine this is?
On Tuesday, April 5, 2016, Sergey Safarov <s.safarov@gmail.com> wrote:
Look at http://www.unimrcp.org/ and https://wiki.freeswitch.org/wiki/Mod_unimrcp
On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com> wrote:
We are building an IVR completely driven by ASR. ASR is performed in a distant location by a HTTP service (supporting chunked transfer) and adds an incompressible latency. We would like to stream the record to the ASR service as soon as it starts to reduce the overall latency before response. Does it make sense ?
On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com> wrote:
> I would stay away from mod_vlc. Its audio portions with recording > have known issues. We do use the rtmp streaming in mod_av heavily but > thats obviously not wav. Can you explain a bit more why you have this > requirement? > > On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> > wrote: > > Ok thanks. > It looks promising ! > I'll dig into mod_vlc. > > Best > > Vincent > > On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov <s.safarov@gmail.com > > wrote: > >> Think is requred streaming feature of freeswitch. >> Look at mod_esf >> <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> and >> mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> >> Instruction to compile mod_vlc on provided link is to old but >> helpfull to undestand how to stream media to http server. >> >> For compiling mod_vlc please use vlc repo >> <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos >> instruction >> <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> >> . >> After you intall vlc, then you can enable mod_vlc module in >> freeswitch sources(SPEC file) and compile freeswitch. >> >> Sergey >> >> >> вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>: >> >>> webdav, mod_http_cache or mod_httapi all results in sending the >>> recording only *after* it is complete. >>> They all write the recording to a file, wait for the recording to >>> complete and the file to close and then send it over HTTP. >>> >>> I would like to start sending the recording to the remove server >>> as soon as it starts (max 1 sec latency). >>> mod_http_cache or mod_httapi would be perfect if they were >>> streaming the recording like mod_shout. >>> >>> >>> On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov < >>> s.safarov@gmail.com> wrote: >>> >>>> Input/output latency is not problem. I use Kazoo on my servers >>>> and call recording is stored to database during 5 seconds after hangup. >>>> What is broken in your case if save file using webdav or >>>> http_cache? >>>> >>>> On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> >>>> wrote: >>>> >>>>> Hello Sergey, >>>>> >>>>> Thank you for your answer. >>>>> I've looked into webdav mounted filesystem. >>>>> >>>>> Unfortunately, most WebDav clients (especially davfs2 on debian) >>>>> do a lot of buffering, caching and even lock-null requests (lock a non >>>>> existent resource before writing to it). I also suspect that they wait for >>>>> the end of the write operation. >>>>> The result is a latency of a few seconds witch is not much >>>>> better than what I achieve with mod_shout if I transcode the MP3. >>>>> >>>>> Any other idea ? >>>>> >>>>> Thank you ! >>>>> >>>>> Best regards >>>>> >>>>> Vincent >>>>> >>>>> On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov < >>>>> s.safarov@gmail.com> wrote: >>>>> >>>>>> Please look at webdav mounted filesystem. >>>>>> >>>>>> On Sun, Apr 3, 2016, 19:17 Vincent Gire <vincent.gire@gmail.com> >>>>>> wrote: >>>>>> >>>>>>> Hi all, >>>>>>> >>>>>>> Thank you to all contributing to FreeSWITCH ! >>>>>>> >>>>>>> I'm working on a IVR project where logic is implemented on a >>>>>>> HTTP server. >>>>>>> We are leaving Twilio because we now need to record and live >>>>>>> stream the session to the HTTP server in WAV format (chunked transfer >>>>>>> encoding). >>>>>>> >>>>>>> *mod_httapi* looks great (HT TAPI very similar to Twilio's) >>>>>>> but it seems that the records are first saved to disk before there are sent >>>>>>> to the server as chunked data. >>>>>>> We need the transfer to start as soon as the recording starts. >>>>>>> >>>>>>> *mod_shout* does start the request almost as the records >>>>>>> starts but it does not support WAV file and shout:// is not exactly a HTTP >>>>>>> request (SOURCE method instead of PUT). >>>>>>> >>>>>>> Is there a way to use these modules to achieve our goal ? >>>>>>> >>>>>>> If not, we are willing to author a specific module or rather >>>>>>> contribute to the existing ones. >>>>>>> >>>>>>> We've identified two approaches: >>>>>>> >>>>>>> 1. From *mod_httapi* Modify mod_httapi to directly stream >>>>>>> the record instead of completely saving it to disk before the HTTP chunked >>>>>>> transfer starts. >>>>>>> This seems the most logical but with more than 3000 lines, >>>>>>> mod_httapi does not seem to be the easiest module to build upon for >>>>>>> newcomers! >>>>>>> >>>>>>> 2. From *mod_shout* >>>>>>> 1. Modify libshoot to replace the custom SOURCE method >>>>>>> with standard HTTP PUT method >>>>>>> 2. Modify mod_shout to support wav files >>>>>>> 3. Implement our IVR in script (javascript/lua) >>>>>>> >>>>>>> What do you think ? >>>>>>> >>>>>>> Thank you for your help. >>>>>>> >>>>>>> Cheers, >>>>>>> >>>>>> > > _________________________________________________________________________ > 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 >
-- Vincent Gire
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire _________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
-- Anthony Minessale II ♬ @anthmfs ♬ @FreeSWITCH ♬
☞ http://freeswitch.org/ ☞ http://cluecon.com/ ☞ http://twitter.com/FreeSWITCH ☞ irc.freenode.net #freeswitch ☞ *http://freeswitch.org/g+ <http://freeswitch.org/g+>*
ClueCon Weekly Development Call ☎ sip:888@conference.freeswitch.org ☎ +19193869900
https://www.youtube.com/watch?v=9XXgW34t40s https://www.youtube.com/watch?v=NLaDpGQuZDA
_________________________________________________________________________ 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
-- Vincent Gire
On Wed, Apr 6, 2016 at 11:47 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
On Wed, Apr 6, 2016 at 6:10 PM, Anthony Minessale < anthony.minessale@gmail.com> wrote:
Firstly,
You cannot stream WAV. Its not possible. WAV is not a streaming format, its a container. Its basically a raw audio data stream with a header explaining the characteristics of the stream.
Yes. Sorry for the language abuse and thank you for clearing this out.
Next,
You need to compare your specific needs to the goal of the software to be flexible and usable to all.
The reason the speech interface was suggested is that the point of the speech interface is precisely to tap into live audio streams and perform asynchronous ASR operations while the original call does whatever it wants like talk to another caller or listen to a file playing etc.
What you envision is more of a synchronous single instruction to record a file in realtime to the remote server over httapi.
Exactly.
The considerations to implement this would be what format to send the audio in. The source audio could be any combination of sample rate, channel count and encoding format. To record we would translate the encoded format into signed linear at the original input sample rate and channels. These variances must be communicated before you start sending the data so it would be a matter of figuring out an existing standard or implement a proprietary one.
I am not sure I am following here. I might again have misused the word streaming. I was thinking that when we specify a http target for a record in mod_httapi, it could start sending the corresponding file (WAV or other) to the HTTP server as chunked data as soon as the record starts. In the case of a WAV target, the WAV header would be sent first as part of the first HTTP data chunks and the raw audio data stream would follow in the next data chunks. Does it make sense ?
I agree that you are not following as I gave a relatively lengthy and specific response. I know exactly what you are asking for so there is no need to re-explain. The way you ask for it is very specific to your own needs so it would not make a very useful feature in FS. Are you asking for someone to make this feature for you or are you asking for advice on how you can implement it? I think you are underestimating the complexity and potential scalability of the overall project you are doing (not just the FS portion) and you should review your entire architecture.
On Wed, Apr 6, 2016 at 10:42 AM, Vincent Gire <vincent.gire@gmail.com> wrote:
We already handle the multiple providers on the HTTP server. So you would recommend to implement a module against the speech interface streaming to the HTTP server ? Wouldn't it be easier (and less redundant) to add streaming capabilities to mod_httpapi ?
On Wed, Apr 6, 2016 at 12:22 AM, Michael Jerris <mike@jerris.com> wrote:
The best way to handle this is probably to write a custom module against the speech interface that sends the streams like you are describing, and supports multiple providers. How exactly do you go about sending to multiple at the same time and combining the results, as this might actually be the trickiest part as youll need to integrate that into a module that uses the speech interface to be able to at all sanely handle that in freeswitch
On Apr 5, 2016, at 1:24 PM, Vincent Gire <vincent.gire@gmail.com> wrote:
We use multiple ones to compare confidence. They all support chunk transfer and provide a HTTP transaction similar to : http://developer.att.com/apis/speech/docs/v3 Look at 3/ Make API Calls
On Tue, Apr 5, 2016 at 7:14 PM, Michael Jerris <mike@jerris.com> wrote:
Unimrcp isn't going to provide the interface he's talking about. Can I ask which engine this is?
On Tuesday, April 5, 2016, Sergey Safarov <s.safarov@gmail.com> wrote:
Look at http://www.unimrcp.org/ and https://wiki.freeswitch.org/wiki/Mod_unimrcp
On Tue, Apr 5, 2016, 19:38 Vincent Gire <vincent.gire@gmail.com> wrote:
> We are building an IVR completely driven by ASR. > ASR is performed in a distant location by a HTTP service (supporting > chunked transfer) and adds an incompressible latency. We would like to > stream the record to the ASR service as soon as it starts to reduce the > overall latency before response. > Does it make sense ? > > > On Tue, Apr 5, 2016 at 5:39 PM, Michael Jerris <mike@jerris.com> > wrote: > >> I would stay away from mod_vlc. Its audio portions with recording >> have known issues. We do use the rtmp streaming in mod_av heavily but >> thats obviously not wav. Can you explain a bit more why you have this >> requirement? >> >> On Apr 5, 2016, at 5:36 AM, Vincent Gire <vincent.gire@gmail.com> >> wrote: >> >> Ok thanks. >> It looks promising ! >> I'll dig into mod_vlc. >> >> Best >> >> Vincent >> >> On Tue, Apr 5, 2016 at 10:04 AM, Sergey Safarov < >> s.safarov@gmail.com> wrote: >> >>> Think is requred streaming feature of freeswitch. >>> Look at mod_esf >>> <https://freeswitch.org/confluence/display/FREESWITCH/mod_esf> >>> and mod_vlc <https://wiki.freeswitch.org/wiki/Mod_vlc> >>> Instruction to compile mod_vlc on provided link is to old but >>> helpfull to undestand how to stream media to http server. >>> >>> For compiling mod_vlc please use vlc repo >>> <https://freeswitch.org/stash/projects/SD/repos/vlc/browse> and centos >>> instruction >>> <https://freeswitch.org/confluence/display/FREESWITCH/Creating+RPM+repositary> >>> . >>> After you intall vlc, then you can enable mod_vlc module in >>> freeswitch sources(SPEC file) and compile freeswitch. >>> >>> Sergey >>> >>> >>> вт, 5 апр. 2016 г. в 10:31, Vincent Gire <vincent.gire@gmail.com>: >>> >>>> webdav, mod_http_cache or mod_httapi all results in sending the >>>> recording only *after* it is complete. >>>> They all write the recording to a file, wait for the recording >>>> to complete and the file to close and then send it over HTTP. >>>> >>>> I would like to start sending the recording to the remove server >>>> as soon as it starts (max 1 sec latency). >>>> mod_http_cache or mod_httapi would be perfect if they were >>>> streaming the recording like mod_shout. >>>> >>>> >>>> On Mon, Apr 4, 2016 at 8:50 PM, Sergey Safarov < >>>> s.safarov@gmail.com> wrote: >>>> >>>>> Input/output latency is not problem. I use Kazoo on my servers >>>>> and call recording is stored to database during 5 seconds after hangup. >>>>> What is broken in your case if save file using webdav or >>>>> http_cache? >>>>> >>>>> On Mon, Apr 4, 2016, 21:10 Vincent Gire <vincent.gire@gmail.com> >>>>> wrote: >>>>> >>>>>> Hello Sergey, >>>>>> >>>>>> Thank you for your answer. >>>>>> I've looked into webdav mounted filesystem. >>>>>> >>>>>> Unfortunately, most WebDav clients (especially davfs2 on >>>>>> debian) do a lot of buffering, caching and even lock-null requests (lock a >>>>>> non existent resource before writing to it). I also suspect that they wait >>>>>> for the end of the write operation. >>>>>> The result is a latency of a few seconds witch is not much >>>>>> better than what I achieve with mod_shout if I transcode the MP3. >>>>>> >>>>>> Any other idea ? >>>>>> >>>>>> Thank you ! >>>>>> >>>>>> Best regards >>>>>> >>>>>> Vincent >>>>>> >>>>>> On Sun, Apr 3, 2016 at 7:30 PM, Sergey Safarov < >>>>>> s.safarov@gmail.com> wrote: >>>>>> >>>>>>> Please look at webdav mounted filesystem. >>>>>>> >>>>>>> On Sun, Apr 3, 2016, 19:17 Vincent Gire < >>>>>>> vincent.gire@gmail.com> wrote: >>>>>>> >>>>>>>> Hi all, >>>>>>>> >>>>>>>> Thank you to all contributing to FreeSWITCH ! >>>>>>>> >>>>>>>> I'm working on a IVR project where logic is implemented on a >>>>>>>> HTTP server. >>>>>>>> We are leaving Twilio because we now need to record and live >>>>>>>> stream the session to the HTTP server in WAV format (chunked transfer >>>>>>>> encoding). >>>>>>>> >>>>>>>> *mod_httapi* looks great (HT TAPI very similar to Twilio's) >>>>>>>> but it seems that the records are first saved to disk before there are sent >>>>>>>> to the server as chunked data. >>>>>>>> We need the transfer to start as soon as the recording starts. >>>>>>>> >>>>>>>> *mod_shout* does start the request almost as the records >>>>>>>> starts but it does not support WAV file and shout:// is not exactly a HTTP >>>>>>>> request (SOURCE method instead of PUT). >>>>>>>> >>>>>>>> Is there a way to use these modules to achieve our goal ? >>>>>>>> >>>>>>>> If not, we are willing to author a specific module or rather >>>>>>>> contribute to the existing ones. >>>>>>>> >>>>>>>> We've identified two approaches: >>>>>>>> >>>>>>>> 1. From *mod_httapi* Modify mod_httapi to directly stream >>>>>>>> the record instead of completely saving it to disk before the HTTP chunked >>>>>>>> transfer starts. >>>>>>>> This seems the most logical but with more than 3000 >>>>>>>> lines, mod_httapi does not seem to be the easiest module to build upon for >>>>>>>> newcomers! >>>>>>>> >>>>>>>> 2. From *mod_shout* >>>>>>>> 1. Modify libshoot to replace the custom SOURCE method >>>>>>>> with standard HTTP PUT method >>>>>>>> 2. Modify mod_shout to support wav files >>>>>>>> 3. Implement our IVR in script (javascript/lua) >>>>>>>> >>>>>>>> What do you think ? >>>>>>>> >>>>>>>> Thank you for your help. >>>>>>>> >>>>>>>> Cheers, >>>>>>>> >>>>>>> >> >> _________________________________________________________________________ >> 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 >> > > > > -- > Vincent Gire > > _________________________________________________________________________ > 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
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
-- Anthony Minessale II ♬ @anthmfs ♬ @FreeSWITCH ♬
☞ http://freeswitch.org/ ☞ http://cluecon.com/ ☞ http://twitter.com/FreeSWITCH ☞ irc.freenode.net #freeswitch ☞ *http://freeswitch.org/g+ <http://freeswitch.org/g+>*
ClueCon Weekly Development Call ☎ sip:888@conference.freeswitch.org ☎ +19193869900
https://www.youtube.com/watch?v=9XXgW34t40s https://www.youtube.com/watch?v=NLaDpGQuZDA
_________________________________________________________________________ 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
-- Vincent Gire
_________________________________________________________________________ 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
-- Anthony Minessale II ♬ @anthmfs ♬ @FreeSWITCH ♬ ☞ http://freeswitch.org/ ☞ http://cluecon.com/ ☞ http://twitter.com/FreeSWITCH ☞ irc.freenode.net #freeswitch ☞ *http://freeswitch.org/g+ <http://freeswitch.org/g+>* ClueCon Weekly Development Call ☎ sip:888@conference.freeswitch.org ☎ +19193869900 https://www.youtube.com/watch?v=9XXgW34t40s https://www.youtube.com/watch?v=NLaDpGQuZDA
participants (4)
-
Anthony Minessale
-
Michael Jerris
-
Sergey Safarov
-
Vincent Gire