Auto Attendant timeout of 0 takes 3 seconds

[mcbsys]

Hi,

I'm testing Vodia, working in 68.0.26. I'm trying to set up an auto-attendant for a small church. Besides dialing user extensions, it provides several prompts for getting pre-recorded info (service times, directions, youth program, web site, etc.) and one for leaving a voicemail for the pastor.  The pre-recorded info changes from time to time (like for special services), so they need to be in separate recordings, not one huge announcement.

I like the option of using text-to-speech for the prompts, not available in the IVR module. Also, I don't see in the IVR the options for direct-dialing extensions or dialing by name.  So I guess I have to do this with Auto Attendants, right?

I set up an Auto Attendant 810 to play a welcome greeting. Besides accepting user input (1-5, 7, 9, 0, or extension), it should just drop through to play the next voice prompt in attendant 820.  So I set 810's Timeout handling to 0 seconds and turned off the three options under IVR > Entering the destination.  The problem is, with the timeout set to 0, it waits 3 seconds before playing the 820 prompt.

How do I "chain" auto-attendants together with no delay?

The Direct Destinations for each Auto Attendant in the chain need to be the same--I don't want to force callers to listen to everything before they can press buttons.  Can I clone Auto Attendants?

Thanks,

Mark

[RichardDCG]

for no delay in AA I have the Redirect set to the next AA number, 0 for timeout and hangup timeout, 1 for repeats and digit timeout.  This seems to work for me. On V68.0.24

[mcbsys]

@RichardDCG, thank you for your reply. The only difference in my setup is that I had Hangup Time 0 and Repeats = Repeat (both defaults I believe). But even after changing to 0 and 1 respectively, I still have a 2-3 second delay after each prompt.

Are you using pre-recorded prompts or are you using text-to-speech? I'm using text-to-speech. It occurred to me that the speech engine might be appending silence at the end of the recording, and in fact I do see about 0.3 - 0.5 seconds of silence at the end of the recordings, but the delay seems longer.

[RichardDCG]

I use pre-recorded.  I then manage it by adding extensions to Record audio for groups. Then update the message as required from a phone.

 

Is your setup - a call is answered on AA 810 that plays a message and please select from  1,2,3,4,5?  Selecting 1,2,3,4,or 5 then moves to another AA e.g. 820 and plays another message and so on?  The delay being after a key selection?

 

I also had a small hiss before answering calls that was apparently 'comfort noise' ... maybe part of your gap? There is a file noise.wav in the audio_moh directory that I changed for a blank file and now there is no hiss.

[mcbsys]

Thanks. Tried switching to pre-recorded and I still have the short delay. I'm not hearing static though. This is a "fall-through" or "waterfall" part of the AA:

800 Welcome (scheduled service flag varies message for open or closed)
810 Worship Times (a different, manual service flag varies message if it's a special week)
820 Please Press 1, 2, 3, etc.

I haven't finished setting up 820 yet; I'm listening to the transition from 800 to 810 and 810 to 820.

I turned media logging up to 9 and see this in the System log when I call 800 from a desk phone x.200:

[6] 8:42:47.340	Port 53: Sending RTP to 162.199.111.222:12608, codec PCMU/8000ⓘ
[6] 8:42:47.340	Last message repeated 2 timesⓘ
[9] 8:42:47.340	Port 53: Received first RTP packetⓘ
[9] 8:42:48.790	Last message repeated 2 timesⓘ
[9] 8:42:48.790	Port 53: RTCP SR time=658361:730150000 timestamp=3885714988 packets=73 octets=11680ⓘ
[9] 8:42:53.789	Last message repeated 2 timesⓘ
[9] 8:42:53.789	Port 53: RTCP SR time=658366:730150000 timestamp=3885754988 packets=323 octets=51680ⓘ
[8] 8:42:55.934	Media: Dropping audio_en/aa_enter_extension_number.wav from cacheⓘ
[8] 8:42:55.934	Media: Dropping audio_en/synth/efef07e31f8f8dd5c507691716593307.wav from cacheⓘ
[8] 8:42:55.934	Media: Dropping recordings/att9.wav from cacheⓘ
[4] 8:42:57.934	Call 43: Dial number 810 from user 800 and dial plan Defaultⓘ
[8] 8:42:57.935	Port 53: state code from 200 to 200ⓘ
[8] 8:42:58.581	Port 53: Clearing port with SIP Call-ID 0_885738201@192.168.1.141ⓘ

It doesn't give a start and stop time for each .wav file it plays. Oddly, att9.wav is the pre-recorded message for x.810, so why is it dropped from cache before transferring to 810? I do wonder what it's doing from 8:42:55 to 8:42:57.

The delay isn't awful, just odd. I was just hoping there was something I missed in terms of how to configure for no delay.

[Vodia PBX]

If you are synthesizing text to WAV this will always take a little while the first time — but then it in the file system and the delay should be negligible. Setting up two-way media also takes a while. Anyhow, I would generate a PCAP and there you can see in all detail what is going on.

[mcbsys]

Well that's pretty cool--I hadn't tried a PCAP from Vodia before. I generated one at the extension level. It does a nice job of showing graphically what I've been trying to describe:  a 2.5-second gap between audio files played by two AAs:

[mcbsys]

Okay I think I figured this out. On the IVR tab, the Gap time (sec) is appended to the welcome message, whether you Play default welcome message or, as I'm doing, use Service flags and Upload content. A Gap time of Default adds about 2.5 seconds after playing the welcome message. A Gap time of 0 eliminates that blank space.

In the UI, maybe Gap time could be moved down below the list of service flag recordings since it is played after whichever welcome message is used.

 

 

[RichardDCG]

there are a few areas in Vodia where there is a 'default' option, but nowhere says what that default value is or how to change it.  I tend to manually set values now rather than accept the 'default'.